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
|
|---|---|---|---|---|---|
Plasmodium falciparum is the major human malaria agent responsible for 200 to 300 million infections and one to three million deaths annually , mainly among African infants . The origin and evolution of this pathogen within the human lineage is still unresolved . A single species , P . reichenowi , which infects chimpanzees , is known to be a close sister lineage of P . falciparum . Here we report the discovery of a new Plasmodium species infecting Hominids . This new species has been isolated in two chimpanzees ( Pan troglodytes ) kept as pets by villagers in Gabon ( Africa ) . Analysis of its complete mitochondrial genome ( 5529 nucleotides including Cyt b , Cox I and Cox III genes ) reveals an older divergence of this lineage from the clade that includes P . falciparum and P . reichenowi ( ∼21±9 Myrs ago using Bayesian methods and considering that the divergence between P . falciparum and P . reichenowi occurred 4 to 7 million years ago as generally considered in the literature ) . This time frame would be congruent with the radiation of hominoids , suggesting that this Plasmodium lineage might have been present in early hominoids and that they may both have experienced a simultaneous diversification . Investigation of the nuclear genome of this new species will further the understanding of the genetic adaptations of P . falciparum to humans . The risk of transfer and emergence of this new species in humans must be now seriously considered given that it was found in two chimpanzees living in contact with humans and its close relatedness to the most virulent agent of malaria .
Malaria is a major parasitic worldwide scourge , infecting and killing several million people each year [1] . Among the numerous Plasmodium species that infect reptiles , birds and mammals , four of them are human-specific: P . falciparum , P . vivax , P . malariae and P . ovale . The most virulent agent is P . falciparum , which kills up to three million people each year , mainly in Africa [1] . In spite of persistent control efforts set up since the end of the fifties , the disease is far from being under control . Even though numerous articles are published every year about the parasite and the disease , progress in controlling malaria has been limited . Resistance has evolved against virtually all drugs currently available [2] , so that the disease frequently reemerges in different parts of the world [3] , [4] . The recent availability of complete Plasmodium genomes [5]–[9] has generated new hopes in the fight against this parasite . Thanks to their comparison we have now a far better understanding of their genomic architecture and of the genes that may help the parasite to escape the host immune response [5]–[10] . This approach remains unfortunately limited regarding the main malignant agent of malaria , P . falciparum . One problem is the lack of knowledge about other closely related apicomplexan models that can serve as reference and comparison [11] . At present , only one species , P . reichenowi is known as a close sister lineage of P . falciparum [8] , [12] , [13] . Other Plasmodium species ( P . rodhaini and P . schwetzi ) were in the past described as parasites of the African great apes ( i . e . chimpanzee and gorilla ) , but they were considered as closely related , for the first , to P . malariae and , for the second , to P . vivax [14] or P . ovale [15] , which are very divergent from P . falciparum [12] , [13] , [16] . The development of comparative genomics for P . falciparum depends therefore on obtaining additional information about the diversity of P . reichenowi [17] and other Plasmodium species parasitic to the African Great Apes-Human lineage ( the AGAH-lineage ) , currently represented by only two known species , P . falciparum and P . reichenowi . In this manuscript , we report the discovery of a new Plasmodium species infecting Hominids in Africa . This new species was isolated from two chimpanzees and is a close relative of P . falciparum , the most virulent agent of human malaria .
To explore the diversity of species belonging to the Plasmodium AGAH-lineage in Africa , we collected blood samples from 17 chimpanzees recently trapped from the wild and kept as pets in villages of Gabon by hunters and their families ( see Figure S1 ) . Considering that only the subspecies Pan troglodytes troglodytes has been found in Gabon , these 17 animals are likely to belong to this chimpanzee subspecies . Among them , two were found to be infected with Plasmodium by means of PCR assay or microscopy . The other 15 animals were found negative both by microscopy and PCR assay . For these two chimpanzees ( named B and K ) , observed parasites under microscopy were falciparum-like ( ring stages with two chromatin dots and presence of multiply-infected red blood cells [15] ) . Thick blood smears revealed low parasitemia in both individuals , approximately 300 parasites/µl for chimpanzee B and 2000 parasites/µl for chimpanzee K . For both , we amplified and sequenced the parasite's Cytochrome b ( Cyt b ) gene . The Cyt b sequences obtained were similar between the two samples ( identity of 99 . 8% based on 866 nucleotides ( nt ) ) , but different from all other Plasmodium Cyt b sequences known to date . The most similar sequences obtained using BLAST were Cyt b sequences from P . reichenowi and P . falciparum , which show 92% and 91% identity , respectively . Because the Cyt b sequences were partial , we studied the whole mitochondrial DNA ( mtDNA ) of these two new isolates ( named P . sp_K and P . sp_B ) . For isolate K , we amplified 5529 nt including three main genes: Cytochrome oxydase I ( Cox I ) , Cytochrome oxydase III ( Cox III ) and Cytochrome b ( Cyt b ) . Apart from short missing segments amounting to 420 nt , the mtDNA sequenced corresponds to the whole P . falciparum 3D7 mtDNA ( 5949 nt ) . For technical reasons ( certainly due to the very low parasitemia and degraded DNA ) , we were unable to accomplish this sequencing for isolate B . In order to determine the evolutionary relationships of this new Plasmodium relative to other species , we compared its sequence to 17 known complete Plasmodium mitochondrial genome sequences , with the bird apicomplexan parasite Leucocytozoon caulleryi as an outgroup ( see Table S1 ) . Maximum likelihood ( ML ) phylogenetic trees were reconstructed at both the nucleotide and amino acid levels on the whole mitochondrial genome sequence , considered as a single genetic unit [18] . DNA and protein analyses provided identical results: the parasite collected in chimpanzee K belongs to the AGAH-lineage but is more divergent from P . falciparum than is P . reichenowi ( Figure 1; see also Figure S3 for the tree reconstructed from the partial Cyt b sequence ( 866 nt ) and including both P . sp_K and P . sp_B ) . Over the entire mitochondrial genome , the genetic distance observed between the new taxon ( P . sp_K ) and P . falciparum ( d = 0 . 213 substitutions per nucleotide site on the ML phylogram ) or P . reichenowi ( d = 0 . 215 ) is almost four times higher than the distance observed between P . falciparum and P . reichenowi ( d = 0 . 058 ) . To estimate the divergence time of the plasmodium AGAH-lineage , we used a calibration chosen within the hominid hosts . The age of the P . falciparum/P . reichenowi split is generally considered to be similar to the one separating humans from chimpanzees [12] , [13] , [19] , that is , between four and seven million years [20]–[22] . Because of pervasive variations of mitochondrial substitution rates among malaria parasite lineages ( Figure 1 ) , a Bayesian relaxed molecular clock was used , which revealed a divergence time of 21±9 Myrs between the new Plasmodium species and the clade constituted by P . falciparum and P . reichenowi . Interestingly , this estimated time frame fits with the radiation of hominoids during the Miocene [23] . Our results suggest therefore that the plasmodium AGAH-lineage may have been present in early hominoids [23] and that this lineage may have also experienced a diversification during the early Miocene period as it occurred for their hosts [23] . Obviously , this estimated time of divergence is dependent on the calibration used . Recently , Martin and colleagues [24] suggested that the split between P . falciparum/P . reichenowi might have occurred far earlier than previously considered . They propose that P . falciparum originated from a recent transfer of P . reichenowi to humans during the last 2 . 8 Myrs [25] . Under this hypothesis , the new species would have diverged from P . falciparum/P . reichenowi about 10 Myrs ago . As divergence data are lacking for those parasites from the fossil record , we are unable to distinguish between these two hypotheses . Further data on the diversity of Plasmodium infecting great apes in Africa will certainly help resolve this particular aspect of the evolution of P . falciparum . In conclusion , we bring to light the existence of a new Plasmodium species that infects chimpanzees in Gabon . We propose to name this new species Plasmodium gaboni sp . nov . in reference to the country where we obtained it . Our discovery suggests that great apes and perhaps simian primates may host a far higher diversity of Plasmodium species in Africa than previously recognised . Beyond the interest of this new species in the understanding of the evolution of this group of parasites , its position in the AGAH-lineage as the sister-group of P . falciparum/P . reichenowi opens up the possibility of exploring lineage-specific evolution using comparative genomics , and hence , to look for the genes responsible for the adaptation of these parasites to their specific hosts . Comparison between genomes will advance understanding of the differences in pathology and the processes at work in the interaction with the vertebrate or the mosquito hosts [9] , [17] . It is thus essential to complete the nuclear genome sequence of this new species of phylogenetic importance within the AGAH-lineage , in order to enhance our knowledge of the functional genomics of human malaria parasites . Finally , this new species was discovered in two chimpanzees conserved as pets by villagers in Gabon . Given the recent history of primate to human shifts in several pathogens ( e . g . HIV [26]; Ebola [27]; for Plasmodium , the most recent involved P . knowlesi and occurred from macaques to humans in Asia [28] ) and the close proximity between P . gaboni and the most virulent agent of malaria , P . falciparum , we think that the risk of transfer of this species to humans must be seriously considered .
Blood aliquots of 17 chimpanzees were collected from different parts of Gabon ( Figure S1 ) . The samples were collected from wild-born animals kept as pets by hunters and their families . The investigation was approved by the Government of the Republic of Gabon and by the Animal Life Administration of Libreville , Gabon ( no . CITES 00956 ) . All animal work has been conducted according to relevant national and international guidelines . Blood samples were collected in 7 ml EDTA vacutainers from chimpanzees under ketamine anaesthesia . Clots and plasma were obtained by centrifugation and stored at −20°C until they were transported to the Centre International de Recherches Médicales de Franceville ( CIRMF ) , Gabon , where they were stored at −80°C until processed for testing . Total DNA ( Plasmodium and host ) was isolated and purified using the DNeasy blood kit ( Qiagen , Hilden , Germany ) according to the manufacturer's instructions . The DNA was eluted in 100 µl of sterile water . Microscopic analyses of the blood samples by thin blood smears revealed that two chimpanzees were infected by falciparum-like parasites ( ring stages with two chromatin dots and presence of multiply-infected red blood cells [15] ) . The two infected individuals were respectively a young chimpanzee ( Pan troglodytes , male of 3 years old ) from the area of Bakoumba ( 1°28′0″S/13°0′0″E , Haut-Ogooué Province ) and a young chimpanzee ( Pan troglodytes , male of 2–3 years old ) from the village of Koulamoutou ( 1°23′59″S/12°13′0″E , Ogooué-Lolo Province ) ( Figure S1 ) . The two Plasmodium isolates were named P . sp_B and P . sp_K , respectively . In both chimpanzees ( named B and K ) , parasitemia were estimated using thick blood smears . The number of parasites per microliter was estimated using the ratio of the number of observed parasites by the number of observed lymphocytes . To extrapolate to the number of parasites per microliter , we considered that there were about 8000 lymphocytes per microliter of whole blood . The infection status of the two chimpanzees was then confirmed by PCR of the Cytochrome b gene ( see below for PCR conditions ) . To obtain whole mitochondrial genome sequences from P . sp_B and P . sp_K , seven primer pairs were designed based on the mitochondrial genome sequence of Plasmodium falciparum ( GenBank accession number AY282930 ) using Primer 3 ( v . 0 . 4 . 0 ) [29] and eight primer pairs already published [30] ( Table S2 ) . PCR amplification and sequencing were performed on seven and eight overlapping regions covering a complete linear copy of the mtDNA genome , respectively . The Cytochrome oxydase I ( Cox I ) , the Cytochrome oxydase III ( Cox III ) and the Cytochrome b ( Cyt b ) genes were also amplified specifically using published primers [30]–[32] . All amplification reactions were performed using a MJ Research PTC100 thermal cycler . The amplified products ( 5 µl ) were run on 1 . 5% agarose gels in TAE buffer to detect the correct band . The PCR-amplified products were used as templates for sequencing . DNA sequencing was performed by CoGenics Genome Express ( Meylan , France ) . For P . sp_B , we were only able to amplify and sequence a part of Cyt b ( 866 nt ) ( deposited in the GenBankTM Database under the accession number FJ895308 ) . For P . sp_K , sequences obtained from all primer datasets were aligned and compared using ClustalW ( v 1 . 8 . 1 in BioEdit v . 7 . 0 . 9 . 0 . software [33] ) and a mtDNA consensus sequence of P . sp_K was created ( GenBank Accession number FJ895307; see Figure S2 ) . In comparative analyses , we used 17 previously published mitochondrial genome sequences from P . falciparum , P . reichenowi , P . gallinaceum , P . juxtanucleare , P . knowlesi , P . simiovale , P . simium , P . vivax , P . cynomolgi , P . yoelii , P . berghei , P . chabaudi , P . ovale , P . malariae , P . gonderi , and P . sp . DAJ-2004 , with Leucocytozoon caulleryi , an avian malaria parasite used here as outgroup . Hosts and GenBank accession numbers for these taxa are given in the Table S1 . The multiple alignment of the 18 sequences was conducted using ClustalW ( see e . g . Figure S2 ) . Phylogenetic relationships between mtDNA haplotypes ( for whole mitochondrial genome ) were inferred from all codon positions and non-coding regions . Non-sequenced sites and sites with gaps ( or missing sites ) ( when gaps were present in more than 5% of the species ) were removed , yielding a total of 5 805 sites available for subsequent inferences . Maximum Likelihood ( ML ) tree reconstruction was conducted from the whole mitochondrial genome . For this , Cyt b , Cox I , Cox III , and non-coding sequences were concatenated and analysed under a single model of nucleotide or amino acid evolution . The best-fitting ML model under the Akaike Information Criterion was GTR ( General Time Reversible ) +Γ ( Gamma distribution ) +I ( Invariable sites's distribution ) for nucleotides as identified by ModelTest [34] and mtART ( replacement matrix developed for arthropod mitochondrial proteins ) +Γ+I for amino acids as identified by ProtTest [35] . The highest-likelihood DNA and protein trees and corresponding bootstrap support values were obtained by PhyML ( freely available at the ATGC bioinformatics platform http://www . atgc-montpellier . fr/ ) using NNI ( Nearest Neighbor Interchange ) +SPR ( Subtree Pruning Regrafting ) branch swapping and 100 bootstrap replicates [36] . The ML analysis evidenced pervasive variations of mitochondrial DNA substitution rates among malaria parasite lineages . In this context , we used a Bayesian relaxed molecular clock approach to estimate the divergence times of Plasmodium species . The log-normal rate-autocorrelated model [37] was adopted to relax the molecular clock hypothesis as it has been shown to reasonably fit various data sets [38] . We assumed a calibration interval of 4–7 Myrs for the split between P . falciparum and P . reichenowi [12] , [19] to reflect the one among their human and chimpanzee hosts [20]–[22] . Dating estimates were computed by the Bayesian procedure implemented in the PhyloBayes software [38] , [39] , version 3 . 0 ( http://www . atgc-montpellier . fr/phylobayes/ ) , with a uniform prior on root age and divergence times . We used the CAT Dirichlet process with the number of components , weights and profiles all inferred from the ML topology , with a general time reversible ( GTR ) matrix of exchangeability among nucleotides , and a 4-category discrete Gamma ( Γ ) distribution of substitution rates across sites . Two independent MCMC runs were conducted for 1 , 000 , 000 generations , with sampling every 10 cycles . After a burn-in of 100 cycles , divergence times were computed , and were virtually identical for the two chains .
|
In 2002 , the publication of the genome of Plasmodium falciparum , the most malignant agent of malaria , generated hopes in the fight against this deadly disease by the opportunities it offered to discover new drug targets . Since then results have not lived up to the expectations . The development of comparative genomics to further understanding of P . falciparum has indeed been hindered by a lack of knowledge of closely related species' genomes . Only one species , P . reichenowi , infecting chimpanzees , was hitherto known as a sister lineage of P . falciparum . Here we describe a new Plasmodium species infecting chimpanzees in Africa . Based on its whole mitochondrial genome , we demonstrate that this species is a relative of P . falciparum and P . reichenowi . The analysis of its genome should thus offer the opportunity to explore P . falciparum specific adaptations to humans . Our results bring new elements to the debate surrounding the origin of this lineage . They suggest that it may have been present in early hominoids and may have experienced a radiation congruent with that of its hosts . Our discovery highlights the paucity of our knowledge on the richness of Plasmodium species infecting primates and calls for more research in this direction .
|
[
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
] |
[
"infectious",
"diseases",
"evolutionary",
"biology/evolutionary",
"and",
"comparative",
"genetics"
] |
2009
|
A New Malaria Agent in African Hominids
|
Myosins are ATP-driven linear molecular motors that work as cellular force generators , transporters , and force sensors . These functions are driven by large-scale nucleotide-dependent conformational changes , termed “strokes”; the “power stroke” is the force-generating swinging of the myosin light chain–binding “neck” domain relative to the motor domain “head” while bound to actin; the “recovery stroke” is the necessary initial motion that primes , or “cocks , ” myosin while detached from actin . Myosin Va is a processive dimer that steps unidirectionally along actin following a “hand over hand” mechanism in which the trailing head detaches and steps forward ∼72 nm . Despite large rotational Brownian motion of the detached head about a free joint adjoining the two necks , unidirectional stepping is achieved , in part by the power stroke of the attached head that moves the joint forward . However , the power stroke alone cannot fully account for preferential forward site binding since the orientation and angle stability of the detached head , which is determined by the properties of the recovery stroke , dictate actin binding site accessibility . Here , we directly observe the recovery stroke dynamics and fluctuations of myosin Va using a novel , transient caged ATP-controlling system that maintains constant ATP levels through stepwise UV-pulse sequences of varying intensity . We immobilized the neck of monomeric myosin Va on a surface and observed real time motions of bead ( s ) attached site-specifically to the head . ATP induces a transient swing of the neck to the post-recovery stroke conformation , where it remains for ∼40 s , until ATP hydrolysis products are released . Angle distributions indicate that the post-recovery stroke conformation is stabilized by ≥5 kBT of energy . The high kinetic and energetic stability of the post-recovery stroke conformation favors preferential binding of the detached head to a forward site 72 nm away . Thus , the recovery stroke contributes to unidirectional stepping of myosin Va .
Myosin is an ATP-driven linear molecular motor that produces force and unidirectional movement along actin filaments . The “swinging lever arm” hypothesis proposes that small nucleotide-dependent movements at the catalytic ATPase active site are amplified by rotation of the myosin “lever arm , ” or “neck , ” light chain–binding domain that extends from the motor domain , or “head” [1] , [2] . In the myosin chemomechanical cycle , the lever arm swing that propels the myosin motor forward along actin is referred to as the “power stroke” and is accepted as a general mechanism for myosin contractility . The “recovery stroke” is the essential motion that primes , or “cocks , ” the lever arm in the pre-power stroke position while myosin is detached from actin . These strokes are the basis for the physiological functions of all characterized myosin motors . Myosin Va is a cargo transporter in cells [3] that has two heads , each connected to a long and relatively stiff neck [4] reinforced with six calmodulins ( Figure 1A ) . Myosin Va moves processively along actin filament and takes unidirectional “steps” [5] in which it alternately places its two heads in forward positions ∼72 nm away from a previous binding site [6] , analogous to human bipedal walking . A mechanism for unidirectional stepping has been investigated and proposed as follows ( Figure 1A ) . When a head is detached off actin , the detached neck undergoes rotational Brownian fluctuations around a free joint at the neck–neck junction [7] , [8] . Although the fluctuations are random [7] , the power stroke of the bound head [4] , [9] tilts the neck via “lever action” and moves the junction ( i . e . , the pivot point for the fluctuations ) forward , thereby favoring binding of the detached head to a forward site . This mechanism explains how a detached head can access a forward site , but not why it binds preferentially to a forward site 72 nm away as opposed to other accessible sites as , for example , a site adjacent to the bound head . For a detached head to bind actin , the actin-binding site of myosin must be properly oriented with respect to the actin filament . Therefore , since the position of the neck–neck junction relative to the actin filament is constrained by the bound neck , the orientation ( angle ) and stability of the detached head relative to its neck ( head–neck angle ) dictate the binding site along a filament . The detached head orientation is determined by the recovery stroke that occurs after ATP-induced detachment from actin . If the role of the recovery stroke were just to prime myosin , the head–neck angle could fluctuate significantly . This could allow for the unbound head to bind to a site near or adjacent to the bound head as well as to a site 72 nm away with similar frequency . Such a distribution of the step size , however , has never been observed in the absence of applied external load [5] , [6] , [10] . Therefore , another mechanism must exist . We anticipated that the recovery stroke plays a critical role in orientating the unbound head so that binding to a ∼72-nm forward site occurs preferentially [11] , [12] . In addition , it has been reported that myosin Va moves forward under ∼2 pN of backward load [5] , [10] which would bring the junction back beyond the neutral position [13] or reverse the power stroke [14] , and cancels the bias introduced by the attached head power stroke . The additional role of the recovery stroke above can be another bias for forward stepping even in the presence of the load . Thus , the properties of the recovery stroke are critical for the myosin Va stepping mechanism . Several recent structural and kinetic studies have demonstrated the existence and implications of the myosin recovery stroke . High-resolution crystal structures of muscle myosin II [15] identified different nucleotide-dependent head–neck angles in the absence of actin; these are thought to correspond to pre- and post-recovery stroke angles . Bulk Förster resonance energy transfer assays of myosin II revealed two [16] or three [17] nucleotide-dependent ( averaged ) transient angle distributions . In addition , electron microscopic analysis of myosin Va [18] showed two different orientations ( i . e . , projection angles ) of heads relative to the neck , depending on the nucleotide in solution . These observations have contributed to a general model in which ATP binding triggers the recovery stroke , and phosphate ( Pi ) release after hydrolysis leads to relaxation of the recovery stroke ( i . e . , generation of the power stroke ) . However , the energetic and kinetic angle stability of the pre- and post-recovery stroke conformations of myosin ( Figure 1A ) and the manner in which they contribute to actin binding specificity during processive stepping of myosin Va remains unknown . We present in this study , to the best of our knowledge , the first direct observations of the myosin recovery stroke ( angle change at head–neck junction ) in real time and at the single molecule level . We developed a novel light-induced ATP-concentration controlling system and single motor molecule assay that enables the direct observation of the nucleotide-dependent dynamics and fluctuations of the myosin motor domain . Our observations and analysis indicate that the myosin Va motor conformation adopted after the recovery stroke is kinetically and energetically stable , which allows for the detached head to bind preferentially to a forward site 72 nm away , thereby providing the grounds for biased forward stepping of myosin Va along actin filaments .
We constructed an optical microscope observation system ( Figure 1B ) to directly visualize in real time the nucleotide-dependent swings ( i . e . , strokes ) and fluctuations of the myosin head–neck angle using an engineered monomeric ( single-headed , “S1-like” ) myosin Va ( Figure S1 ) . We anticipated that a monomeric myosin Va molecule with a 50-nm bead ( gray ) attached at its neck ( configuration depicted in Figure 1B ) would permit transient swinging of a 0 . 29-µm bead duplex ( cyan ) attached to the distal head region . To determine how the head orientation , assayed from the bead position of the 0 . 29-µm bead duplex , responds to ATP , we included 200 µM caged ATP and 1 . 7 mU µl−1 apyrase in the solution , such that ultraviolet ( UV ) irradiation generated an ATP transient that was rapidly removed ( hydrolyzed to AMP ) by the apyrase with a time constant of 2–3 s ( Figures S2 and 2B ) . We imaged a duplex ( or a larger aggregate ) of beads , and initiated a full-intensity ( ∼2 nW µm−2 , defined as 100% ) UV pulse for 0 . 1 s that yielded a peak ATP concentration ( [ATP]peak ) of ∼2 µM ( Figure S2 ) . Approximately 0 . 1% of duplexes made a distinct angular ( >30° judged in real time ) swing within several seconds of the UV flash . Such a low frequency is not unexpected given the low probability of an unobstructed configuration , as illustrated in Figure 1B ( drawn to scale in Figure S4 ) . Myosin Va predominantly adopts two distinct conformations during an experiment: a resting angle in the absence of ATP ( i . e . , before UV irradiation; cyan in Figures 1C , 3A , and S5 ) and a metastable angle ( yellow ) accessible only after ATP generation ( save rare excursions driven by Brownian fluctuations ) , interpreted as the post-power stroke and pre-power stroke conformations of myosin Va , respectively . A large fraction ( ∼50% ) of the beads that swung returned to the original angle in less than 2 min , and the UV-induced transient swings could be repeated multiple ( >2 ) times ( Figures 1C and 3A; Video S1 ) . We monitored 15 such duplexes ( i . e . , myosin Va molecules ) and analyzed a total of 121 swing–return pairs as detailed below . A subset ( ∼20% ) made two return swings and then detached from the surface or remained immobile . The remaining ∼30% did not return or did not respond to the second UV flash . Excursions to the post-recovery stroke “state” ( see Figure 1C legend ) are ATP ( UV flash ) –dependent . Every UV irradiation lasting 0 . 1 s at 100% intensity ( [ATP]peak ∼ 2 µM ) induced a bead swing within a few seconds ( 0 . 78 s on average; 29 flashes in six molecules; e . g . , Figure 1C ) . Shorter and/or weaker irradiations yielded longer delays before a swing ( Figure S5A and S5B ) or no bead swings . UV irradiation while the bead duplex was in the post-recovery stroke state , in contrast , never induced a swing: 37 flashes of 0 . 1- or 0 . 2-s duration at 100% intensity failed to induce bead rotation in six molecules ( Figure S5C ) . Thus , swings from the pre-recovery stroke state are initiated and limited by ATP binding , and myosin Va in the pre-recovery stroke state prior to a swinging event is free of bound nucleotide . To quantitate the ATP dependence of swings , we developed a new technique that generates a nearly constant level of ATP in a chamber using caged ATP , which was first applied to a biological system by Trentham and colleagues [19] , and evaluated the method using the rotary F1-ATPase ( GT mutant ) motor [20]–[22] ( Figure S3A ) . Stepwise UV pulse sequences with pulse width modulation ( Figure 2A ) of varying intensity ( 14% , 4 . 4% , and 0 . 7% ) repeatedly generated intensity-dependent rotations of a given F1-ATPase molecule ( Figure 2B ) . Averaged traces of rotations are smooth , indicating that the UV pulse sequences generate nearly constant ATP levels in the sub-second time scale ( Figure S3B ) . Rotational rates in the presence of known [ATP] ( Figure S3C ) yielded UV intensity–dependent ATP concentrations ( Figure 2C ) . These calibrations for constant ATP level allow us to analyze the kinetics of ATP-induced myosin Va swinging . In the myosin Va swing assay , we turned on the sequence at different UV intensities ( i . e . , [ATP] ) until a swing occurred ( orange bars in Figures 3A and S5D ) . The time before a swing was inversely proportional to the [ATP] ( Figure 3B ) , yielding an apparent ATP binding rate constant of 2 . 5×106 M−1s−1 , comparable to the value of 1 . 7×106 M−1s−1 measured in solution ( Protocol S1 ) . These qualitative measurements strongly suggest observed swinging events are those of functional myosin motors . Except for occasional , short reversals in the post-recovery stroke state ( e . g . , arrow heads in Figures 1C and S6; discussed below ) , the post-recovery stroke state is characterized by exponentially distributed dwell times with an average of 40±4 s ( standard error ) ( Figure 4 ) . Note that the post-recovery stroke state is quite stable kinetically , particularly in comparison to the stepping intervals of 60–80 ms at physiological [ATP] , during which Pi release is accelerated by binding to actin [10] , [23] . Our bulk , biochemical assays indicate that ATP is rapidly ( >100 s−1; [23] ) hydrolyzed into ADP and inorganic Pi is released with a rate constant of ∼0 . 02 s−1 ( τ = ∼50 s ) ( Figure S7 ) . Measurements with a shorter-neck , 1IQ construct [23] , [24] indicated a Pi release rate constant of ∼0 . 02 s−1 and a subsequent ADP release rate of ∼1 . 2 s−1 . Collectively , these measurements indicate that myosin Va in the post-recovery stroke conformation has ADP and Pi bound in its active site and that Pi release limits the return swing ( i . e . , power stroke off actin ) , consistent with bulk Förster resonance energy transfer assays with myosin II in solution [16] and electron microscopy of myosin Va bound to actin [25] . The angular fluctuations in both the pre-recovery stroke and post-recovery stroke states are well fitted to Gaussian distributions ( Figures 5A and S8 ) , with a peak separation yielding an average swing amplitude ( θswing ) of 85°±19° ( standard deviation [s . d . ] for 15 molecules ) . The measured amplitudes reflect projections in the image plane , and thus the actual amplitudes will differ if out-of-plane swinging occurred . However , the observation that the appearance of most ( ∼2/3 ) of the bead amplitudes is independent of the swing angle ( e . g . , Figure 5B ) , as confirmed by the constancy of the axial ratio ( Figure 5C ) , indicates that the recorded swings used in the analysis were in a near horizontal plane . Electron micrographs of myosin Va without actin show comparable ( ∼90° ) nucleotide-dependent angular changes [18] , thereby strengthening the interpretation that transitions between pre-recovery stroke and post-recovery stroke conformations of myosin Va are being observed . Both the pre-recovery and post-recovery stroke conformations display considerable conformational flexibility , as indicated by the standard deviation of the angular fluctuations ( Figures 5A and S8 ) . The magnitudes of fluctuations in both states are comparable , with the Gaussian width ( s . d . , σ ) averaging 24°±10° ( s . d . for 15 molecules; σpre/θswing = 0 . 29±0 . 09 ) for the pre-recovery stroke conformation and 26°±9° ( σpost/θswing = 0 . 31±0 . 10 ) for the post-recovery stroke conformation . These observed fluctuations include contributions from flexibility in the myosin–bead junctions as well as experimental image noise , so they represent an upper limit , with actual angle fluctuations being smaller . The Gaussian width of the thermally driven fluctuations ( σ ) measured here , with the equipartition principle [26] , ( 1 ) where kBT ( 4 . 1 pN•nm ) is thermal energy and k is the myosin Va head–neck joint stiffness ( spring constant ) , allows us to determine the spring constants , kpre = 23 pN•nm•rad−2 , and kpost = 20 pN•nm•rad−2 . With this spring constant , the energy required for bending of the head in the post-recovery stroke conformation to the pre-recovery stroke conformation ( i . e . , the energy needed to bend the spring by θswing ) expressed in terms of the elastic potential energy ( E ) , ( 2 ) is 5 . 2 kBT . The post-recovery stroke conformation is stabilized at least to this extent: because the experimental σ in equation 1 includes the fluctuations of other components described above , the spring constant k for the head–neck junction must be underestimated , and thus the energy difference , E , of 5 . 2 kBT between pre- and post-recovery stroke conformations is a lower limit . There were occasions where we observed momentary swings back to the pre-recovery stroke angle in the post-recovery stroke state ( e . g . , arrow heads in Figures 1C and S6 ) . These are unlikely to be purely Brownian excursions , because the bead tended to remain at the pre-recovery stroke angle for a second or longer . A natural return followed by immediate ATP binding that would induce a second swing is also unlikely , because [ATP] must be negligibly low and these momentary swings happened irrespective of the time after UV irradiation . The observed momentary returns may represent reversal of the reaction responsible for the swing to post-recovery stroke conformation , ATP hydrolysis [23] , or subsequent myosin isomerization [16] . We note that the return frequency in the absence of drag from the attached beads could possibly be higher .
The natural assumption is that the detached head accessing a forward site in the post-recovery stroke conformation will have its actin binding site properly oriented for productive binding to actin ( Figure 1A ) . Conversely , when the detached head goes back to the post-recovery stroke conformation , the actin binding site is predicted to be oriented incorrectly , thereby precluding actin binding ( Figure 6A ) . The kinetic stability of the post-recovery stroke state observed here indicates that this proper head orientation is maintained for ∼40 s , much longer than the stepping intervals . Even if the head in the post-recovery stroke state accidentally touches a backward site at a moment when the head adopts a near pre-recovery angle by fluctuation or momentary reversal , the binding should be unstable by at least by 5 kBT compared to forward binding . Thus , the kinetic and energetic stabilities of the post-recovery stroke state together ensure forward binding of an unbound head . Momentary binding of a head with incorrect orientation will be unstable from intramolecular strain [11] , [12] . Consistently , a quantitative model has shown that the lever arm ( neck ) elasticity and its strain influence the position of the next binding site on actin , therefore the detached head preferentially binds to the forward site [27] . This model assumes that the unbound neck with bound ADP–Pi rigidly takes post-recovery stroke conformation , which we report here . The key for directional movement is to bias the completely random Brownian rotations of a detached neck toward forward binding . The power stroke and its angle stability of the attached rear head contribute approximately half of the bias by moving the pivot for the Brownian rotation of the unbound neck forward , which allows the detached head to access positions 36- to 72–nm distant on an actin filament [4] , [7] ( Figure 1A ) . The remaining bias between positions ∼36 nm and ∼72 nm from the detached site is provided by the recovery stroke and its stability . Under a high backward external load , the power stroke would fail to produce a bias: owing to the compliance in the neck and/or neck–head junction [13] , [14] , the neck–neck junction would be pulled back to the neutral position , immediately above the bound head ( Figure 6B ) . Even under this circumstance , the bias by recovery stroke still works , favoring forward binding . Therefore , for ∼72-nm discrete unidirectional steps of myosin Va , the recovery stroke and its angle stability of the detached head contributes to the bias , in addition to the power stroke and its angle stability of the attached head . This mechanism may contribute to transport cargos in a cell since some cellular components could be obstacles to hinder the movement of cargo at times . An alternative mechanism has recently been proposed for myosin VI [28] , which is thought to function as a force sensor as well as a transporter [29]: stable lead head binding is facilitated by a backward load on the head , and hence internal strain between the two necks promotes forward binding of an unbound head . Myosin VI is the only reverse motor known to date , moving in the direction opposite to all other myosins studied so far . It is of interest to study whether the other myosins , including myosin Va , also adopt a similar , strain-dependent binding for forward bias . The stability of the post-recovery stroke conformation would also be important for muscle myosin II , which can produce tension without contraction ( isometric tension ) by repeatedly “scratching” actin . Forward binding is required for efficient force production , but the base of the necks does not move in this situation , and thus myosin II may rely entirely on the head orientation being stabilized in the post-recovery stroke state . Other linear motors may also rely on an effective swing to the post-recovery stroke conformation [11] , [12] . To study ligand-dependent motion of molecules , caged nucleotides ( uncaged by UV irradiation ) have been combined with microscopic observations . UV pulse irradiation allows one to trigger motion of the molecular motor and to clearly show its nucleotide dependence [30] , [31] , and modulation of UV irradiation time allows one to control motor velocity and total movement [32] . This assay design has the advantage over conventional flow/mixing assays in that solution conditions ( e . g . , nucleotide concentration ) can be altered rapidly and with minimal perturbation . However , caged nucleotide measurements have been limited to kinetic analysis because the concentration of uncaged compound can change significantly during the course of an experiment , particularly if consumed by the system being examined ( i . e . , diffusion , enzyme–substrate interaction , or apyrase ) . We have developed a new technique to keep ATP level constant in which the concentration and time evolution can be modulated by light intensity and irradiation time ( Figures 2 , 3A , S3 , and S5D ) . This method for visualization of a nucleotide-linked conformational change in a motor protein under the controlled delivery of ATP should be generally applicable to ligand-induced conformational changes of macromolecules .
Monomeric Gallus gallus myosin Va truncated at Leu-909 ( containing all six IQ motifs ) with an N-terminal myc tag ( EQKLISEEDL ) positioned directly replacing Met-1 and a C-terminal FLAG tag ( DYKDDDDK ) with a single glycine linker ( Figure S1 ) was co-expressed with Lc-1sa in Sf9 cells and purified by FLAG affinity chromatography in the presence of excess calmodulin as previously described [24] , [33] . The calmodulins on the expressed protein were exchanged for 6× his-tagged calmodulin , expressed in Escherichia coli , as previously reported [34] and modified [7]: the his-tagged calmodulin and monomeric myosin Va at the molar ratio of 6∶1 were mixed and incubated for 10 min on ice in 20 mM imidazole-HCl ( pH 7 . 6 ) , 4 mM MgCl2 , 100 mM KCl , 0 . 04 mM EGTA , 0 . 5% ( v/v ) β-mercaptoethanol , and 400 µM CaCl2 . The reaction was terminated by the addition of 4 mM EGTA followed by >20 min incubation on ice . Monomeric myosin Va carrying his-tagged calmodulin was mixed with an anti-his monoclonal antibody ( Clontech Laboratories ) at the antibody:myosin molar ratio of 17∶1 in buffer A ( 25 mM imidazole-HCl [pH 7 . 6] , 4 mM MgCl2 , 100 mM KCl , 1 mM EGTA , 5 mM DTT ) , and incubated at room temperature for >5 min to allow binding . A flow chamber , in all experiments under a microscope , was made of two coverslips separated by two spacers of ∼100-µm thickness , and , after the last infusion , the chamber was sealed with silicone grease or nail liquid . The following infusions ( 2–3 chamber volumes ) , all in buffer A , were made with 1–2 min of incubation in between: 2 mg ml−1 unphosphorylated α-casein for surface blocking , buffer A for washing , 5 . 6% ( w/v ) 0 . 05-µm silica beads ( Polysciences ) , buffer A for washing , monomeric myosin Va ( 10 nM ) complexed with anti-his antibody ( for binding to the silica beads through the antibody ) or myosin Va alone without the calmodulin exchange ( for direct binding ) , 2 mg ml−1 unphosphorylated α-casein , 25 µg ml−1 biotinylated anti-myc monoclonal antibody ( Millipore ) , and buffer A for washing . Finally , 0 . 29-µm streptavidin-coated beads ( Seradyn ) , washed three times by centrifugation in buffer A , were infused together with 200 µM caged ATP ( Dojindo ) , 1 . 7 mU µl−1 apyrase ( Sigma ) , 1 . 1 mg ml−1 unphosphorylated α-casein , and 0 . 5% ( v/v ) β-mercaptoethanol . The purpose of the anti-his antibody was to let it serve as a cushion between the myosin neck and a silica bead so as to keep the myosin intact . Direct binding , though , worked as well , and some results , e . g . , in Figures 1C and S5A–S5C , were obtained with direct binding . In both cases , most of the 0 . 29-µm beads on the surface were bound to the head of myosin Va through a biotin–avidin linkage , because the bead density decreased significantly without myosin , with non-biotinylated anti-myc antibody instead of the biotinylated one , or by mixing excess biotin with the streptavidin-coated beads before infusion . When we infused short actin filaments instead of the 0 . 29-µm beads , they attached ( presumably ) to myosin Va , and a flash of 100% UV light for 0 . 2 s released >97% of them from the surface within a few seconds . We used an Olympus IX70 microscope equipped with a 100× objective ( UPLSAPO100× O IR , N . A . 1 . 4 , Olympus ) , a stable sample stage ( KS-O , ChuukoushaSeisakujo ) , a dual-view system [35] for simultaneous observation of fluorescence and bright-field images [36] , a regular epi-fluorescence port , and an additional UV excitation port consisting of a mercury lamp , an extension tube ( IX2N-FL-1 ) that forms an intermediate ( conjugate ) image plane outside the microscope body , and a computer-controlled shutter with 5-ms open–close time ( Uniblitz ) . Fluorescence of Alexa 488 was excited at 475–490 nm , and images at 500–535 nm were captured with an intensified ( VS4-1845 , Video Scope ) CCD camera ( CCD-300-RCX , Dage-MTI ) . Bright-field images ( 650–730 nm ) were recorded with another CCD camera . UV excitation ( 300–400 nm ) for uncaging ATP was confined in a circle of diameter ∼90 µm at the image plane . A mask was placed on the conjugate plane in the extension tube such that the central ∼30-µm square in the image plane did not receive UV light . The swing assay was always made near the center of the masked area to protect myosin from possible UV damage , although we found that direct UV irradiation at the maximum intensity ( see below ) for tens of seconds did not affect the motile activity of myosin Va . The rotation speed of F1-ATPase ( for estimation of ATP concentration; Figures 2 and S3 ) did not depend on the position in , and even outside , the masked area , and short actin filaments bound to myosin Va were released by a UV flash with indistinguishable kinetics at all positions . Note that oblique UV beams illuminated the solution above the masked area except for the immediate vicinity of the coverslip surface . To record correlation of events and UV irradiation , a part of the UV beam was recorded with the intensified CCD camera above , or with the camera for bright field at an edge of the image . The UV power was measured above the objective lens , and the estimated intensity in the image plane was ∼2 nW µm−2 for unattenuated ( maximal ) excitation ( defined as 100% intensity ) . Observations were made at 23 °C . The orientation of a bead duplex was determined as previously reported [7] . When another bead came nearby , the orientation was judged by eye or abandoned . Ellipticity of a bead image was estimated as the ratio of the long axis length to the short one , calculated from the second moments of the intensity distribution as <Ix2>1/2/<Iy2>1/2 where x and y are pixel coordinates measured along the long and short axes and with the origin at the image centroid , I is the pixel intensity minus a threshold value , and <> denotes averaging . UV-generated ATP concentrations were estimated by both gliding bead assay for native myosin Va and rotational assay for F1-ATPase ( Protocol S1 ) . ATP binding rate and Pi release rate of myosin Va were measured using stopped flow apparatus ( Protocol S1 ) .
|
Myosin Va is a “two-legged” ATP-dependent linear molecular motor that transports cellular organelles by “stepping” along actin filaments in a processive manner analogous to human walking , the two “feet” alternating between forward and backward positions . During stepping , the lifted leg undergoes rotational Brownian movements around a free joint at the leg–leg junction . Although these movements are random , the lifted foot lands preferentially on forward sites and rarely steps backward . This directional bias arises in part from the forward movement of the junction bending the “ankle” of the attached leg . Here , we show that the lifted foot also plays a role in the direction of stepping by controlling the orientation of its actin-binding site ( the “sole” ) , which dictates the accessibility of potential stepping positions . We observed the ATP-dependent foot orientation and its stabilizing on individual myosin Va molecules in real time under an optical microscope; we show that the lifted foot of walking myosin Va is oriented in a “toe-down” conformation so that binding to a forward site on actin is preferred largely over backward or adjacent sites . Thus , the great kinetic and energetic stability of the myosin Va lifted foot conformation contributes to unidirectional stepping along actin filaments .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biomacromolecule-ligand",
"interactions",
"biochemistry",
"enzyme",
"structure",
"cell",
"motility",
"actin",
"filaments",
"enzymes",
"biology",
"enzyme",
"kinetics",
"biophysics"
] |
2011
|
Direct Observation of the Myosin Va Recovery Stroke That Contributes
to Unidirectional Stepping along Actin
|
Apicomplexa tick-borne hemoparasites , including Babesia bovis , Babesia microti , and Theileria equi are responsible for bovine and human babesiosis and equine theileriosis , respectively . These parasites of vast medical , epidemiological , and economic impact have complex life cycles in their vertebrate and tick hosts . Large gaps in knowledge concerning the mechanisms used by these parasites for gene regulation remain . Regulatory genes coding for DNA binding proteins such as members of the Api-AP2 , HMG , and Myb families are known to play crucial roles as transcription factors . Although the repertoire of Api-AP2 has been defined and a HMG gene was previously identified in the B . bovis genome , these regulatory genes have not been described in detail in B . microti and T . equi . In this study , comparative bioinformatics was used to: ( i ) identify and map genes encoding for these transcription factors among three parasites’ genomes; ( ii ) identify a previously unreported HMG gene in B . microti; ( iii ) define a repertoire of eight conserved Myb genes; and ( iv ) identify AP2 correlates among B . bovis and the better-studied Plasmodium parasites . Searching the available transcriptome of B . bovis defined patterns of transcription of these three gene families in B . bovis erythrocyte stage parasites . Sequence comparisons show conservation of functional domains and general architecture in the AP2 , Myb , and HMG proteins , which may be significant for the regulation of common critical parasite life cycle transitions in B . bovis , B . microti , and T . equi . A detailed understanding of the role of gene families encoding DNA binding proteins will provide new tools for unraveling regulatory mechanisms involved in B . bovis , B . microti , and T . equi life cycles and environmental adaptive responses and potentially contributes to the development of novel convergent strategies for improved control of babesiosis and equine piroplasmosis .
The presence of AP2 genes in apicomplexans was initially described by Balaji et al . [22] , who first reported the identification of members of the AP2 gene family in the genomes of Plasmodium , Theileria , Cryptosporidium , and Toxoplasma . Initial genome characterization in the B . bovis T2Bo strain genome resulted in the annotation of 18 genes encoding for AP2 domain-containing proteins [3] . However , Oberstaller et al . [8] , using a highly sensitive Hidden Marcov Model ( HMM ) , recently identified four additional genes encoding for AP2 proteins , thus extending the number of genes encoding for AP2 domain-containing proteins to a total of 22 . General features of the 22 B . bovis genes and their predicted proteins are shown in Table 1 . Because AP2 proteins may have more than a single AP2 domain , the B . bovis AP2 proteins display a total of 26 known AP2 domains . Similar to what was found in other apicomplexan genomes , the AP2 genes are not organized in clusters but dispersed throughout the four chromosomes of B . bovis ( Fig 1A and 1B and Table 1 ) . Bioinformatics analysis performed on the predicted amino acid sequences of the B . bovis AP2 proteins shows that some contain other additional known functional domains ( Table 1 , Fig 1B ) , such as the ACDC domain ( AP2 coincident domain present mostly at the C-terminus of the proteins ) , a conserved PBP1domain ( PAB1-binding protein 1 ) , which is also present in proteins interacting with a poly ( A ) -binding protein , and in the Topoisomerase II-associated protein ( PAT1 ) , a protein that facilitates accurate chromosome separation during cell division ( Table 1 ) . Consistently , and together with the AP2 domain , all these additional domains are known to function in a nuclear environment . Predicted intracellular localization and routing of B . bovis AP2 proteins into the cell nucleus is consistent with the lack of signal peptides in all the putative B . bovis AP2 proteins as determined by sequence analysis using the SMART programs ( http://smart . embl-heidelberg . de/smart/set_mode . cgi ? NORMAL=1 ) . In addition , cellular localization predictions using the program Cello v2 . 5 ( http://cello . life . nctu . edu . tw/ ) predicted an intranuclear subcellular localization for all B . bovis AP2 proteins . The predicted molecular size and isoelectric points of the B . bovis AP2s are also highly diverse , ranging from ~21 to 103 kDa to 5 . 15 to 11 . 21 kDa ( Table 1 ) . In general , there appears to be an association between isoelectric points ( pI ) and size of the molecules , and , thus , molecules with higher pI are of a relatively smaller size than the ones with a lower pI ( Table 1 ) . This association is consistent with a previous study by Kiraga et al . [23] , although its biological relevance remains unknown . While 19 out of the 22 known B . bovis AP2 proteins contain a single AP2 domain , the genes BBOV_II007120 and BBOV_III004740 contain two AP2 domains , and gene BBOV_I004850 has three AP2 domains ( Table 1 , Fig 1B ) . Similar to AP2 proteins in plants , two of the three domains in the putative protein encoded by BBOV_I004850 are separated by 25 amino acids in the amino terminal part of the molecule , whereas the third domain is distally localized , separated by 160 amino acids from the second domain and 30 amino acids apart from the C-terminal end of the molecule . The AP2 protein encoded by gene BBOV_II007120 contains the two AP2 domains separated by 21 amino acids , whereas the two AP2 domains of the protein encoded by gene BBOV_III004740 are just 17 amino acids apart . It is possible that proteins containing multiple AP2 domains are able to bind to distinct DNA regions either separately or simultaneously , thus adding increasing functional versatility for these molecules . In general , and consistent with what was found for other AP2 proteins , there is low sequence identity or similarity among the AP2 proteins , and , thus , their similarities are just restricted to the conserved 60 amino acid domain [8 , 22 , 24] . The percent identities found among the full AP2 proteins after their alignment is shown in S1 Table . The alignment and the identity results were obtained by using Clustal omega ( http://www . ebi . ac . uk/Tools/msa/clustalo/ ) . The more significantly related AP2 proteins are BBOV_I004850 and BBOV_II005480 , sharing 25 . 59% identity ( S1 Table ) , followed by BBOV_I000100 and BBOV_III003770 , with 23 . 68% identity ( S1 Table ) . Overall , these data suggest that , with few exceptions , the B . bovis AP2 proteins are not highly related in sequence outside the AP2 domains . Alignments of the AP2 domain among all B . bovis revealed 100% identity between the AP2 domains in BBOV_III003770 and BBOV_I003560 , suggesting the possibility of shared DNA binding specificities . Interestingly , the highly related domains BBOV_I004850 . 3 and BBOV_I004850 . 2 ( sharing 51 . 02% identity ) are both localized in the same protein ( gene BBOV_I004850 ) . Alignments among all B . bovis AP2 domains ( Fig 2 ) show that certain amino acid residues have a high degree of sequence conservation and may be functionally required in the B . bovis AP2 proteins . For instance , and similar to what was found for other apicomplexan AP2 proteins , all B . bovis AP2 domains contain highly conserved W and F residues ( labeled with asterisks in Fig 2 ) . It is known that these positional conserved residues are likely to help stabilize hydrophobic interactions between the AP2 domain and its recognized DNA target [22] . Consistently , and as described in more detail below , these residues are also conserved in the AP2 domains identified in the B . bovis related intra-erythrocytic apicomplexan B . microti and T . equi ( S1 and S2 Figs ) . In addition , other amino acids are also highly conserved ( Fig 2 ) among the B . bovis AP2 domains . Just 20 AP2 genes were annotated as containing AP2 domains in the published T . equi genome [6] . However , using further bioinformatics analysis , we found that genes BEWA_041620 and BEWA_018840 also contain AP2 domains . Thus , we propose that T . equi contains at least 22 AP2 genes . The organization and orientation of such genes into the four nuclear T . equi chromosomes are depicted in Fig 3A and S2 Table . Similar to what was observed for B . bovis , the T . equi AP2 genes are scattered among all four chromosomes ( Fig 3A ) . As it was found for B . bovis , the AP2 genes of T . equi may also contain 1 , 2 or 3 AP2 domains . Similar searches performed on the published B . microti genome [5] resulted in the identification of 21 AP2 genes ( S3 Table ) . All the Ap2 domain-containing genes present in the B . microti genome were previously annotated as such , except gene BBM_III08920 coding for a protein with a single previously unnoticed Ap2 domain , which is reported here for the first time . Fig 3B describes the organization as well as the orientation of the 21 AP2 genes into the four chromosomes of B . microti . Similar to what was found for B . bovis , the T . equi and B . microti AP-2 proteins contain other conserved domains , such as the ACDC and the PBP1 domains ( S2 and S3 Tables ) . Sequence comparisons among all the Ap2 domains identified in the B . bovis , T . equi , and B . microti putative AP2 proteins ( Table 2 ) revealed high levels of identity among some domains . The identity reaches 100% among domains from proteins BBOV_III008870 ( B . bovis ) , and BEWA_010510 ( T . equi ) , and BBM_III06770 ( B . microti ) . Interestingly , the proteins encoded by the B . bovis gene BBOV_I004850 and the T . equi BEWA_011980 gene have three highly similar domains . They share 100% identity for their first domain , which is also highly conserved in the B . microti protein encoded by gene BBM_III05870 . 1 ( 95 . 74% identity ) . Additional domain similarities are described in Table 2 . The functions and DNA-binding specificities of the B . bovis , B . microti , and T . equi AP2 domains remain unknown , and they will need to be defined experimentally . Remarkably , the specificity of binding of some AP2 proteins to certain short DNA target motifs ( usually six to seven base-pairs long ) appears to be quite conserved among distinct Plasmodium species and , furthermore , among other related apicomplexans [7 , 25] . These findings suggest that Plasmodium binding specificity data together with bioinformatics analysis on the 5′ upstream gene coding regions could guide the design of future experiments aimed at establishing the DNA binding specificities of the AP2 proteins in the three parasites examined in this study . Recent research focused on the identification of specific AP2 proteins involved and required for regulating the expression of some stage-specific genes in Plasmodium [9 , 10 , 15 , 16] . The related malaria parasites start differentiating into gametocytes while the parasites are still replicating inside erythrocytes in mammalian hosts . This crucial step requires a developmental decision , resulting in parasites that continue to replicate asexually or to differentiate into non-dividing male or female gametocytes , a life cycle event that is required to assure generation of genetic diversity and further transmission of the parasite upon mosquito acquisition . It was recently demonstrated that this developmental transition in P . falciparum parasites is regulated by the activity of the AP2 protein identified as pfAP2-g ( PFL1085w ) ( Fig 4 Panel A ) . It was thus postulated that pfAP2-G functions as a transcriptional switch , stimulating the commitment to sexual development in this parasite [26] . Recent studies also supported the role of AP2 factors as candidate regulators driving the commitment to merozoite production in T . annulata [27] . Using a combination of techniques including transcriptome analysis and phenotypic characterization of AP2 gene knock outs , Yuda et al . [28] identified the AP2-O transcription factor , which is involved in the formation of invasive kinetes in Plasmodium berghei and P . falciparum ( PB000572 . 01 . 0 and PF11_0442 ) ( Fig 4 , Panel B ) . Orthologues of the AP2-O gene have been also identified in other Plasmodium spp parasites . In addition , the same study also defined the sequence of the DNA involved in the binding to the AP2-O as the six-base motif TAGCTA . In a different study , Yuda et al . [29] also identified AP2-Sp ( PB000752 . 01 . 0 ) ( Fig 4 , Panel C ) , a protein that is required for the regulation of the expression of P . berghei sporozoites and also defined the sequence TGCATG as a cis-acting element that is specific for its binding to DNA . Interestingly , the Ap2 domains involved in the binding of all these functionally defined Plasmodium AP2s are found to be well conserved in B . bovis , B . microti , and T . equi AP2 proteins , as shown in Fig 4 . Therefore , and based on the sequence similarities of the AP2 domains shown in Fig 4 , we hypothesize that the proteins encoded by genes BBOV_II005480 ( ~72% identity ) , BBM_I03085 ( ~76% identity ) , and BEWA_022490 ( ~77% ) are functionally equivalent to the Plasmodium G ( AP2-G ) protein ( PFL1085w ) . This is supported by previous findings demonstrating that the divergent T . annulata AP2-G protein containing AP2 motifs that are orthologous with the P . falciparum AP2-G protein are able to bind identical GxGTACxC motifs [27] . Data in S4 Table illustrates the orthologous relationships of putative AP2-G motifs of Theileria and Babesia parasites . The recently identified AP2-G T . annulata TA13515 gene [27] encodes for an AP2 motif that is 77 . 36% identical to the motif encoded by the functionally defined AP-G PFL1085w gene . However , this motif is more related in identity to the putative AP-G proteins in Theileria parva , Theileria orientalis , T . equi , B . bovis , and B . microti addressed in this study . These findings further support the testing of these AP2 as candidates for modulators in the transition of these parasites into sexual stages . Consistently , we also hypothesize that the genes identified as BBOV_I004280 ( ~70% identity ) , BBM_II03250 ( ~79% identity ) , and BEWA_041620 ( ~74% identity ) are the functional equivalents of the Plasmodium AP2 proteins PF11_0442 and PB000572 . 01 . 0 , which are both involved in Plasmodium ookinete development . Similarly , the AP2 proteins encoded by genes BBOV_II001610 ( ~65% identity ) , BBM_II02455 ( ~68% identity ) , and BEWA_008880 ( ~65% identity ) might also be functional equivalents of AP2- Sp PB00752 . 01 . 0 , which is involved in sporozoite development in malaria ( Fig 4 ) . These domain homology-driven predictions could help in prioritizing and selecting candidates for functional testing of these hypotheses , leading to define B . bovis , B . microti , and T . equi regulation pathways involved in gametocyte , ookinete , and sporozoite development . It is possible that the proteins containing these highly conserved domains share similar DNA binding specificities among these three parasites , but this will have to be confirmed experimentally . Full transcriptome analysis in the life cycle of these organisms is not yet available , and it will be needed in order to perceive the possible role of AP2 proteins influencing life cycle transitions in these parasites . The Myb proteins , which are highly conserved in eukaryotes , belong to the tryptophan cluster family and are also known to regulate gene expression . Similar to AP2 factors , Myb proteins are involved in differentiation and growth control by binding to DNA in a sequence-specific manner through a DNA-binding domain [10 , 30] . Importantly , Myb proteins have been confirmed to be essential for parasite growth , cell cycle regulation , and progression in Plasmodium parasites [18] . Myb families containing eight genes each are present in the B . bovis , T . equi , and B . microti genomes ( Table 3 ) . Interestingly , a full set of eight Myb genes appears to be well conserved in sequence among the three parasites , and the Myb proteins of these three parasites appear to have similar domain architectures ( S3 Fig ) . Their phylogenetic relationships are shown in Fig 5 and their orthologous relationships confirmed by using Bidirectional Best Blast hit analysis [31] . The orthologous Myb proteins BBOV_II001770 , BEWA_009170 , and BBM_I02995 contain an additional DnaJ motif located at their N-terminus region , while the DNA binding domain typical of the Myb proteins is located in their C-terminus ( S3 Fig ) . In general , Myb genes are unlinked and dispersed among these three parasites’ chromosomes . However , this is not the case for the T . equi Myb genes BEWA_008190 and BEWA_008180 , which are contiguous in chromosome 3 of T . equi . Protein sequence comparisons revealed limited sequence identity among the Myb proteins encoded in each of these three parasites . The possible ortholog relationships among all Myb genes identified in these three parasites are illustrated in the phylogenetic tree shown in Fig 5 . The highly conserved gene BBOV_IV003030 encodes for a Myb protein that is 60 . 43% identical to the one encoded by gene BEWA_044120 in T . equi , and 50% identical to the protein encoded by gene BBM_III01265 found in in the B . microti genome . It is thus possible that these three proteins are functional homologues . In conclusion , these relationships indicate that a core of eight Myb genes is conserved among these three parasites , and perhaps this is also the case in other related apicomplexan parasites as well . Consistently , searches performed on the genome of T . annulata , T . parva , and T . orientalis revealed full conservation of the set of eight Myb genes in these classical Theileria parasites ( S5 Table ) . The complement of eight Myb genes from B . bovis , B . microti , and T . equi grouped in the phylogenetic tree together with the three classical Theileria parasites is shown in S4 Fig . It is possible to infer from these data that an ancestor organism existing previous to speciation among Babesia and Theileria also contained an eight Myb gene family . The high mobility group box proteins ( HMG ) is a group of DNA-binding transcription factors required for the maintenance of structural alterations in DNA during transcription . The HMG superfamily is divided into three families of proteins according to their functional motifs , known as HMGA , interacting with the AT hook; HMGN , involved with nucleosomes; and HMGB , containing one or several copies of HMG box DNA binding domain [20] . In contrast to the AP2 and Myb proteins , the HMG proteins have the ability to bind A-T—rich regions of DNA rather than sequence-specific targets , in a process mediated by basic amino-acid residues of the proteins [31] . There appears to be just one HMG gene in B . bovis ( BBOV_IV001910 ) in chromosome 4 . This HMG gene has been previously cloned and characterized in yeast and B . bovis [32 , 33] . The size of the predicted protein , domain and secondary structure predictions , and sequence comparisons indicate that the B . bovis BBOV_IV001910 gene is similar to the Pf HMGB genes [20] and , thus , it can be considered as a member of the HMGB family . The binding specificity of the P . falciparum HMGB proteins to four-way DNA junctions was also previously established [20] . In addition , a single HMG gene copy in T . equi BEWA_012790 was found on chromosome 4 . The B . bovis BBOV_IV001910 and the T . equi BEWA_012790 predicted proteins are 65% identical and contain just 92 amino acids and a single HMG domain , lacking the typical acidic C-terminal tail [20 , 33] . This putative HMG gene is well conserved among apicomplexans [20] and in other cells but was not annotated as such in the B . microti genome . However , BLAST analysis of the B . microti genome with the BBOV_IV001910 sequence demonstrated the occurrence of a gene present in an unannotated region of the genome ( http://protists . ensembl . org/Babesia_microti_strain_ri/Tools/Blast ? db=core ) , encoding for a homologous HMG protein . This novel putative HMG gene is located in the ~2829bp non-coding region between bp 676455 and 676744 of chromosome 1 of B . microti ( Fig 6A upper part ) . Furthermore , synteny among B . microti , B . bovis , and T . equi in genomic regions encoding this gene was identified ( Fig 6A bottom part ) . Similar to B . bovis and T . equi , the non-coding region of B . microti , which contains the HMG domain , was found to be followed by gene BBM_I01880 ( Fig 6A bottom part ) encoding for a protein containing an AAA domain [cd00009] , an ATP binding motif present in ATPases ( Fig 6A bottom part ) . Furthermore , we also found conservation and consistent synteny of the HMG gene in other related apicomplexa ( T . annulata , T . parva , P . falciparum , Plasmodium knowlesi , and Plasmodium vivax ) ( S5 ) . In Fig 6B , the defining amino acids for the HMG domain are shown , as well as a sequence alignment of three putative HMG proteins and the predicted secondary structures of B . microti , B . bovis , and T . equi . Interestingly , the predicted secondary structures for the in silico translated HMG proteins of B . microti , B . bovis , and T . equi shows three identical alpha-helixes comprising all amino acids involved in the HMG domain ( Fig 6B ) , identical to what was described for their Plasmodium HMGB homologues [20] . It is likely that this conserved secondary structure is essential for access of the HMG proteins to its DNA binding target and for effecting protein function . In P . falciparum , the HMG proteins are present in the nucleus and induce DNA bending [20] . However , the binding targets and exact functions of the Babesia and Theileria HMG proteins remain to be defined . Considering these observations , together with the facts that gene BBOV_IV001910 is relatively highly expressed in B . bovis erythrocyte stages , as described below and shown in Fig 7C , and that key residues defining the HMG domain are also fully conserved in the B . microti putative protein ( Fig 6A and 6B ) , we propose that the region in chromosome 1 of B . microti represented in Fig 6A represents a novel HMG gene . If the presence of an HMG gene in B . microti is confirmed experimentally , then annotation in this region of chromosome 1 of B . microti should be revised . Studies in Plasmodium , T . annulata , and Toxoplasma indicated that most AP2 genes are differentially expressed during the life cycle of the parasites [27] . B . bovis parasites have a complex life cycle involving at least two distinct hosts , the mammal bovine and arthropod tick hosts . B . bovis parasites developing in the bovine hosts only invade and reproduce in erythrocytes , and it remains unclear whether they start committing into gametogenesis while residing in the erythrocyte . However , the life stages of the parasite developing in the definitive tick vector appear to be more diverse and complex , including sexual stages and sexual reproduction , in addition to the development of kinete and sporozoite stages . Furthermore , because of their trans-ovarian mode of transmission , Babesia parasites are able to survive in additional stages of the tick host ( adult , egg , larva , and nymph , with each of these tick stages occurring in dramatically distinct physical surroundings ) . We propose that this feature reflects a high degree of plasticity for this parasite , which enables radical adaptive morphological transitions during changing temperatures , surviving the non-adaptive immune system of the tick and other variable environmental factors while replicating in the tick . Based on the known role of AP2 proteins in related apicomplexans , it is possible that these changes are correlated with unique patterns of expression of AP2 proteins , in order to fulfill their role as stage-specific transcriptional regulators . Analysis of the currently available transcriptome of B . bovis in the blood stages supports this notion , as , at least , expression of two of the AP2 genes , such as BBOV_II005480 and BBOV_II004230 , are significantly elevated in blood stage parasites of attenuated and virulent B . bovis T2Bo strains , while some of the AP2 genes are silenced ( Fig 7A ) . Interestingly , and as shown in Fig 4 , sequence comparisons suggest that the AP2 gene BBOV_II005480 , highly transcribed in blood stages of B . bovis , is a possible correlate of the P . falciparum gene AP2-G ( PFL_1085w ) , which was shown to be involved in the transition of P . falciparum blood stage parasites into sexual forms [26 , 34] . It was recently shown that PFAP2-G functions as a master regulator controlling sexual-stage differentiation decision in Plasmodium parasites [26] . It is currently unknown whether the B . bovis AP2 gene BBOV_II005480 is also involved in the regulation of the expression of genes involved in sexual stage transitions and whether such stage transition also occurs in blood-stage parasites of B . bovis . However , the general currently accepted paradigm is that commitment of B . bovis to sexual forms might start with the formation of pre-gametes while the parasites reside in the bovine hosts [35 , 36] , which would be associated with the high level of expression of the AP2 gene BBOV_II005480 gene in the blood stages of the parasite . It is possible that B . bovis blood-stage parasites need to be primed before developing into sexual stage while still developing into the mammalian host , but this remains unknown . Alternatively , it is also possible that the expression of the gene BBOV_II005480 in blood stages is required for functions unrelated to sexual stage development . Other AP2 genes found to be highly expressed in blood stage parasites include BBOV_II004230 , BBOV_III008870 , BBOV_I002320 , and BBOV_III009600 . Interestingly , levels of transcription for the putative gene AP2-O ( BBov_I004280 ) are negligible in the blood stage , whereas the levels of transcript for the putative AP2-Sp gene , although higher than AP2-O ( BBov_II001610 ) , are also significantly lower than AP2-G ( BBOV_II005480 ) . It could be predicted that the levels of expression of both genes are elevated in tick stages of B . bovis , as its differentiation to kinete and sporozoite stages occurs in the tick . Comparative multistage global transcriptome analysis , together with proteomic analysis , remains to be performed in order to fully understand the patterns of expression of the Babesia AP2 genes among its different life stages . Taken together , these studies should provide a framework for deciphering the gene regulation networks operating during the life cycle of B . bovis and may also contribute to the design of novel methods for the control of this parasite . Myb transcript analysis performed on two distinct B . bovis strains ( T2bo attenuated and virulent ) shows that seven out of the eight gene members are transcribed at relatively low levels in B . bovis blood stages ( Fig 7B ) , whereas the Myb gene BBOV_IV011350 appears to be expressed at significantly higher levels , and , thus , members of this family are also differentially expressed by the parasite . In addition , the HMG gene BBOV_IV001910 is also consistently and relatively highly expressed in the two distinct B . bovis strains analyzed ( T2bo attenuated and virulent strains ) ( Fig 7C ) . The relative high levels of expression of the AP2 , Myb , and HMG genes in B . bovis blood stages can be compared in Fig 8 . Transcripts of the AP2 gene BBOV_II005480 were detected at levels that are at least an order of magnitude higher than the Myb gene and at twice the levels of the highest expressed HMG gene BBOV_IV001910 . The functional significance of these observations remains unknown and requires further study . However , microarray data does not show significant differences in the level of expression of the genes analyzed in this study among the attenuated and virulent strain pairs so far analyzed . There is no experimental evidence supporting the hypothesis that differential expression of these regulatory genes has any correlation with the virulence phenotype of Babesia strains .
Described here is the structure of the AP2 genes of B . bovis as well as the general organization of this family in the related T . equi and B . microti parasites . AP2 genes that are differentially expressed during the blood stages were identified and , based on domain sequence similarities , correlated with already functionally characterized Plasmodium AP2 proteins . A previously unknown gene family with an eight-gene core encoding for proteins , including the DNA binding domain that is characteristic for the transcription factors , known as Myb , was found conserved in B . bovis , B . microti , and T . equi . Remarkably , a conserved HMG gene was also described in these three parasites for the first time , although expression of the B . microti HMG gene identified in this study remains to be confirmed experimentally . The Myb and HMG genes of B . bovis might also be differentially expressed in the blood stages of the parasite . The pattern of expression of AP2 , Myb , and HMG genes in multiple B . bovis , T . equi , and B . microti parasite stages should also be compared in order to start unraveling mechanisms involved in the regulation of gene expression in these parasites . Overall , the findings described in this study suggest conservation of regulatory genes in the face of large divergence of genome size , content and organization , and host specificities among these three apicomplexan parasites . Taking advantage of transfection and gene editing techniques , it is now possible to design KO and overexpression studies aimed at defining the resulting phenotype of mutated or genetically altered transfected parasites , leading to a correlation between gene and protein function for the AP2 , HMG , and Myb proteins . In addition , experiments leading to the identification of the binding specificities for each of the B . bovis , B . microti , and T . equi AP2 proteins , as well as the Myb and HMG transcription factors , should also be performed . Finally , the ability to genetically manipulate genes encoding for transcription factors should result in a better understanding of the biology of these parasites and to the rational design of attenuated and non-tick transmissible parasite strains that can be used for the development of the next generation of live attenuated vaccines and chemotherapeutics . Conservation of key gene regulation mechanisms may lead to future development of novel converging control strategies that can be applied to apicomplexan parasites .
|
The tick-borne apicomplexan parasites Babesia and Theileria are responsible for costly and devastating diseases globally . Improved control is needed , but the biology of these parasites remains poorly understood . Significant gaps include better understanding of the mechanisms involved in control of gene expression and the events leading to parasite development among hosts , including the production of sexual stages in their definitive tick vector hosts . Similar to other better-studied eukaryotic cells , it is likely that regulatory genes coding for DNA binding proteins such as members of the Api-AP2 , HMG , and Myb families play crucial roles as transcription factors in these processes , but these genes remain uncharacterized in these three related parasites . In this study , we describe the presence and genomic organization of these three types of genes in Babesia bovis , Babesia microti , and Theileria equi , highlighting the importance of the conservation of these genes and their possible contributions to parasite development through their different life stages . We also describe the occurrence of a previously unreported HMG gene in B . microti , an important emerging human pathogen; define the repertoire of eight conserved Myb genes; and describe the pattern of transcription of the regulatory AP2 , HMG , and Myb genes in B . bovis intra-erythrocytic stages for the first time . It is expected that these findings will elicit additional research in this field and contribute to the development of converged intervention strategies for the improved control of these devastating and generally under-studied diseases .
|
[
"Abstract",
"Introduction",
"Concluding",
"Remarks"
] |
[
"sequencing",
"techniques",
"parasite",
"groups",
"plasmodium",
"gene",
"regulation",
"regulatory",
"proteins",
"dna-binding",
"proteins",
"parasitic",
"protozoans",
"parasitology",
"developmental",
"biology",
"apicomplexa",
"review",
"protozoans",
"transcription",
"factors",
"sequence",
"motif",
"analysis",
"molecular",
"biology",
"techniques",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"proteins",
"gene",
"expression",
"life",
"cycles",
"molecular",
"biology",
"biochemistry",
"babesia",
"protein",
"domains",
"genetics",
"biology",
"and",
"life",
"sciences",
"organisms",
"parasitic",
"life",
"cycles"
] |
2016
|
Comparative Bioinformatics Analysis of Transcription Factor Genes Indicates Conservation of Key Regulatory Domains among Babesia bovis, Babesia microti, and Theileria equi
|
Little is known about the genetic basis of ecologically important morphological variation such as the diverse color patterns of mammals . Here we identify genetic changes contributing to an adaptive difference in color pattern between two subspecies of oldfield mice ( Peromyscus polionotus ) . One mainland subspecies has a cryptic dark brown dorsal coat , while a younger beach-dwelling subspecies has a lighter coat produced by natural selection for camouflage on pale coastal sand dunes . Using genome-wide linkage mapping , we identified three chromosomal regions ( two of major and one of minor effect ) associated with differences in pigmentation traits . Two candidate genes , the melanocortin-1 receptor ( Mc1r ) and its antagonist , the Agouti signaling protein ( Agouti ) , map to independent regions that together are responsible for most of the difference in pigmentation between subspecies . A derived mutation in the coding region of Mc1r , rather than change in its expression level , contributes to light pigmentation . Conversely , beach mice have a derived increase in Agouti mRNA expression but no changes in protein sequence . These two genes also interact epistatically: the phenotypic effects of Mc1r are visible only in genetic backgrounds containing the derived Agouti allele . These results demonstrate that cryptic coloration can be based largely on a few interacting genes of major effect .
Animal pigmentation has attracted substantial evolutionary interest because changes in color , be they driven by natural or sexual selection , can have profound effects on fitness . Dissecting the genetic basis of morphological variation , such as adaptive pigmentation , allows us to answer several long-standing evolutionary questions: How many genes contribute to adaptive phenotypes ? What are the relative sizes of their effects ? Are adaptive alleles generally dominant , semidominant , or recessive ? What types of genes are involved in adaptive change ? Do adaptive mutations generally occur in coding or regulatory regions ? What is the role of epistasis in evolutionary change ? To understand the genetic processes involved in generating adaptive color patterns , we revisited a series of classic natural history studies [1–3] that described geographic variation in coat-color pattern of the oldfield mouse ( P . polionotus ) . The extreme coat-color variation within this species is driven by selection for camouflage [4] , yielding a strong geographical correlation between coat color and reflectance of the substrate [5 , 6] . We focused on the two subspecies of P . polionotus showing the greatest difference in color pattern: P . p . subgriseus and P . p . leucocephalus . The mainland subspecies ( P . p . subgriseus ) occupies oldfield habitats in the southeastern United States and has a coat that is dark brown on top and light gray on the belly , as well as a striped tail . In contrast , the light-colored Santa Rosa Island beach mouse ( P . p . leucocephalus ) , like other “beach mice” that have colonized Florida's barrier islands and sandy coastal dunes , lacks visible pigmentation on its face , flank , and tail ( Figure 1 ) .
To analyze the genetic basis of color-pattern difference , we made reciprocal genetic crosses between three mainland and three beach mice , yielding 28 F1 hybrids that were then intercrossed to produce 465 F2 progeny . A genome-wide linkage map was generated using both anonymous microsatellite markers and single nucleotide polymorphisms ( SNPs ) in candidate pigmentation genes ( Figure 2 ) . This represents the first genome-wide linkage map for Peromyscus , with the exception of an allozyme-based map with few markers [7] . We scored all F2 progeny for 113 microsatellite markers fixed within but differing between the two subspecies ( Table S1 ) . We also scored F2 progeny for SNPs in 11 pigmentation genes chosen because of their chromosomal location and their known mutational effects on pigmentation in Mus ( Tables S2–S4 ) . In sum , we analyzed the linkage of 124 informative markers scored in all 465 F2 progeny ( 57 , 660 genotypes ) using JoinMap software [8] . The markers were ordered in 27 linkage groups ( LG ) based on a log likelihood of odds ( LOD ) ratio of linkage threshold of 4 . 2 ( permutation test , p = 0 . 05 ) . The combined LGs span 1 , 103 cM ( by comparison , the Mus genome comprises ∼1 , 300 cM [9] ) , with a mean interval length between markers of 8 . 9 cM . Because P . polionotus has 24 chromosomes [10] , we expect that the markers will collapse into 24 LGs when additional regions are screened . To identify which genomic regions were statistically associated with the pigmentation differences , we determined the phenotypes of F2 progeny in seven regions of the body . These regions show the most divergence in pigmentation between the subspecies and together accurately encapsulate the difference in color and pattern . We measured total pelage reflectance ( brightness ) and scored pigment pattern on individual hairs for four facial traits ( rostrum , cheek , eyebrow , and earbase ) and also calculated the extent of dorsal , rump , and tail pigmentation as three additional traits ( Figure 3A ) . The phenotypic data show no evidence for sexual dimorphism or maternal effects . The phenotypic correlation ( r ) between traits ranged from 0 . 29 to 0 . 82 ( the highest value between earbase and cheek ) , suggesting that while some genes cause similar pigmentation differences among different body parts , other genes have more localized effects ( Figure S1A ) . The distribution of phenotypic scores among F2 individuals was not consistent with simple Mendelian inheritance for any of the traits with the exception of tail stripe , which shows a bimodal distribution of scores ( Figure S1B ) . We analyzed these phenotypic values , along with the molecular marker data , using MapQTL 5 [11] . Only three LGs harbored quantitative trait loci ( QTL ) that influence pigmentation differences between the subspecies ( Figure 3B–3D ) . Because pigmentation has served as a model pathway for studies of gene action and interaction in a variety of biological processes , there are over 100 well-characterized genes known to affect pigmentation in laboratory mice [12 , 13] . Each of the three QTL regions contains only a single pigmentation gene from the homologous regions ( bounded by homologous microsatellite markers ) of the closely related model organisms Mus musculus and Rattus norvegicus: these genes are the Agouti signaling protein ( Agouti; LG 7 ) , the melanocortin-1 receptor ( Mc1r; LG 1 ) , and the c-kit receptor ( Kit; LG 14 ) . When mapped in Peromyscus , markers in these three candidate genes showed the highest LOD values for all seven pigmentation traits compared to other markers in the same LG ( Table 1 ) . Application of Multiple QTL model ( MQM ) mapping methods shows that none of the other eight candidate genes or 113 microsatellites is significantly associated with pigmentation variation . Our results suggest that nearly all of this difference is likely due to the three pigmentation genes Agouti , Mc1r , and Kit , although it is formally possible that other closely linked loci affect the color difference between subspecies . Below , we refer to these three QTLs using the names of the candidate genes . Each of the two regions of largest effect , Agouti and Mc1r , influence all seven pigmentation traits ( LOD > 5 . 8 ) . Agouti explains the greatest amount of pigment variation for three traits ( cheek , eyebrow , and tail ) , while Mc1r explains the greatest amount of variation for two traits ( rostrum and earbase ) . Both regions contribute equally to the extent of dorsal and rump pigmentation . The relative phenotypic effect of these two regions varies among traits ( Table 1 ) . For example , Agouti explains 78% of the variation in tail striping , but only 9% of the variation in dorsal pigmentation , while Mc1r explains 27% of the variation in rostrum pigmentation but only 1% of the variation in tail striping . Depending on the trait , the combination of these two loci explains between 19% and 80% of the variation for each of the pigmentation traits . The candidate gene Kit mapped to the only QTL of small effect , which explained less than 3 . 2% of the phenotypic variation among traits . This region is associated with only four traits ( rostrum , cheek , earbase , and tail; LOD > 3 . 0 ) , thus showing more spatial specificity than the two regions of major effect . The remaining phenotypic variance is likely attributable to other loci of small effect that are undetectable in a cross of this size and to environmental and/or epigenetic variation . Thus , a small number of chromosomal regions—and perhaps only a few genes—are responsible for most of the difference in color pattern between subspecies . One of the classical ways to determine the effects of genetic variation on pigmentation is to analyze the allelic composition of extreme classes in an F2 or backcross ( e . g . , 14 ) . An analysis of the most extreme phenotypes among our F2 progeny shows a striking association between phenotype and the allelic variation ( “light” allele derived from the beach parents [L] and “dark” allele derived from the mainland parents [D] ) at Agouti and Mc1r . Of the 50 F2 progeny with the lightest dorsal pigmentation , 42 had at least one light Mc1r allele ( LL or LD Mc1r genotypes; χ2 test , p < 0 . 0001 ) . Similarly , of the 113 F2 progeny lacking a tail stripe , 112 had at least one light Agouti allele ( LL or LD Agouti genotypes; χ 2 test , p < 0 . 0001 ) . The direction and magnitude of QTL effects were gauged by comparing phenotypic means among the F2 offspring . For the two major QTLs , the derived Agouti and Mc1r alleles increase the average coat reflectance ( i . e . , produce lighter color ) and reduce the extent of dorsal , rump , and tail pigmentation ( Table 1 ) , changes consistent with the idea that these alleles were fixed by natural selection in beach mice . In addition , population-specific alleles of Mc1r and Agouti show differences in dominance for all traits . For example , 30 F2 progeny lack pigmentation on the rump . Among these mice , the distribution of Mc1r alleles ( DD = 0 , DL = 2 , and LL = 28 ) suggests that the light Mc1r allele is largely recessive . By comparison , the distribution of Agouti alleles in the same 30 F2 progeny ( DD = 0 , DL = 15 , and LL = 15 ) suggests that the light Agouti allele is dominant to the dark allele . These patterns are consistent with the dominance hierarchy of these genes seen in laboratory mice , in which either recessive loss-of-function mutations in Mc1r or dominant gain-of-function mutations in Agouti yield lighter pigmentation [15 , 16] . Mc1r is an integral membrane protein of melanocytes , which are pigment-producing cells . Agouti , the ligand of Mc1r , is an inverse agonist that , when bound , reduces Mc1r activity ( via lowered cAMP signaling ) resulting in lighter pigmentation . Thus , it is the biochemical interaction between these two proteins that controls the switch between dark eumelanin and light phaeomelanin production in melanocytes [17] . Previous work showed that laboratory populations of beach and mainland mice differ by a fixed , single amino acid mutation that reduces Mc1r's signaling potential [18] , but additional changes in Mc1r expression levels have not been ruled out as a contributor to the difference in pigmentation ( see below ) . To identify whether a coding change in the Agouti gene itself might also contribute to the pigmentation differences between our subspecies , we sequenced Agouti's three translated exons ( encoding a 139 amino acid protein ) in the six original parents of our cross ( Figure S2 ) . The beach and mainland sequences did not differ by any fixed nonsynonymous mutations , demonstrating that amino acid changes in Agouti are not responsible for the color differences . In addition , sequencing of Agouti cDNA products showed that both mainland and beach mice produced an intact and spliced Agouti transcript similar to that observed in Mus . Because most of the Agouti mutations that produce light coloration in laboratory mice involve gain-of-function cis-regulatory mutations [19] , we also tested the prediction that an increase in Agouti expression contributes to the light coloration of beach mice . To examine whether differences in expression level of Mc1r or Agouti influence color patterning , we conducted gene expression assays ( reverse transcriptase PCR [RT-PCR] and quantitative-PCR [q-PCR] ) on adult skin taken from five of the seven assayed pigmentation areas ( Figure 4 ) . Specifically , we performed parallel expression analyses for Mc1r , Agouti , and beta-Actin ( a ubiquitously expressed control gene ) mRNAs in the two polionotus subspecies . We also included mRNA from their fully pigmented sister species P . maniculatus , to determine which subspecific expression pattern is derived and which is ancestral . As a control , we compared patterns of Mc1r and Agouti expression in skin taken from the dorsum , a region that shows similar levels of pigmentation in beach and mainland subspecies and thus should show little difference in Mc1r and Agouti expression . Dorsal skin showed no significant difference in Mc1r expression among the three taxa ( analysis of variance , p = 0 . 96 ) . Agouti expression on the dorsum did not differ significantly between beach and mainland subspecies ( p = 0 . 13 ) , but Agouti was expressed at a lower level in P . maniculatus than in either polionotus subspecies ( p < 0 . 01 ) , consistent with P . maniculatus's darker dorsal pigmentation . We also compared levels of Mc1r expression between the three taxa for four body regions ( rostrum , cheek , eyebrow , and earbase ) that show large differences in pigmentation between beach and mainland mice . There was no difference in Mc1r expression level among taxa or among body regions ( analysis of variance , p > 0 . 05 ) and no correlation between Mc1r expression level and reflectance among taxa across body regions when all taxa were included ( Figure 4; r = 0 . 45 , R2 = 0 . 20 , and p = 0 . 10 ) . Thus , taken together with earlier functional analyses [18] , these data suggest that a single amino acid mutation in the coding region of Mc1r—and not mutations in neighboring cis-regulatory regions—produces light pigmentation in beach mice . Finally , in the same four body regions , Agouti expression was always significantly higher in tissues from beach mice than in tissues from mainland mice ( Student's t-test , p < 0 . 05 , two-tailed test ) . Comparing Agouti expression in P . polionotus to its sister species P . maniculatus , we find that the increased expression in beach mice is a derived trait because both P . p . subgriseus and P . maniculatus have similarly low levels of Agouti expression . In addition , Agouti expression is significantly correlated with pelage reflectance when all three taxa are compared ( Figure 4C; R2 = 0 . 65 , and p < 0 . 001 ) . Agouti also explains spatial variation in light coloration within a subspecies; there is a significant positive correlation between pelage reflectance and Agouti expression across body regions in beach mice ( R2 = 0 . 91 , and p < 0 . 05 ) . Together , these results suggest that increased expression of Agouti , caused by either mutation ( s ) in its cis-regulatory region or in closely linked trans-factors , also contributes to the light phenotype of beach mice . Because the Mc1r and Agouti proteins interact physically , we tested for epistasis by performing gene interaction analyses ( MapManager QTXb [20] ) . We found evidence of epistasis in several pigmentation traits ( e . g . , eyebrow: LOD score = 11 . 28; χ2 test , p = 0 . 001 and rostrum: LOD score = 10 . 32; χ2 test , p = 0 . 001 ) . We also examined the effect of Mc1r genotypes on different Agouti backgrounds ( and vice versa ) using a categorical measurement of pigmentation . We detected epistasis for all seven of the traits but most strikingly for cheek pigmentation ( Figure 5 ) : F2 mice with the Agouti DD genotype always have fully pigmented hairs regardless of Mc1r genotype . In fact , for all seven traits there is no difference in phenotype between individuals who have either the DD or DL Mc1r genotypes on an Agouti DD genetic background ( Student's t-test , p > 0 . 05 ) , and for only a few traits ( e . g . , tail ) is the Mc1r LL genotype visible on the Agouti DD background . In contrast , in F2 mice having the Agouti LL genotype , the Mc1r genotype explains all of the variation in cheek pigmentation ( r = 1 . 0 and p < 0 . 0001 ) ; and double mutants ( Agouti LL and Mc1r LL ) always lack visible pigment in their cheek hairs ( Figure 5 ) . Clearly , Mc1r genotype has a significant effect on pigmentation only on the light Agouti ( LL and LD ) background , a pattern mirrored in all seven traits . Thus , the effect of each of the two genes on phenotype clearly depends on the genotype at the other locus . It is striking that the interaction between Agouti and Mc1r in these mice is just the reverse of what we would predict from the classical genetics of the laboratory mice . In the pigmentation pathway , Mc1r is downstream of Agouti action , and in the laboratory variation in Mc1r has been shown to mask the action of Agouti alleles [21 , 22] . In contrast , we show here that Agouti can mask Mc1r , even though the dominance hierarchy of alleles remains identical to that seen in laboratory mice . We can explain this pattern of “reverse epistasis” mechanistically . The single mutation in Mc1r significantly decreases agonist ( αMSH ) binding , and hence cAMP signaling , but does not eliminate the receptor functionality of its protein product [18] . However , it is only with increased expression of Mc1r's antagonist Agouti that the phenotypic effects of this weakened Mc1-receptor are revealed , a result consistent with Agouti's ability to decrease cAMP production independent of αMSH . Thus , it is clear that epistasis is a property of particular alleles rather than of loci themselves , and thus epistatic interactions observed in the laboratory may differ from those seen in natural populations [see 23] . In sum , our genome-wide linkage map for Peromyscus allowed us to identify three genetic regions , two of which have major phenotypic effects on the adaptive difference in color pattern between subspecies of P . polionotus . Moreover , these regions contain the well-studied Agouti and Mc1r pigmentation genes [13] . While mutations in Mc1r are correlated with pigmentation in a number of wild vertebrates [see 13] , our results are , to our knowledge , the first example of variation at the well-studied Agouti locus being associated with adaptive variation of animal coloration in nature . Our results also have several implications for understanding the genetic basis of adaptation . First , this subspecific difference in color pattern is produced by a few interacting genes of large effect , supporting the idea that adaptations can involve relatively few genes rather than , as is often believed , many genes of small effect [24] . When an animal suddenly invades a novel habitat , their ancestral phenotype is often very different from the new optimal phenotype ( as was almost certainly true for beach mice ) . Indeed , population genetic theory predicts that mutations of large effect will often be involved in adaptation in these circumstances [25] , a prediction consistent with several studies on the genetic basis of mimicry and crypsis [26–28] . Second , both structural mutations ( a single amino acid change reducing Mc1r signaling potential ) and regulatory mutations ( a derived increase in Agouti expression ) contribute to adaptive change , and this change involves both recessive ( Mc1r L ) and dominant ( Agouti L ) alleles . These results support the idea that adaptation is not necessarily driven largely by cis-regulatory changes [29 , 30] or by ( semi ) dominant alleles [31 , 32] . Third , we show that the nature of epistasis between Mc1r and Agouti in wild populations does not mirror that seen in the laboratory , suggesting that one should be cautious not only about extrapolating the genetics of laboratory strains to evolution in nature , but also about inferring the directionality of biochemical pathways from patterns of gene interactions . Finally , most genetic studies of morphological change have concentrated on the loss of phenotypic traits through loss-of-function mutations ( e . g . , reduced armor in stickleback fish [33 , 34] , absence of wing spots in Drosophila [35] , and lack of pigment in cavefish [36] ) . This study provides a novel example of how adaptation can result from mutations involving a gain of function .
Parental stocks were maintained at the Peromyscus Genetic Stock Center ( University of South Carolina , United States ) . Maintenance of stocks and the crossing design have been described previously [18] . A total of seven pigmentation traits ( rostrum , cheek , eyebrow , earbase , and the extent of pigmentation on the dorsum , rump , and tail ) were scored in all 465 F2 individuals . A spectrophotometer ( Ocean Optics , http://www . oceanoptics . com ) was used to capture reflectance spectra from the four facial traits ( rostrum , cheek , eyebrow , and earbase ) . A reflectance probe was held at a 45 ° angle to the surface , and the program OOIbase32 ( Ocean Optics ) was used to capture reflectance measurements from 300–700 nm . Brightness was calculated by summing the area under the reflectance curve and converting to a normalized reflectance [37] . The extent of dorsal and tail pigmentation was measured as the proportion of the area that was pigmented . Rump pattern was scored using five categories from minimally ( 0 ) to fully ( 4 ) pigmented ( following [1–3] ) . In addition , categorical pigmentation values ( 0–2 ) were scored for all seven pigmentation traits in the F2 progeny ( following [18] ) . All statistical correlation analyses for the color traits were performed using JMP version 5 . 1 . 2 statistical software package ( SAS Institute , http://www . sas . com ) . All F2 individuals were genotyped for a total of 113 anonymous microsatellite markers and 11 SNPs in pigmentation genes . Microsatellites were cloned from enriched partial genomic libraries developed for P . maniculatus bairdii and P . polionotus subgriseus [38] . Cloned sequences were edited in Sequencher 3 . 1 . 1 ( Genecodes , http://www . genecodes . com ) , and microsatellite motifs were identified by eye . PCR primers designed to amplify the repeat motifs were used to genotype the six parental mice ( three beach and three mainland parents ) . Out of 400 microsatellite loci tested , 113 showed diagnostic differences between individuals from the two subspecies , and these were scored in the 465 F2 progeny . All microsatellite loci were inherited in a codominant manner and were anonymous ( with the exception of one microsatellite identified in a pigmentation gene , t-box protein 15 [Tbx 15] ) . Microsatellite markers used to construct the linkage map are listed in the Table S1 . All PCRs were performed in a 15 μl volume using Eppendorf Mastercycler Gradient thermal cyclers ( http://www . eppendorf . com ) . Each reaction included 30 ng of template DNA , 10× Taq Buffer with 1 . 5 mM MgCl2 ( Eppendorf ) , 0 . 3 μL of 10 mM dNTPS , 0 . 6 μM each of a fluorescently labeled forward primer , unlabeled reverse primer , and 0 . 15 units Taq DNA polymerase ( Eppendorf ) . The majority of microsatellite primers were synthesized with a known CAG ( 5′-CAGTCGGGCGTCATCA-3′ ) or M13R sequence ( 5′-GGAAACAGCTATGACCAT-3′ ) attached to the 5′ end . The PCR master mixes used in this system included 0 . 06 μM of the sequence-tagged primer , 0 . 6 μM of the untagged primer , and 0 . 54 μM of the fluorescently labeled probe . The cycling conditions for all primer pairs followed a touchdown protocol ( successively lower annealing temperatures ) . PCR parameters were: 94 °C for 90 s , followed by 21 cycles of denaturation at 94 °C for 30 s , 55 °C annealing for 30 s , and 72 °C for 1 min . The initial annealing temperature decreased by 0 . 5 °C for each of 20 cycles . An additional 15 cycles were performed as follows: 94 °C for 30 s , followed by 30 s at the last temperature , and 72 °C for 1 min . The final extension occurred at 72 °C for 5 min . Amplification products were scored on an ABI 3100 ( http://www . appliedbiosystems . com ) in a 96-well format and genotyping was multiplexed by labeling loci with different 5′ fluorescent dyes: 6-FAM ( blue ) , VIC ( green ) , and NEB ( yellow ) . Rox Genescan 400HD ( Applied Biosystems ) was used as internal size standard , and PCR products were analyzed with Genemapper version 3 . 5 software ( Applied Biosystems ) . In addition to microsatellite markers , 11 candidate pigmentation genes were screened for SNPs that were diagnostic between the two P . polionotus subspecies ( Table S2 ) . Candidate genes were chosen based on their known phenotypic effects , both on pigmentation and pleiotropic effects on other traits , in laboratory mice . For each candidate gene , PCR primers were designed in conserved exonic regions based on alignments of mouse , rat , and human sequences . Amplification primers were designed to span introns to maximize variation between subspecies . Following PCR optimization , introns were amplified in all six parents to identify diagnostic polymorphisms . Sequences were edited using Sequencher , and diagnostic markers were identified by eye . PCR primers and amplification conditions are listed in Table S3 . Genotyping of three candidate loci , Kit , Kitl , and Hps4 , was performed using a restriction enzyme digest assay . One microgram of Kit amplification product was digested at 60 °C for 60 min with 1 unit of BsiEI , 1× NEBuffer , and 100× BSA ( 10 mg/ml ) in a total reaction volume of 50 μL . Kitl and Hps4 amplicons were digested in a total reaction volume of 15 μl at 37 °C for 4 h using Hpy188 III and PspOM I enzymes , respectively . Digestion products were visualized on a 1 . 5% agarose gel stained with ethidium bromide . A polymorphic microsatellite was identified in the first intron of Tbx 15 . Genotyping of Tbx 15 was scored on an ABI 3100 in a 96-well format . Genotyping of seven candidate pigmentation genes ( Mc1r , Agouti , Tyr , Atrn , Slc24a5 , Pldn , and Mgrn1 ) was performed on an ABI 7000 using a TaqMan assay . A total of 60 ng of genomic DNA was used in each reaction , and cycling parameters were as follows: 40 cycles of 50 °C for 2 min , 95 °C for 10 min , and 92 °C for 15 s followed by an allelic discrimination step of 60 °C for 2 min . The TaqMan primer sequences are listed in Table S4 . A genetic linkage map was generated using JoinMap version 3 . 0 [8] on a locus file containing genotypes of a total of 124 molecular markers in 465 F2 progeny , with the population type set for segregation of two alleles per locus ( F2 population ) . JoinMap was used with an LOD score threshold of 6 . 0 to assign 120 of 124 loci to 27 LGs . For each LG , a map was created considering: Kosambi mapping function , default LOD ( 1 . 0 ) and recombination ( 0 . 4 ) thresholds , jump threshold of 5 . 0 , and not fixed order . A ripple analysis was performed after all markers on the LG were added to the map . This analysis attempts to improve the order of the loci in a chromosome by testing alternative orders created by local permutations of the locus order . All quantitative measures of pigmentation traits were analyzed with MapQTL 5 [11] using the interval mapping ( IM ) method , which fits a single QTL model ( additive versus dominant model ) . Using likelihood ratio tests in MapManager QTXb [20] , we verified that the additive versus dominance model was the best model of allelic effects . Similar mapping results were observed for the quantitative and categorical datasets . The MapQTL 5 parameters used were: mapping step size of 2 . 0 cM , maximum of 200 interactions , functional tolerance value of 1 . 0e−8 , and a minimum of five flanking markers to resolve incomplete genotypes . MQM mapping was performed in LGs where several QTLs were detected . Cofactors for MQM analyses were automatically selected with a p-value of 0 . 02 . Results from MQM analyses improved initial IM outputs by identifying from multiple to a single QTL per LG . Significance thresholds for determining linkage were chosen using conservative criteria for genome-wide linkage mapping in noninbred individuals: significant linkage of LOD ≥ 4 . 5 [39] . Significance of LOD values for each trait was confirmed by permutation tests in MapQTL 5 , with a genome-wide significance level of α = 0 . 05 and 1 , 000 iterations . Calculation of the percentage of phenotypic variance explained ( p . v . e ) by a QTL was performed in MapQTL 5 on the basis of the population variance found within the progeny of the cross . Gene interaction analyses to identify epistasis between QTLs were performed using MapManager QTXb . Probability of association was set at p = 0 . 0001 , and the LOD thresholds for each quantitative trait were estimated by permutation tests . Total RNA was isolated for four facial regions and the dorsum from shaved adult skin tissue of P . polionotus leucocephalus , P . p . subgriseus , and P . maniculatus using TRIzol reagent ( Invitrogen , http://www . invitrogen . com ) following the manufacturer's protocol . Total RNA was treated using DNase I ( New England BioLabs , http://www . neb . com ) . Subsequent reverse transcriptase reactions were performed using the Titan One tube RT-PCR kit ( Roche , http://www . roche . com ) with specific Peromyscus primers of Agouti ( forward 5′-TCTCTGGTGGGTGGGACTTC-3′ and reverse 5′-TGATTTTAGCCTCCATTAGGTTTCC-3′; exons 2–4 ) , Mc1r ( forward 5′-TGGACATACAGAATTGCCATGAG-3′ and reverse 5′-CAACCACACAGCCGTCCTAA-3′; exon 1 ) , and beta-Actin ( forward 5′-TCCTGACTGAGCGTGGCTATAG-3′ and reverse 5′-TCTCTTTGATGTCACGCACGAT-3′; exon 4 ) genes . Although Agouti has two differentially expressed transcripts , these primers were designed to measure expression of both isoforms simultaneously . For all experiments , both no-RT and no-Template controls were included . PCR products were visualized on a 1 . 5% agarose gel . RT-PCR were also performed using the SuperScript III Reverse Transcriptase ( Invitrogen ) , RNaseOUT ( Invitrogen ) , and the oligo ( dT ) 20 . qPCR amplifications were conducted in 20 μl reactions containing approximately 100 ng of total cDNA , 10 μl 2× TaqMan Universal PCR Master Mix , and 1 μl 20× TaqMan gene expression assay of Agouti , Mc1r , and beta-Actin . The amplification protocol used was as follow: initial 10 min denaturation at 95 °C followed by 50 cycles of 95 °C for 15 s and 60 °C for 1 min . Amplification signals were detected continuously with an ABI 7000 sequence detection system . All expression assays were done in either duplicate or triplicate . Analysis of the qPCR data was conducted by calculating the average Ct value across replicate experiments for each target gene ( Mc1r and Agouti ) and normalizing by the average Ct value of the reference gene ( beta-Actin ) for a specific tissue . Significance of the qPCR data was determined by one-way ANOVA and Student's t-tests using the JMP statistical package .
The GenBank ( http://www . ncbi . nlm . nih . gov/Genbank ) accession numbers for the sequences discussed are: EU020066 ( Peromyscus maniculatus ) , EU020067 ( Peromyscus polionotus subgriseus ) , and EU020068 ( Peromyscus polionotus leucocephalus ) .
|
The tremendous amount of variation in color patterns among organisms helps individuals survive and reproduce in the wild , yet we know surprisingly little about the genes that produce these adaptive patterns . Here we used a genomic analysis to uncover the molecular basis of a pale color pattern that camouflages beach mice inhabiting the sandy dunes of Florida's coast from predators . We identified two pigmentation genes , the melanocortin-1 receptor ( Mc1r ) and its ligand , the agouti signaling protein ( Agouti ) , which together produce a light color pattern . We show that this light pigmentation results partly from a single amino acid mutation in Mc1r , which reduces the activity of the receptor but does not affect the gene's expression level , and partly from the derived Agouti allele , which shows no change in protein sequence but does exhibit an increase in mRNA expression . We also show that these two genes do not act additively to produce pale color; rather , the derived Agouti allele must be present to see any effect of Mc1r on pigmentation . Thus , the light color pattern of beach mice largely results from the physical interaction between a structural change in a receptor ( reducing Mc1r activity ) and a regulatory change in the receptor's antagonist ( increasing Agouti expression ) .
|
[
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
] |
[
"evolutionary",
"biology",
"mus",
"(mouse)"
] |
2007
|
Adaptive Variation in Beach Mice Produced by Two Interacting Pigmentation Genes
|
In the context of an ageing population , understanding the transmission of infectious diseases such as scabies through well-connected sub-units of the population , such as residential care homes , is particularly important for the design of efficient interventions to mitigate against the effects of those diseases . Here , we present a modelling methodology based on the efficient solution of a large-scale system of linear differential equations that allows statistical calibration of individual-based random models to real data on scabies in residential care homes . In particular , we review and benchmark different numerical methods for the integration of the differential equation system , and then select the most appropriate of these methods to perform inference using Markov chain Monte Carlo . We test the goodness-of-fit of this model using posterior predictive intervals and propagate forward the resulting parameter uncertainty in a Bayesian framework to consider the economic cost of delayed interventions against scabies , quantifying the benefits of prompt action in the event of detection . We also revisit the previous methodology used to assess the safety of treatments in small population sub-units—in this context ivermectin—and demonstrate that even a very slight relaxation of the implicit assumption of homogeneous death rates significantly increases the plausibility of the hypothesis that ivermectin does not cause excess mortality based upon the data of Barkwell and Shields .
Stochastic ( random ) models play a pivotal role in the description of transmission and control of infection within small closely-connected populations . The typical example of such populations is households [2–5] , although the household methodology is generalisable to any well-connected population sub-unit such as the residential and nursing care homes ( RNCs ) we consider here . These models are increasingly becoming useful tools for studying disease transmission and control in structured populations , which can in part be attributed to the availability of household-stratified infection data that can be used to parameterise the models [6–8] as well as affordability in the amount of computing power . A type of household transmission model that is currently growing in popularity is a stochastic Markovian household model [2 , 9 , 10] . In this type of model , individuals are assumed to have two levels of mixing: one representing transmission between people sharing/living in the same household and the other representing global contacts within the population . These kind of models have the advantage that they capture the temporal behavior of the epidemic and offer a computational trade-off between simpler whole-population models [11 , 12] and more complex , computationally-intensive individual-based models [13 , 14] . Stochastic models have traditionally been studied via simulation and estimation based upon a large number of event-driven integer-based simulations [15 , 16] . These methods are powerful but require a large number of replicates to reduce Monte Carlo error , since it is typically unclear if a single simulation represents the average behaviour of the system or the outcome of a combination of rare events . They therefore quickly become computationally intensive due to the combination of the sheer number of replicates required and the number of possible events that can occur in a given time step . In this work , we have presented a method that allows for the complete range of stochastic behaviours to be captured by a large set of ordinary differential equations ( ODEs ) which we will refer to as the master equation ( also known as the forward equation ) . The master equation is a set of linear ODEs representing the probability of being in each possible state with the dynamics driven by the rates of transition between states . This method has previously been applied to the study of stochastic disease dynamics [9 , 17–19] . The use of a master equation has existed for quite a long time but has not been widely used in epidemiology and this is partly due to the algorithmic difficulties involved in solving the resulting large system of linear ODEs . There are a range of methods that can be used to solve the system and in this work we concentrate on so-called series-expansion and projection-based methods . Both method classes are based on the efficient approximation of the action of the matrix exponential on a vector , a problem that has attracted a lot of research and for which various numerical methods have been proposed [20–22] . Besides the existence of these numerical methods used to solve the master equation , there still lacks a body of work that computationally benchmarks the algorithms for use in mathematical epidemiology , an area where these algorithms can be of great utility . In this work , we benchmark a number of competing algorithms and then use the best algorithm to solve a system of linear ODEs describing the transmission of scabies in care homes in the UK to enhance the efficiency and quality of an inferential and policy-driven modelling study . We also consider how the safety and efficacy of pharmaceutical interventions can be assessed using probabilistic models . Sarcoptes scabiei is an ectoparasite that infests human skin , where it burrows and lays eggs causing intense itching and scratching , which may in turn lead to secondary bacterial infection [23] . Global prevalence in 2015 was estimated at 204 million [24] but varies enormously both between and within nations [25] . All age prevalence over 30% has been reported in some countries ( Papua New Guinea , Panama , Fiji ) [26] and overcrowded settings such as slums and refugee camps can have very high prevalence [27] . In most low and middle income settings it particularly affects children [27] . However in developed high-resource settings , the incidence and prevalence in children and schools has declined , while outbreaks are commonly reported in residential and nursing care homes for the elderly [28 , 29] . The mite is transmitted mainly through skin to skin contact , and also to a lesser extent through “fomites” ( e . g . , bedding , skin flakes ) [23 , 30] . Itching begins 4–6 weeks after exposure for a first episode , but can start within 24 hours in subsequent infections [30 , 31] . An infested person becomes infectious in most cases around 10–14 days after becoming exposed , when newly fertilised adult female mites become ready to seek new burrows in which to lay eggs [30] . Scabies is commonly misdiagnosed since the classical physical signs ( burrows and papules ) are variably present . This is a particular problem in RNCs since many residents have dementia and may not be able to say they are itching , while their increased agitation may be attributed to other causes . As a consequence , recognition of cases and outbreaks in RNCs is often delayed [28] . In the absence of interventions , scabies is generally not self-limiting [30 , 32] with a study in Bangladesh observing that children could remain infected for more than six months [33] . In the UK , first-line therapy for scabies is topical permethrin , which is applied all over the body , left on for 8 hours before being washed off , and repeated 7 days later . In RNC outbreaks , residents and staff need to be treated simultaneously . This can be distressing and logistically challenging [28] , especially as some guidelines recommend prophylactic mass treatment of all residents and staff once an outbreak is declared [34] . Oral ivermectin has been suggested as an effective alternative in healthcare settings , and is included in French national outbreak management guidelines for RNCs [34] . However , its wider usage as a mass treatment in RNCs has been limited in part due to safety concerns raised by Barkwell and Shields [1] . The retrospective study was carried out in a 47-bed closed unit for residents with behavioural tendencies over a period of six months between June and November 1995 . A scabies outbreak occurred during this period and the individuals were treated with two different topical agents , lindale and permethrin , but scabies symptoms re-occured . Consequently , the patients were treated with a single oral dose of ivermectin and all the rashes and symptoms had cleared within five days with the individuals requiring no further treatment . However , during the following six months , the authors observed an increased pattern of excess deaths among the residents who had received ivermectin . Barkwell and Shields , in their conclusion , subsequently advised against using ivermectin to treat scabies in the elderly and/or those with an underlying medical condition , suggesting a potential causal association with deaths in the facility . We re-examine their statistical analysis and interpretation and provide a rigorous method for more careful analysis .
The natural history of scabies infection in the absence of interventions is highly dependent on the history of previous exposure as well as immunological competency of the individuals . Walton [35] has reported that spontaneous recovery of scabies in healthy adults can occur only with subsequent re-infestations . Additionally , parasite numbers can be reduced and in approximately 60% of cases re-infestation of sensitised hosts was unsuccessful . It is still unknown how long this capacity for some level of acquired immunity persists , though 15-24 months after infestation with scabies mite extracts injected intradermally have failed to induce immediate wheal reactions in patients . So in the elderly and especially those in care homes with high co-morbidities and compromised immunological responses , we do not include the possibility of spontaneous recovery in the absence of treatment . This would mean that following exposure to scabies , the individuals would have a protracted infection that is not self limiting . We therefore assume that scabies follows an SEI model framework in which individuals are initially susceptible ( S ) , then following an infection event spend some time in a latent ‘exposed’ class ( E ) . Once a fertilised female mite is transferred to a susceptible individual , mite generation time means there is a delay , between 7 and 14 days , before the host can become infective . However , during this period , the mite burrows can still be observed on the host’s skin [30] . Eventually , the individuals become infectious , ( I ) , and are able to infect others . Our starting point is therefore the stochastic SEI model in a closed population of size N . This consists of three non-independent random variables , S ( t ) , E ( t ) and I ( t ) such that S ( t ) + E ( t ) + I ( t ) = N , for all values of t , representing the number of individuals who are uninfected with scabies ( Susceptible ) , who have been infected but are not able to infect others ( Exposed ) , and who have been infected and are able to infect others ( Infectious ) respectively . The state transitions and rates in this model are ( S , E , I ) → ( S - 1 , E + 1 , I ) at rate λ S I , ( S , E , I ) → ( S , E - 1 , I + 1 ) at rate γ E . ( 1 ) If we define the expectations S ¯ = E [ S ] , E ¯ = E [ E ] and I ¯ = E [ I ] then in the limit of large N the dynamics of this model will be governed by the more familiar deterministic differential equations: d S ¯ d t = - λ S ¯ I ¯ , d E ¯ d t = λ S ¯ I ¯ - γ E ¯ , d I ¯ d t = γ E ¯ . ( 2 ) To model the finite-population dynamics , let Ps , e , i ( t ) represent the probability that there are s , e , i numbers of susceptibles , exposed and infected respectively in the population at time t , then the complete dynamics will be modelled by considering all the possible infection configurations as shown in the system ( 3 ) : d P s , e , i d t = γ ( - e P s , e , i + ( e + 1 ) P s , e + 1 , i - 1 ) + λ ( - s i P s , e , i + ( s + 1 ) i P s + 1 , e - 1 , i ) + τ I ( t ) ( - s P s , e , i + ( s + 1 ) P s + 1 , e - 1 , i ) . ( 3 ) Eq ( 3 ) can be equivalently represented by counting the number of events of each type that occur rather than the number of individuals in each compartment . This is known as the DA ( Degree of Advancement ) representation and has previously been described elsewhere [36 , 37] . If we define Z1 and Z2 as the number of exposure ( E ) and progression to active infection ( I ) events respectively , then the state space of the process at time t can be denoted as Z ( t ) = ( Z1 , Z2 ) . If we index the states of the system as zi = ( z1 , z2 ) with i = 1 , … , n where n = ( N + 1 ) ( N + 2 ) 2 is the size of the state space , we can then order the states of the system such that zi < zi+1 . The within-household transmission parameter λ in the system ( 3 ) is modelled as λ = β ( N - 1 ) α , ( 4 ) where β > 0 is an overall scaling for transmission and α represents the different ways that mixing behaviour can change with population size , N . If α = 0 then every pair of individuals in the same population makes contacts capable of spreading disease at the same rate regardless of N; and if α > 0 then larger populations reduce the rate of transmission as if each infective individual had a certain demand for contacts that are evenly spread throughout the population . If α < 0 then larger populations enhance transmission—while this would not normally be considered in the context of households , for RNCs we consider there is the possibility that larger facilities will have more opportunities for contact due to , for example more activities in larger communal areas . The parameter τ , represents the transmission between general members of the community i . e . between household mixing , and the proportion of the overall population that is infective is given as I ( t ) = ∑ s , e , i i P s , e , i ( t ) ∑ s , e , i ( s + e + i ) P s , e , i ( t ) . As we are considering a small number of carehomes in a large population [28] , we assume that there is no contact between members of different carehomes and therefore no between carehome transmission . It follows then that if carehomes are independent then τ = 0 . A more rigorous derivation of Eq 3 can be found in literature [2 , 38] . More insight can be gained by representing the system ( 3 ) in vector notation . Let p be the column vector of the probabilities of a household being in a certain configuration at time t . Then ( 3 ) can be expressed more succinctly as a linear constant-coefficient initial value problem , the so-called master equation , d p d t = Q p , p ( 0 ) = p 0 , ( 5 ) where Q ∈ R n × n is the household transition matrix of order n ( with n being equal to the total number of states the system can occupy; for our SEI model n = ( N + 1 ) ( N + 2 ) /2 as detailed above ) and the probability column vector p 0 ∈ R n represents the initial configuration at time t = 0 . The household transition matrix has the property that its elements sum to zero column-wise . The solution vector p ( t ) represents the transient behavior of the finite-state Markov chain and is easily shown to be a probability vector for all t ≥ 0 . The solution of the master Eq ( 5 ) is given by p ( t ) = exp ( t Q ) p 0 , ( 6 ) where exp ( tQ ) = I + ( tQ ) + ( tQ ) 2/2 ! + ⋯ denotes the matrix exponential; see , e . g . , [39 , Chapter 10] . In what follows we sometime drop time t for notational simplicity . Note that the matrix Q is typically sparse because there is a limited number of transitions that a household in a certain configuration can make , i . e . , there are few epidemiological state changes compared to the size of the matrix Q . Because a household can only move to a state following the current one in the DA representation , the states can be ordered so that Q is an upper triangular matrix . This leads to computational savings in some of the algorithms discussed later . The matrices we consider here also have a small bandwidth because we consider epidemiological processes with a limited number of events , though this is not a general feature of epidemiological models . Technically , we are concerned with the fast and sufficiently-accurate computation of the matrix exponential in Eq ( 6 ) . For scalar problems ( i . e . , n = 1 ) the computation of the exponential is trivial . However , the problem becomes challenging as n gets larger in which case the matrix Q is hopefully sparse or otherwise structured as in our case . We make use of the fact that the full matrix exponential exp ( tQ ) is not required , but merely the vector-matrix product exp ( tQ ) p0 . Computationally , these two are different problems and this section focuses on methods which compute this product directly without forming the matrix exponential itself . Methods based on polynomial and rational approximants have proven to be particularly efficient for this task . They have in common that exp ( tQ ) p0 ≈ r ( Q ) p0 where r is a well-chosen polynomial , or more generally a rational function , which depends on t and the required approximation accuracy . In the following we review a number of methods which fit into this framework . We refer the reader to [20] and [39 , Chapter 10] for further reading . It is easy to ensure that all discussed methods return probability vectors by adding a procedure at each time step that zeros-out all negative numbers and renormalizes the result to have unit 1-norm . If the computed vector is sufficiently accurate , such a normalization procedure does not affect the error significantly . More precisely , let p = exp ( tQ ) p0 be the exact probability vector such that p ≥ 0 component-wise and ‖p‖1 = 1 . Further , let p ˜ ≈ p be a numerical approximation such that ‖ p - p ˜ ‖ 1 ≤ ε ⪡ 1 . We define P to be the operator that zeros out negative entries of a vector , i . e . , ( P p ˜ ) i = { p i ˜ , if p ˜ i ≥ 0 , 0 , if p i ˜ < 0 , where the subscript i refers to the ith component of a vector . Then , using basic vector norm inequalities and the fact that | p i - ( P p ˜ ) i | ≤ | p i - p ˜ i | , we have | 1 - ‖ P p ˜ ‖ 1 | = | ‖ p ‖ 1 - ‖ P p ˜ ‖ 1 | ≤ ‖ p - P p ˜ ‖ 1 ≤ ‖ p - p ˜ ‖ 1 ≤ ε , and hence ‖ P p ˜ ‖ 1 = 1 + δ with |δ| ≤ ε . Now , for the the normalized vector p ^ = P p ˜ / ‖ P p ˜ ‖ 1 we have ‖ e - p ^ ‖ 1 = 1 1 + δ ‖ ( 1 + δ ) p - P p ˜ ‖ 1 ≤ 2 ε 1 - ε . Hence , the normalization procedure guarantees a probability vector p ^ and only increases the approximation error ‖ e - p ^ ‖ 1 by a factor ≈2 compared to the error of the non-normalized approximation p ˜ . In general , we work in a Bayesian framework , which has the benefit of dealing with the statistical challenges of the small datasets we consider in a systematic manner . We do this by calculating the posterior distribution , f , over parameters θ , given data D , using Bayes’ theorem: f ( θ | D ) = L ( D | θ ) π ( θ ) ∫ L ( D | ϑ ) π ( ϑ ) d ϑ , ( 10 ) where L is the likelihood of the data given the parameters and π is the prior distribution over parameters . If the integral in the denominator of the right-hand side of ( 10 ) above is tractable , then we can simply evaluate f directly , but if it is not then we can use Markov chain Monte Carlo ( MCMC ) methods to produce samples from the posterior distribution [51 , 52] . We fitted the stochastic SEI model above to scabies infection data from a study that enrolled carehomes in the UK [28] . In the study , the authors investigated a series of suspected scabies outbreaks in residential care homes , exploring barriers to early recognition and optimal management . Seven care homes agreed to participate and questionnaires were administered requiring details about dates of onset , diagnosis and treatment , clinical features , underlying illness , pre-existing skin conditions and mobility . An outbreak was determined if a report of two or more clinically suspected cases of scabies in a residential care home were reported to the Surrey and Sussex Health Protection Teams of the Public Health England ( PHE ) by a GP or a carehome manager . Case definitions included suspected cases because a definite diagnosis of scabies by dermatoscopy or microscopy is rare within the carehome setting and not all symptomatic cases had been seen by a doctor . The data we used to fit to the model are tabulated in Table 1 . These data involves scabies outbreaks in seven different care homes , for which the resident population ( N ) , days from onset to diagnosis ( T ) which is also the point at which treatment is initiated , and the number of scabies cases treated ( C ) are recorded . Days from onset of symptoms was defined as the first reported day of itching or rash . Frequently an exact date was not available and participants stated that symptoms began e . g . ‘over a year ago’ . The number of cases included suspected cases and hence would include individuals who have been exposed to the mite but not yet infectious , E , and the infectious individuals , I . As a result , the predicted number of cases from the model , comprised of E and I , is then compared to the number of cases reported , C . Formally , we write the data D , as the number of scabies cases {ci} that are observed at time {ti} in a care home with population {ni} where i ∈ {1 , … , 7} , and assume the stochastic SEI model as our generative model for the data . Our likelihood , assuming that carehomes are independent , therefore takes the form L ( { c i } | { n i } , { t i } ; α , β , γ ) = ∏ i = 1 7 P [ E ( t i ) + I ( t i ) = c i | N = n i ; α , β , γ ] ( 11 ) where P [ E ( t i ) + I ( t i ) ] = P s , e , i ( t i ; α , β , γ ) which is obtained by solving Eq 6 . Our MCMC procedure was Random-Walk Metropolis Hastings with hand-tuned Gaussian proposals , which was used to obtain samples from the posterior distribution of the model parameters . We ran 16 MCMC chains in parallel each of length 2 . 5 × 104 , burn-in time for each chain was 104 and samples were thinned by a factor of 20 . Mixing was assessed using trace plots and the total number of samples visualised is 1 . 2 × 104 . We consider here two costs associated with a scabies outbreak with the first case at time 0 and with an intervention starting at time t . Barkwell and Shields [1] reported 172 deaths in a population of size 210 over a 36 month period , and 15 subsequent deaths over 6 months in a sub-population of 47 who had received ivermectin treatment , as well as 10 deaths in the remaining population of 163 over that 6 month period . They reported deaths for each month in the two sub-populations over the six months following ivermectin treatment . Barkwell and Shields performed two statistical tests on these data: chi-squared and Fisher’s exact . Of these , Fisher’s exact test is more accurate for small populations and answers the following question: if two groups , one of size 163 and one of size 47 , are formed by picking individuals from the total population of 210 ( with 25 deaths ) uniformly at random , then what is the probability p of the pattern of deaths observed , or one with more deaths in the population of size 47 . This test gives p < 0 . 0001 when applied to the data . This work received criticism from a more standard biostatistical and epidemiological perspective—in particular due to the absence of control for illnesses other than scabies—shortly after its publication [56 , 57] . Here we do not comment on these issues , but rather focus on the extent to which more general heterogeneity between individuals can invalidate methods such as Fisher’s exact test . Mathematically , we model heterogeneity by assuming that the mortality rates in the population are variable , and that in particular the probability of k deaths in a unit of size n over a time period t are given by a Poisson mixture L ( k | μ , θ ; n , t ) = ∫ λ = 0 ∞ Poisson ( k | n t λ ) Gamma ( λ | ( μ / θ ) , θ ) d λ = ( n t θ ) k ( 1 + n t θ ) - k - ( μ / θ ) Γ ( k + ( μ / θ ) ) k ! Γ ( μ / θ ) . ( 14 ) Here μ is the mean death rate in the population and θ is the variance divided by the mean , which we call the overdispersion . When θ → 0 , we recover the situation where death rates are homogeneous , and larger values of θ imply more heterogeneity . We will carry out two analyses of the Barkwell and Shields data .
We have implemented the exponential integrators discussed above in Matlab [58] using the codes provided by the respective authors . All arising linear systems have been solved using Matlab’s backslash operator ( ∖ ) . The time integrations have been performed over the same interval [0 , tmax] , where tmax = 360 days . The matrix Q corresponds to a Markovian household model with three epidemiological compartments , i . e . , S , E and I ( test results in the main paper ) , and to a more complicated model , Fig A in S1 Text , representing complex transmission dynamics of a multi-strain infection in the population ( test results in the S1 Text ) . Our computations were carried out on a 64-bit desktop computer running Ubuntu 14 . 04 LTS with 32GB RAM and 16 processors each capable of 3 . 30GHz . To compare the error of our algorithms’ output at the final time point with the reference solution , we have chosen to use the relative infinite norm ( ℓ∞-norm ) , i . e . , the infinite norm of the difference between the approximate and the reference solution at the final time point divided by the infinite norm of the reference solution . The result of the reference simulation was obtained by solving the ODE system using Matlab’s function expm ( ) which gives a machine precision estimate of the matrix exponential . We performed the benchmarking by assuming a population sub-unit with a small ( N = 10 ) , medium ( N = 30 ) and large ( N = 99 ) number of individuals with corresponding Q matrices of size 66 , 496 and 5 , 050 respectively for the SEI model . For the more complex multi-strain model , the matrix Q is of size 120 and 11 , 440 corresponding to a household with 3 and 8 individuals respectively . We run each algorithm with 100 replicates and report the mean for the computational time against the error , i . e . , the relative ℓ∞-norm . The results for numerical performance for the SEI model with the small Q matrix of size 66 × 66 are presented in Fig 1 with the x-axis showing the log time in seconds and the y-axis the error as measured by log of relative ℓ∞-norm . From the figure , we can observe that the accuracy of DA1 , DA2 and DA3 increases with an increase in the step size in what seems to be linear in log scale . The higher the degree to which the polynomial is expanded , the greater the accuracy . However , for comparable accuracy , DA1 takes almost 2 orders of magnitude more time than DA3 . This is because DA1 needs to be evaluated over a much smaller step size in order to achieve the same accuracy as DA3 which consequently increases the computational time . The RK4 approximation appears to be relatively less accurate than all of the other methods for large step sizes but the accuracy increases with a reduction in the step size . The computational time does not appear to be greatly influenced by varying the tolerance for Chebyshev expansion although it does take more time for a lesser accuracy in the solution vector compared to DA1 , 2 , 3 , RATKRYL and RK4 . RATKRYL has the advantage of a steep increase in the accuracy of the solution for very minimal increase in computational time . ODE45 is the slowest of all the methods followed by EXPM2010 and EXPOKIT . However , for a small matrix , EXPOKIT , results in the most accurate solution . These results are however dependent on the size of the system . For a medium sized matrix of size 496 , all the algorithms take more computational time with the accuracy of CHEBY and EXPOKIT decreasing significantly , see Fig 1 . For a large matrix , RATKRYL takes the shortest time albeit at the expense of reduced accuracy . However , the error can be further decreased to the desired level by reducing the relative tolerance . Fig B and C in S1 Text . Complex transmission dynamics model show comparable results when the integrators are applied to a more complex multi-strain model . The outcome is consistent with that of the simple SEI model considered here and the efficiency of the methods depends on the system size in both cases . Since the size of the care homes in our data range between 18 and 99 , leading to matrix sizes of between 190 and 4 , 371 , we opt to use the RATKRYL method in the following model fitting computations . This method seems to strike a good trade-off between the accuracy and the computational time for both small and large system sizes . Since it does not guarantee a probability vector , we implement the projection mapping discussed above , guaranteeing probability vectors at an almost negligible computational cost . The error tolerance used for the model fitting is taken to be 10−3 . Using RATKRYL method , we solve the SEI scabies model and fit it to data collected from 7 carehomes in the UK [28] . We used MATLAB’s kdensity , to produce kernel posterior predictive densities , which gives a smooth probability density function from a finite sample of the random variables β , α and γ . The contour plots in Fig 2 show the joint posteriors with the histograms showing the marginal posterior densities for the three parameters β , α and γ with the black solid lines showing the prior distributions . The blue dashed lines , in the sub-plots of the first column of Figs 3 and 4 , show the mean of the posterior samples representing how well the model fits the data ( blue circles ) from the seven care homes ( row A to G ) . The black solid lines ( mean ) and the grey dashed lines ( 95% CI ) , in the first column of Figs 3 and 4 , represent the model predictions of the number of scabies cases in the presence of treatment with permethrin and ivermectin respectively . Treatment is implemented immediately when one or more cases have been detected in a care home . In all care homes , we can observe that treatment with permethrin leads to a total eradication of scabies with the exception of care home D in which case a late observation of scabies cases occured . On the other hand , treatment with a single dose of oral ivermectin , first column Fig 4 , does not lead to eradication except in carehome G but leads to a reduction in the number of cases that later rebound and saturate at long time . In the second column sub-plots of Figs 3 and 4 treatment with permethrin and ivermectin is seen to lead to a reduction in the QALY cost , computed as the cumulative person-days of symptomatic infection during the epidemic , with permethrin leading to a greater reduction with the exception of care home D . Fig 5A shows the full posterior we obtain for μ and θ . Given the level of uncertainty for such a small dataset , we perform an analysis for given fixed overdispersal θ ( Fig 5B , 5D and 5F ) which indicates that as θ is increased , the uncertainty in μ also increases and that the probabilities of a given number of deaths in 6 months in a population of size 47 as defined in ( 16 ) also become more spread out , which we assume is the salient fact to explain in the Barkwell and Shields data . In particular , the probability that κ ≥ 15 is 0 . 0033 for overdispersal 0 , which is consistent with the very low p-values reported by Barkwell and Shields . As we increase the overdispersal , we obtain probabilities that κ ≥ 15 of 0 . 20 for overdispersal θ = 0 . 01 , 0 . 31 for overdispersal θ = 0 . 02 , 0 . 52 for overdispersal θ = 0 . 1 and 0 . 58 for overdispersal θ = 1 . Therefore , the mortality pattern reported by Barkwell and Shields is not a particularly low probability event . These conclusions are confirmed by our analysis over the full posterior including overdispersal ( Fig 5C , 5E and 5G ) .
Expansion and projection based methods to compute the exponential of a matrix were used to compare how well they perform when applied to a Markovian SEI household model describing the transmission of scabies . The computational accuracy and efficiency were tested by performing various numerical tests by changing the step size h and the system size . In order to test the accuracy , the reference solution was obtained by EXPM and computing the relative ℓ∞-norm . The methods indicate that for large system sizes , the RATKRYL method was superior both in terms of computational time and accuracy . The DA methods would be appropriate when a fast solution is needed without the need for strict accuracy . Otherwise , for strict accuracy , obtained by reducing the time step , there seems to be a linear increase ( in log scale ) in time . DA1 has the inherent benefit of ensuring the solution is a probability vector negating the need for correction which might come with some computational time saving and a reduction on the error . In instances where the solution is required at multiple time points , the expansion based methods can deal with this by applying them repeatedly using the last obtained approximation; the accuracy of which is dependent on the step size . However , the projection based methods have the advantage that the projection space is chosen independent of t and hence no time stepping is required . To run the Scabies model , we chose RATKRYL as it runs fast for different choices of tolerance over a wide range of matrix sizes . The efficiency of the method allows us to perform efficient Bayesian inference on limited scabies data , allowing us to incorporate prior information and quantify remaining uncertainty in a consistent manner . A proper choice for the best method will however depend upon the details of implementation and upon the particular problem being solved . We have also demonstrated , in the S1 Text . Complex transmission dynamics model , that the results obtained from the SEI model are fairly consistent with those from a more complex multi-strain disease model . The overarching observation is that the outcome is predominantly dependent on the system size which is a function of both the number of disease states in the model and the number of individuals in the household . The appetite for computationally efficient algorithms for solving these type of models has been well documented . In a recent study exploring the use of Bayesian design of experiments in designing epidemiological studies that collect time resolved longitudinal data of infections in households , noted that the failure to adopt such methods stem from the absence of sufficiently efficient methods for computing the likelihood [10] . The ability to reduce problems in household epidemic theory to a set of linear differential equations [2 , 37 , 59] , on which this work builds , can take advantage of the numerical advances in this work with the ability to achieve the modelling ideal of sufficiently accounting for uncertainty in both structural and parameter values and the ability to propagate it forward within a modelling framework . By fitting the scabies model to data , we estimate that the mean of the posterior distributions are β = 0 . 0053 , α = 0 . 68 , and 1/γ−1 = 5 . 4 days with the 95% marginal credible intervals presented in Table 3 . The economic cost as well as the QALY cost increase with time in the absence of interventions , with the former growing linearly and the latter saturating at long time . Our results are relevant to understanding and mitigating the spread of scabies in carehomes in two main ways . First , early detection of the index case is critical in establishing who and when to treat . This can be achieved either by frequently screening for scabies in care homes and a mandatory screening for new residents and staff . This would ensure that the infection is not spread to other carehome residents and staff who act as a frequent link between the carehome and the outside world . We note that there has been reported difficulties in diagnosing scabies , even for trained personnel , due to confounding skin conditions e . g eczema , atypical presentation and lack of a specific diagnostic criteria [28 , 35] . Comprehensive and timely treatment and identification of all cases , both symptomatic and asymptomatic , would help prevent the spread of the disease . From Fig 4 care home G , it is clear that the earlier the intervention is implemented the more QALYs gained . However a delay in administering ivermectin would lead to a rebound of cases albeit with a high degree of uncertainty . Treatment with permethrin leads to a total eradication of scabies but with the drawback that it requires application to both staff and residents and to be left on the body for 8 hours before being washed , which can be distressing and logistically challenging leading to sub-optimal adherence . However , a case can be made for a second and/or third dose with ivermectin in order to achieve better efficacy leading to eradication , which would be easy to do as it is administered orally . Secondly , the degree of uncertainty in the model predictions can be attributed to the lack of sufficient data to parameterize the model . However we are confident that the technique we have adopted , of drawing samples from the posterior distribution and then propagating forward the uncertainty in a simulation model , is the best framework to adopt for studies with limited data . This has also been made possible by the computational efficiency we have achieved in solving the system of ODEs . Although the model would be considered simple , it is possible to make modifications to it in order to accommodate other complex natural histories with minimal or no change to the numerical solving algorithms . This work can therefore be seen as a foundation on which more complex household based models can be built in order to help make public health decisions . The human and economic costs of delayed interventions provide a motivation for the use of all available treatments—here , we have suggested that heterogeneity in death rates can make conclusions about mortality from ivermectin hard to draw on the basis of small population studies such as that of Barkwell and Shields; in particular , we strongly advise against over-interpretation of very small p-values that can arise as ‘highly significant’ [60] as a specific instance of known problems with the use of p-values in general [61] . One of the drawbacks to this is that there does not exist a good and standardized diagnostic test to determine if a person is infected . A study from Japan [62] noted that the time it took for a patient to be diagnosed with scabies was 141 days with a range of between 34 and 313 days despite hospital personnel checking for the condition on the patients skin . In another study [28] which involved a semi-structured survey of managers and affected residents of care homes indicated that most outbreaks were attributable to late diagnosis of the index case . There is therefore a need for improved support to hospital staff , care workers and authorities managing these outbreaks . Any efforts that are directed towards early and correct diagnosis would be beneficial in curbing the spread of scabies . The development of a point-of-care diagnostic test would be a major contribution towards this objective . As there are long term health conditions associated with scabies [63] , failed or wrong diagnosis has extra health economic implications and especially in the context of health institutions , outbreaks , and large epidemics or pandemics . Another important , but maybe slightly understated requirement is the need for a harmonised and carefully coordinated response to scabies outbreaks . There is no national guidance policy for the public health management of scabies in the UK and it is usually left to the individual Public Health England Health Protection Teams to make decisions [34] . This may lead to sub-optimal test and treat strategies that retain some level of background transmission in the community which potentially make control or elimination difficult . The design of such a national level guideline would benefit from mathematical modelling and the model developed in this work is designed to be a platform on which we can build more sophisticated models to support national-level decision making .
|
Our work makes five main contributions . ( 1 ) We study a previously under-modelled scenario—transmission of scabies in residential care homes—that is of significant and growing public health importance in the context of an ageing population . ( 2 ) We develop a Markov-chain-based modelling framework that accurately captures the disease dynamics in well-connected sub-units such as care homes , but whose use has previously been limited due to computational cost . ( 3 ) We demonstrate that appropriate numerical methods ( in particular rational Krylov approaches ) for the solution of the mechanistic model for transmission of scabies in care homes speeds up evaluation by several orders of magnitude compared to other methods . ( 4 ) We demonstrate a Bayesian approach in which the model is fitted to data using computationally-intensive MCMC methods , validated using posterior predictive intervals , and has its uncertainty quantified in forward predictions . ( 5 ) We revisit the question of safety of ivermectin using appropriate methods and demonstrate that relaxation of the assumption of homogeneous death rates can make previous influential conclusions on lack of safety unsound .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"death",
"rates",
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"markov",
"models",
"dose",
"prediction",
"methods",
"applied",
"mathematics",
"tropical",
"diseases",
"social",
"sciences",
"parasitic",
"diseases",
"animals",
"health",
"care",
"simulation",
"and",
"modeling",
"algorithms",
"mathematics",
"ectoparasitic",
"infections",
"pharmaceutics",
"sexually",
"transmitted",
"diseases",
"algebra",
"neglected",
"tropical",
"diseases",
"population",
"biology",
"polynomials",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"health",
"economics",
"scabies",
"mites",
"economics",
"probability",
"theory",
"arthropoda",
"population",
"metrics",
"eukaryota",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"organisms"
] |
2018
|
Scabies in residential care homes: Modelling, inference and interventions for well-connected population sub-units
|
In many sensory systems , transmembrane receptors are spatially organized in large clusters . Such arrangement may facilitate signal amplification and the integration of multiple stimuli . However , this organization likely also affects the kinetics of signaling since the cytoplasmic enzymes that modulate the activity of the receptors must localize to the cluster prior to receptor modification . Here we examine how these spatial considerations shape signaling dynamics at rest and in response to stimuli . As a model system , we use the chemotaxis pathway of Escherichia coli , a canonical system for the study of how organisms sense , respond , and adapt to environmental stimuli . In bacterial chemotaxis , adaptation is mediated by two enzymes that localize to the clustered receptors and modulate their activity through methylation-demethylation . Using a novel stochastic simulation , we show that distributive receptor methylation is necessary for successful adaptation to stimulus and also leads to large fluctuations in receptor activity in the steady state . These fluctuations arise from noise in the number of localized enzymes combined with saturated modification kinetics between the localized enzymes and the receptor substrate . An analytical model explains how saturated enzyme kinetics and large fluctuations can coexist with an adapted state robust to variation in the expression levels of the pathway constituents , a key requirement to ensure the functionality of individual cells within a population . This contrasts with the well-mixed covalent modification system studied by Goldbeter and Koshland in which mean activity becomes ultrasensitive to protein abundances when the enzymes operate at saturation . Large fluctuations in receptor activity have been quantified experimentally and may benefit the cell by enhancing its ability to explore empty environments and track shallow nutrient gradients . Here we clarify the mechanistic relationship of these large fluctuations to well-studied aspects of the chemotaxis system , precise adaptation and functional robustness .
High-resolution microscopy has revealed the exquisite spatial organization of signaling pathways and their molecular constituents . Understanding the computations performed by biological networks therefore requires taking the spatiotemporal organization of the reactants into account [1] . One feature common to many signal transduction pathways is the clustering of receptors in the cell membrane . This arrangement has been observed for diverse receptor types [2] such as bacterial chemoreceptors [3]–[6] , epidermal growth factor receptors [7] , and T cell antigen receptors [8] . Receptor clustering provides a mechanism for controlling the sensitivity [9] , [10] and accuracy [11] , [12] of a signaling pathway . Moreover , by controlling which types of receptors participate in clusters a cell can achieve spatiotemporal control over the specificity of the signaling complexes . While clustering receptors can tune the sensitivity and specificity of a signaling pathway , organizing receptors into clusters also imposes novel constraints on the kinetics of the pathway . Temporal modulations of the activity of signaling complexes , such as adaptation , are typically achieved via posttranslational modification of the cytoplasmic tail of the receptors by various enzymes . The localization of the receptor substrate into clusters implies that trafficking of enzymes between the cytoplasm and the cluster and between receptors within a cluster is likely to be an important determinant of the dynamics of such modulations . Recent theoretical studies of the effect of the localization of enzymes and substrates on signaling kinetics have shown that spatiotemporal correlations between reactants can significantly affect the signaling properties of these pathways [13]–[15] . One well-characterized system in which the spatial organization of receptors plays a significant role is the chemotaxis system of the bacterium Escherichia coli [16]–[18] . E . coli moves by performing a random walk alternating relatively straight runs with sudden changes of direction called tumbles . The probability to tumble is modulated by a two-component system in which transmembrane receptors regulate the activity of a histidine kinase CheA , which in turn phosphorylates the response regulator CheY . Phosphorylated CheY rapidly diffuses through the cell and binds the flagellar motors to induce tumbling . The tumbling rate decreases in response to chemical attractants and increases in response to repellants , allowing the bacterium to navigate its environment . Chemoreceptor clustering affects both signal amplification and adaptation to persistent stimuli , which together enable bacteria to remain sensitive to over five orders of magnitude of ligand concentration [19] . Signal amplification arises from allosteric interactions between clustered receptors [9] , [20]–[23] whereas adaptation is mediated by the activity of two enzymes: CheR methylates inactive receptors , thereby reactivating them , while CheB demethylates active receptors , deactivating them . This arrangement implements an integral feedback mechanism [24] , enabling kinase activity and therefore cell behavior to return to approximately the same stationary point following response to stimulus [25] , [26] . The localization of enzymes to the cluster is facilitated by a high-affinity tether site present on most receptors . This tether , together with the dense organization of the cluster , enables localized enzymes to modify multiple receptors within a range known as an assistance neighborhood [27] . Modeling efforts have shown that assistance neighborhoods are required for precise adaptation when receptors are strongly coupled [28] . Precise adaptation , however , is not by itself sufficient for successful chemotaxis . The dynamics of the adaptation process , including the rate of receptor modification and the level of spontaneous fluctuation in receptor activity , are also critical determinants of chemotactic performance [29]–[35] . Recent measurements of the dynamic localization of chemotaxis proteins have shown that the time scale of CheR and CheB localization to the receptor cluster is comparable to the time scale of adaptation [36] and therefore expected to affect the dynamics significantly . Moreover , dense clustering may enable localized enzymes to perform a random walk over the receptor lattice without returning to the cytoplasmic bulk , a proposed process termed brachiation [37] that may lead to more efficient receptor modification . Here we analyze how the spatiotemporal localization of the adaptation enzymes to the receptor cluster affects the dynamics of the adaptation process . First we build a stochastic simulation of the chemotaxis system taking into account the organization of the receptors into large clusters [4] , [6] , the slow exchange of enzymes between the cytoplasm and the clusters [36] , enzyme brachiation [37] , and assistance neighborhoods [27] , [28] , [38] . This model quantitatively recapitulates experimental observations of the magnitude of the spontaneous fluctuations in single cells [39]–[42] and the kinetics of adaptation averaged over multiple cells [43] . Notably , while localized enzymes in this model operate at saturation , the output of the system nonetheless remains robust to cell-to-cell variation in enzyme expression levels [44] , in contrast to the covalent modification system studied by Goldbeter and Koshland [12] . We therefore resolve the question of how large spontaneous fluctuations might coexist with a robust mean output in the system [30] . We interpret these results in the second part of the paper , using a mean-field analytical model to examine the molecular mechanisms underlying these features and their relation to receptor clustering .
We used the rule-based simulation tool NFsim [45] to create a stochastic model of the bacterial chemotaxis system that accounts for the organization of chemoreceptors into a large , dense , hexagonal lattice [4] . Like the Gillespie algorithm , NFsim computes exact stochastic trajectories , but avoids the full enumeration of the reaction network , which can undergo combinatorial explosion , by using rules to generate reaction events [45] . In the simulation , each chemoreceptor dimer is represented by an object with one tether site , one modification site , and a methylation level ranging from 0 to 8 . We model a single contiguous lattice consisting typically of 7200 dimers , although we consider different sizes as well . The structure of the lattice is fully specified by enumerating for each dimer its six nearest neighboring dimers . Receptor cooperativity is modeled using Monod-Wyman-Changeux ( MWC ) complexes consisting of six receptor dimers ( Fig . 1A ) . The activity a of each signaling complex depends on the methylation and ligand-binding state of the dimers in the complex and is calculated from Eq . ( 13 ) ( Methods ) as previously described [23] , [28] . The implementation of this model in NFsim is discussed in the Supporting Text S1 . Receptor modification occurs through the enzymes CheR and CheB , which are each modeled as having two binding sites , one specific to the receptor tether and one specific to the modification site . In the model , CheR and CheB dynamically bind and unbind both of these sites . CheR participates in the reactions illustrated in Fig . 1B . The possible states of the enzyme are: free and dispersed in the cytoplasmic bulk , or bound to one or both of the tether and modification sites . Enzymes in the bulk localize to the cluster by binding either the tether site or the modification site directly . The time scales of these binding reactions ( Fig . 1B , blue arrows ) are the slowest in the present model: ∼15 s for localization through tether binding , as measured [36] , and longer for modification site binding , reflecting the lower affinity of enzymes for the modification site . Once bound to the tether or modification site , an enzyme may bind the modification site or tether , respectively , of the receptor to which it is already bound ( Fig . 1B , red arrows ) or any of its six nearest neighbors ( green arrows ) . Therefore the assistance neighborhood consists of seven dimers , consistent with measurements [27] . Assistance neighborhoods are unique for each receptor dimer and therefore overlap . Accordingly , in the simulation individual receptor dimers participate in multiple assistance neighborhoods . Since these reactions are confined to small volumes ( given by the ∼5 nm tether radius [46] ) , they proceed at high rates ( 1–10 ms time scales; see Text S1 ) . The activity-dependent binding rate of CheR to the modification site is proportional to 1 - a , while the rates of all other CheR reactions are taken to be independent of activity . Phosphorylated CheB ( CheB-P ) participates in completely analogous reactions except that the rate of binding the modification site is proportional to a . CheB phosphorylation proceeds at a rate proportional to the activity of the receptor cluster ( Text S1 ) . For simplicity we assume that only CheB-P can localize to the receptor cluster since its affinity for the tether is much higher than that of CheB [47] . Our study is the first to incorporate enzyme brachiation [37] , assistance neighborhoods [28] , [38] , cooperative amplification of the input signal [9] , [22] , [23] , activity-dependent adaptation kinetics [25] , and a large contiguous receptor cluster into a single model . This model specifically extends two earlier models . The first of these models considered enzyme brachiation on a large receptor cluster [37] , but did not include activity-dependent kinetics , receptor cooperativity , or any modification of the receptors . The second of these models included activity-dependent kinetics , cooperativity , and assistance neighborhoods [28] , [38] but excluded enzyme brachiation and limited the system size to a single MWC complex consisting of 19 dimers . Here we take advantage of the flexibility and efficiency of NFsim to examine how all of these processes together determine the dynamics of adaptation . Calibration of the model parameters is discussed in the Supporting Text S1 . Supporting Tables S1 and S2 present the full set of simulation parameters . We note that our model includes only Tar receptors . This choice enabled us to compare our model directly to measurements of the adaptation kinetics [43] performed on cells lacking receptors other than Tar . These measurements were obtained by exposing cells to time-dependent exponential ramps of methyl-aspartate , a protocol that we modeled in silico ( Fig . 2A and Fig . S2 ) to verify the calibration of the kinetics of our model . In the remainder of the paper we denote this calibrated model as the reference model M1 . Together with the dense organization of the receptor lattice , the presence of the tether site on each receptor gives rise to assistance neighborhoods [27] and possibly enzyme brachiation [37] . During the brachiation process , enzymes successively bind and unbind the tethers and modification sites on different , neighboring receptors , enabling them to perform a random walk over the lattice without returning to the bulk . Both assistance neighborhoods and enzyme brachiation should increase the distributivity of the methylation process , meaning that sequential ( de ) methylation events catalyzed by a single enzyme will tend to take place on different receptors . In a distributive scheme , therefore , an enzyme will tend to modify multiple receptors during its residence time on the cluster . Moreover , it will tend to methylate receptors in an even fashion , rather than sequentially modifying a single receptor until it is fully ( de ) methylated . Since brachiation enables some randomization of enzyme position between methylation events , it should lead to a more distributive methylation process . To investigate how distributivity affects adaptation we compared our reference model M1 , which includes assistance neighborhoods and brachiation , to a model in which the binding of tethered enzymes to the modification sites of neighboring receptors ( and modification site-bound enzymes to neighboring tethers ) is not allowed , denoted M2 ( Table 1 ) . Disabling these reactions both removes assistance neighborhoods and prevents enzyme brachiation . As a result , methylation is more processive . In this scheme , an enzyme remains bound to and modifies only a single receptor during its residence time in the cluster . This scheme increases the probability that CheR and CheB will become bound to receptors with high or low methylation levels , respectively . Consequently , enzymes will tend to have low affinity for their local modification sites and modification will proceed in an inefficient manner compared to a distributive scheme . In M2 , adaptation to both small ( 5 µM ) and large ( 1 mM ) steps of the attractant methyl-aspartate becomes much slower ( Fig . 2B , light gray ) than in the reference model M1 ( Fig . 2B , black ) . Precise adaptation is also severely compromised for the large stimulus . We also consider the case in which enzyme brachiation is made less efficient , but adaptational assistance is not eliminated . To examine this intermediate model ( M3 ) , we decreased the unbinding rates from the tether relative to M1 . As a result , more methylation events occur before an enzyme moves on the lattice . This leads to less efficient brachiation than in M1 but preserves assistance neighborhoods . As a result , adaptation to the large stimulus is less precise compared to M1 but more precise than M2 ( Fig . 2B ) . The picture that emerges is that the distributivity of the modification process is an important determinant of the precision of adaptation . Adaptational assistance and enzyme brachiation increase the distributivity of modification and lead to more precise adaptation in our model of the full receptor lattice . This result extends previous findings that the ability of tethered CheR and CheB to modify several receptors within an assistance neighborhood is necessary for precise adaptation within a single MWC complex [28] , [38] . In our simulations , as in these previous studies , increasing the distributivity of receptor methylation reduces the time CheR and CheB spend bound to highly methylated and demethylated receptors , respectively . Consequently , the methylation rate in distributive models is largely independent of the methylation levels of individual receptors , resulting in more precise adaptation . Additionally , ( de ) methylation rates are higher than in the more processive schemes because the enzymes spend less time interacting with receptors that are already highly methylated or demethylated . Indeed , plotting the rate of methylation after the step stimulus for the three simulations depicted in Fig . 2B ( bottom panel ) indicates that it is highest in the most distributive model M1 ( Fig . S7 and Text S1 ) . Experiments and modeling efforts strongly suggest that the adaptation mechanism of the bacterial chemotaxis system introduces slow spontaneous fluctuations in the activity of the receptor-kinase complex with a standard deviation of ∼5–10% of the mean [33] , [39]–[42] , [48] , [49] . These fluctuations are thought to lead to long-tailed distributions of run durations [39] , [50] and may enhance navigation in shallow gradients and exploration [30] , [32] , [33] , [35] , [39] . Since distributivity affects the kinetics of adaptation , it is also likely to affect the spontaneous fluctuations of the system . Fig . 2C compares the level of fluctuation in receptor activity about the unstimulated steady-state level for each model at different expression levels of CheR . The model M1 exhibits fluctuations of the same order as those measured experimentally , particularly at low CheR levels for which the standard deviation σa of fluctuations exceeds 7% of the mean activity a0 . Notably , the magnitude of this noise is reduced when receptor modification is made less distributive in models M2 and M3 . These results suggest that the features required for successful adaptation , assistance neighborhoods and brachiation , also lead to experimentally observed levels of signaling noise . The mechanism underlying these relations will be discussed in a later section with insights provided by an analytical model . Cells within an isogenic wild-type population are known to exhibit significant cell-to-cell variability in the level of signaling noise [33] , [39]–[41] . To what extent does this variability arise from cell-to-cell variability in the expression levels of the chemotaxis proteins ? Our simulations of the model M1 indicate that the level of signaling noise is sensitive to the relative amounts of CheR and CheB in the cell ( Fig . 2C ) . However , the multicistronic organization of cheR and cheB on the chromosome ensures that the ratio of CheR to CheB is approximately conserved in each cell within a wild-type population due to cotranscription [44] , [51] . Therefore variability in signaling noise levels must arise largely from correlated variation in the expression levels of the chemotaxis proteins . Using our stochastic simulation of enzyme dynamics on the receptor lattice ( M1 ) , we investigated the effects of covarying the number of CheR , CheB and chemoreceptors . We sampled cells from across a population in which CheR , CheB and chemoreceptor counts all vary according to a log-normal distribution ( Fig . S5 ) obtained from measurements of CheY-YFP levels expressed from the native chromosomal locus [44] . Mean protein expression levels were set according to immunoblotting measurements [52] . To study only the effects of concerted variation in protein levels , we ignored intrinsic noise , thereby preserving the ratio of CheR/CheB/receptors . We found that the level of signaling noise varies widely between each sampled cell , between 3 and 10% of the mean ( Fig . 3A ) . This degree of variation in signaling noise levels agrees well with measurements performed across a wild-type population [40] , [41] . Additionally , we found that cells with low expression levels of the chemotaxis proteins are predicted to exhibit the large fluctuations , ∼10% of the mean level . Consequently , we expect cells with high levels of signaling noise to be present even in populations across which the CheR to CheB ratio is maintained at the single cell level . In previous models of the chemotaxis system in which enzyme localization is not considered , the slow , spontaneous fluctuations in the activity of the system were traced back to the ultrasensitive nature of the methylation and demethylation reactions , which were assumed to operate near saturation [30] . This mechanism , however , is insufficient to explain the large magnitude of the noise observed experimentally in individual cells . Indeed , using a stochastic simulation of a recent representative analytical model ( Model B1 ) in which the authors calibrated the rates of methylation-demethylation using direct measurements of the average response of the receptor activity to ramps of attractant [43] , we observe at most 2–3% relative noise for the individual cell ( Fig . 3B ) . The model B1 incorporates activity-dependent binding of the enzymes to the modification sites , but does not consider any aspects of enzyme localization via tether binding ( Table 1 ) . Additionally , while this model includes cooperative receptor-receptor interactions using a MWC model , given by Eq . ( 13 ) as for M1 , it considers neither the geometry of the receptor cluster nor the resulting features of adaptational assistance and enzyme brachiation . Higher noise levels can be obtained in this model by increasing the enzyme-substrate affinities tenfold ( Model B2 ) . These higher affinities , however , result in a steady-state activity that is ultrasensitive to total enzyme counts ( Fig . 3C , light gray ) . In this case the addition or subtraction of only a few adaptation enzymes in the cell is sufficient to switch the system between the fully active and fully inactive states . This scenario is biologically unacceptable since small fluctuations in gene expression across a population would lead to large numbers of non-functional cells with either fully active or inactive receptors at steady state . Parameter values for models B1 and B2 are given in Tables S4 and S6 . Interestingly , in our model accounting for the localization of enzymes to the receptor cluster , large fluctuations around the steady state activity are present even though the mean activity remains relatively robust to changes in enzyme counts . Fig . 3B shows the dependence of the steady-state fluctuations in M1 on total CheR count with all other parameters fixed . M1 exhibits activity fluctuations that exceed 7% of the mean value a0 for low CheR counts and are significantly larger than those of the model B1 for all CheR values . While the noise level is high , the mean receptor activity at steady state , a0 , is only modestly sensitive to changes in the total CheR count ( Fig . 3C , black ) . The specific features enabling the coexistence of large fluctuations with a robust steady state are discussed in a later section with reference to an analytical model . Finally , we compare the noise levels predicted by the models M1 and B1 across a cell population . When cell-to-cell variability in receptor and enzyme counts is taken into account we observe that B1 , which does not account for receptor clustering or enzyme localization , exhibits insufficiently large fluctuations ( σa/a0<4% ) across the entirety of the population ( Fig . 3A ) . In contrast , M1 exhibits levels of noise similar to those measured experimentally [33] , [40] , [41] , as discussed in the previous section . To investigate the mechanisms underlying our numerical results , we constructed an approximate model that can be solved analytically . Here we provide a mathematical derivation of the model . Analysis of the adaptation mechanism using this model is provided in the next section . At the heart of this model is a covalent modification scheme that describes the kinetics of receptor methylation by CheR and CheB , similar in form to previous models [12] , [25] , [30] , [53] , [54] . In order to modify the receptors , however , we require that CheR and CheB be localized to the receptor cluster by being bound to the tether site . In this treatment , CheR may exist in three states: free and dispersed in the cytoplasmic bulk ( R ) , bound only to the tethering site of a receptor ( R* ) , and bound to both the tether site and modification site of receptors ( ) . The notation for the states ( ) of phosphorylated CheB is analogous . Unphosphorylated CheB is assumed not to interact with the receptors and therefore only exists in the bulk ( B ) . For simplicity , we assume that enzymes in the bulk always bind the higher-affinity tether sites on the receptors prior to binding the modification sites . Since the model includes reactions occurring in multiple volumes and will later be used for stochastic calculations , all molecular species below are quantified by number rather than concentration . Therefore , the binding rates as written implicitly include a factor of the inverse of the reaction volume . In the model , active receptor complexes phosphorylate CheB at a rate ap and CheB autodephosphorylates at rate dp , leading to , which we take to be in the steady state , yielding . We assume that only bulk CheB ( B , Bp ) participates in the phosphorylation reactions . Defining and as the total number of tether-bound CheR and CheB-P , the dynamics of enzymes in the bulk binding to the tether site is modeled by ( 1 ) ( 2 ) Here denote the rates of cytoplasmic enzymes binding the tether site and denote the rates of enzymes bound only to the tether unbinding the tether and dispersing into the bulk . Since the number of tether sites greatly exceeds the number of CheR and CheB [52] , we assume it to be constant and equal to the total number of receptors TTot . Enzymes bound to the tether may bind the modification site according to ( 3 ) ( 4 ) in which are the rates of a tether-bound enzyme to bind the receptor modification site , are the unbinding rates from the modification site , and ( kr , kb ) are catalytic rates for demethylation and methylation of the modification site , respectively . Binding to the modification site is dependent on the activity of the receptor . Eqs . ( 3 , 4 ) employ a mean-field approximation by assuming that the activity of the receptor whose modification site is to be bound is equal to the mean activity of all receptors in the cell , a . This assumption makes the methylation process in this model fully distributive . Therefore the mean-field model represents the limit of a single , maximally large assistance neighborhood , encompassing all receptors , or infinitely fast brachiation , in which enzymes completely randomize their position on the lattice between methylation events . Relaxing this assumption requires a more detailed analytical model , which is explored in the Supporting Text S1 . Since Eqs . ( 3 , 4 ) describe a binding reaction confined to the ∼5 nm radius defined by the tether [46] , the kinetics are fast relative to other reactions in the model ( Text S1 ) . We take , leading to an expression for the number of enzymes bound to both tethers and modification sites ( 5 ) ( 6 ) Here Kr and Kb are dimensionless constants analogous to Michaelis-Menten constants . The rate of change of the total methylation level M of all MWC complexes in the system ( the total number of methylated receptor sites across all receptors in the cell ) is ( 7 ) Using Eqs . ( 5–7 ) , we write the equation describing changes in average methylation level per 2N-receptor MWC complex , m = M ( 2N/TTot ) , in the form familiar from the Goldbeter-Koshland system [12] , [30] , [54] ( 8 ) The tether-binding reactions Eqs . ( 1 , 2 ) may be rewritten in terms of and as ( 9 ) ( 10 ) with an activity-dependent unbinding step . To include variation around the mean , Langevin sources ( ηm , ηr , ηb ) have been added with magnitudes evaluated using the linear noise approximation ( Text S1 ) [55] , [56] . The instantaneous output of the system is the fraction of active receptors a ( t ) = a[m ( t ) , L ( t ) ] with a given by a MWC model , Eq . ( 13 ) , for some external stimulus L ( t ) ( Methods ) [22] , [23] , [43] . The noise statistics of the output a ( t ) at steady state are calculated by linearizing the model and solving it as a multivariate Ornstein-Uhlenbeck process ( Methods and Text S1 ) [57] , [58] . Parameter values for the analytical model ( Tables S1 and S4 ) were taken to be consistent with those of the stochastic simulation M1 . Two important features can be noted from the form of Eqs . ( 8–10 ) . First , Eqs . ( 9 ) and ( 10 ) emphasize that unbinding from the receptor lattice is a two-step process . Since CheR has higher affinity for the modification site as activity decreases , the overall rate of CheR unbinding the lattice and returning to the bulk decreases accordingly . Additionally , a smaller value of Kr , which denotes higher affinity of the localized enzyme for the modification site , leads to slower overall rates of unbinding . The argument for CheB-P unbinding is analogous . Second , since Eq . ( 8 ) depends only on the mean activity of the system and not on methylation or stimulus levels , the analytical model exhibits precise adaptation . This property follows from the mean field assumption or , equivalently , the assumption of fully distributive kinetics . Using this analytical model , we next examine the mechanisms underlying the key observations made using numerical simulations and argue that: ( 1 ) large fluctuations in receptor activity are primarily due to noise in localized enzyme counts amplified by a methylation process ultrasensitive to these counts; ( 2 ) a distributive methylation scheme increases signaling noise by increasing the ultrasensitivity of this process; ( 3 ) the localized enzymes work at saturation without causing the mean activity to be ultrasensitive with respect to total enzyme expression levels . This result contrasts with the covalent modification scheme studied by Goldbeter and Koshland [12] . The analytical model derived above predicts large fluctuations in receptor activity ( Fig . 4A , black ) , similar to those predicted by the stochastic simulation M1 . This level of signaling noise is significantly higher at all CheR levels than the level predicted when enzyme localization is not taken into account ( Fig . 4A , gray; analytical version of model B1 [43] ) . The high level of intracellular signaling noise in this system arises from three key features . First , since the total numbers of CheR and CheB are small [52] , the relative variation in the number of localized enzymes due to Poisson statistics is large . The overall rates of methylation and demethylation are therefore highly variable in time . Second , these fluctuations in localized enzyme counts occur at sufficiently slow time scales [36] to not be filtered out by the methylation process . The possibility of slow fluctuations in the number of tethered enzymes leading to increased fluctuation in receptor activity was previously noted using a model of a single MWC complex [38] . Third , the interaction between the localized enzymes and the substrate occurs at saturation . Since the binding of the localized enzymes to the receptor modification site is activity-dependent , this interaction takes the same form as the covalent modification system studied by Goldbeter and Koshland [12] , as can be seen from Eq . ( 8 ) . Therefore we may analyze the localized enzyme-receptor interaction in the same terms . Since a localized enzyme is confined to the tether radius , the effective local substrate concentration is high and binding to the modification site proceeds at a fast rate . Therefore , Kr , Kb≪1 and , following Goldbeter and Koshland , the steady-state output a0 has ultrasensitive dependence on the ratio of localized CheR to CheB-P ( Fig . 4B , steep curve ) . This steep relationship suggests that the output of the system is in general highly susceptible to changes in the ratio of localized CheR to CheB-P and , consequently , fluctuations in this ratio are the primary source of noise in the output . In the limit in which methylation is fast relative to enzyme localization , dm/dt∼0 , Eq . ( 8 ) yields . In this limit , receptor activity is a function of only the ratio of the localized adaptation enzymes , corresponding to the steep curve of Fig . 4B . Likewise , the variance in receptor activity becomes . Therefore when the catalytic step is fast relative to enzyme localization , fluctuations in the localized enzyme ratio are amplified by exactly this steep curve . This limit case is relevant for understanding the behavior of our analytic and numerical models , in which the rates of enzyme localization are slow relative to all other rates in the system . We may also show that fluctuations in the number of localized CheR and CheB are the dominant noise sources in the system without assuming dm/dt∼0 . To illustrate this point , we use the analytical model to decompose the total variance σaa of the receptor activity into a sum of three terms , each plotted in the inset of Fig . 4B: ( 11 ) fluctuations due to the number of localized CheR , those due to number of localized CheB-P , and fluctuations due to intrinsic variability in the methylation and demethylation rates ( σaa , m ) . Each contribution depends linearly on the intensity of the corresponding noise source ηi in Eqs . ( 8–10 ) , . The magnitude of the third term σaa , m is comparable to the total noise predicted by models without enzyme localization . Fig . 4B ( inset ) shows that the first two terms on the right hand side of Eq . ( 11 ) dominate to the exclusion of the third , confirming that variability in localized CheR and CheB-P is the dominant source of the large fluctuations in receptor activity . This same mechanism underlies the observed large fluctuations in the stochastic simulation of the model M1 , considered previously . Fig . 4C shows mean activity a0 versus the ratio of mean localized CheR to mean localized CheB-P obtained from simulation by varying only the total CheR count . As in the analytical model , this relationship is highly ultrasensitive . To illustrate the dependence between fluctuations in the localized enzyme ratio and fluctuations in receptor activity , the inset of Fig . 4C displays 500 s time traces of receptor activity and the ratio of localized CheR to localized CheB-P taken from simulation . The correlation between the two series is apparent and consistent with activity fluctuations arising from variability in the number of tethered enzymes . In summary , clustering of the receptors leads to a high density of modification sites for the enzymes localized at the cluster . This results in saturated ultrasensitive kinetics of the covalent modification reactions , which strongly amplify the noise due to the slow exchange of enzymes between the cluster and the bulk . In the analytical model , large fluctuations in receptor activity result from the high affinity of localized enzymes for the modification site . Since all receptors in the analytical model are assumed to have the same activity , this affinity is entirely characterized by the small values of the constants Kr and Kb . In the numerical models , in contrast , the binding of enzymes to individual receptor dimers within MWC complexes of varying levels of activity is explicitly simulated . Consequently , the affinity of the enzymes for the modification site depends not just on the values of Kr and Kb ( as derived from the binding , unbinding , and catalytic rates in the simulation ) , but also on the distribution of CheR and CheB within complexes of different activities . If enzymes tend to become localized within regions of the cluster for which they have low binding affinity ( e . g . , CheR within a highly methylated region ) , we expect the ultrasensitive dependence of activity on the ratio of localized enzymes ( Fig . 4C ) to be reduced . This effect may be thought of as increasing the effective values of Kr and Kb . Adaptational assistance and brachiation mitigate this effect to some extent by enabling localized enzymes to sample a number of receptors during their residence time in the cluster . A higher rate of sampling indicates that a given enzyme samples a larger fraction of the cluster between subsequent methylation events and therefore corresponds to more distributive methylation kinetics . A potentially analogous situation has been studied theoretically for a MAP kinase cascade [13] . In this system , slowly diffusing enzymes tended to rebind the same substrate molecule multiple times , leading to a processive modification scheme . Faster diffusion enabled the enzymes to randomize their positions between modification events , corresponding to distributive modification . In the MAP kinase study , faster diffusion led to an ultrasensitive dependence of the output on enzyme levels . Is a similar mechanism at work in the chemoreceptor cluster ? For our numerical models , we quantified the rates at which enzymes sampled different , unique receptors within the cluster and found that this rate was between 4 and 13-fold smaller for the more processive models M2 and M3 than for the reference model M1 ( Table S7 ) . Accordingly , the steady-state activity in the more processive models M2 and M3 is also less dependent on the ratio of localized CheR to CheB-P than in M1 ( Fig . 4D ) . Since this relationship effectively amplifies fluctuations in the ratio of localized enzymes , this decreased steepness leads to lower signaling noise levels in these more processive models , as seen previously ( Fig . 2C ) . For further details regarding the comparison between simulations and the analytical model , see Supporting Text S1 . We conclude that a distributive methylation scheme leads to higher signaling noise levels by increasing the overall affinity of the localized enzymes for the modification site substrate . The mean steady-state activity for the analytical model with enzyme localization is plotted in Fig . 4B as a function of the ratio of both localized and total ( across the entire cell ) adaptation enzymes , and . While the activity is highly ultrasensitive with respect to the localized enzyme ratio , its sensitivity to the total enzyme ratio is significantly less and comparable to the model B1 . Therefore , the mean steady-state activity of the system a0 is robust to changes in the total CheR to CheB ratio caused by noisy gene expression . This result is somewhat surprising because in the classic covalent modification system studied by Goldbeter and Koshland [12] , saturated enzyme-substrate interactions always lead to a steady-state activity that is ultrasensitive to the total CheR to CheB ratio . In Eq . ( 8 ) , which we may analyze in the same manner as the Goldbeter-Koshland system , the sensitivity of the steady-state activity a0 with respect to the ratio of localized CheR to CheB is determined solely by the constants ( Kr , Kb ) that characterize the probability that a localized enzyme will be bound to a modification site . Small values of these constants lead to saturated kinetics and ultrasensitivity of the steady-state activity to the ratio of localized CheR to CheB . Our model differs from the Goldbeter-Koshland system , however , in that in our model these constants only partially determine the sensitivity of a0 to the ratio of total CheR to CheB . The sensitivity of the system to the total enzyme ratio is also determined by the rates at which cytoplasmic enzymes localize to the cluster and at which localized enzymes return to the bulk . Since the rates at which enzymes localize to the cluster are slow [36] , the effective affinities of the enzymes for the modification sites are reduced even though the affinities of enzymes already localized at the cluster are high . The steady-state solutions to Eqs . ( 8–10 ) quantify how the mean steady-state activity depends on the total enzyme counts RTot and BTot . Solving Eqs . ( 9 ) and ( 10 ) for the localized enzyme counts and and inserting the results into Eq . ( 8 ) , we obtain ( 12 ) Eq . ( 12 ) is also of the Goldbeter-Koshland form which indicates that the steepness of the steady-state activity as a function of the total CheR to CheB ratio is determined by the effective inverse affinities and . Values of lead to ultrasensitivity of the steady-state activity with respect to the ratio RTot/BTot . For the steady-state activity to be considered robust , we require . From this condition , we can see that the steady-state a0 can be robust even if the affinity of the localized enzymes for the modification site is extremely high , ( Kr , Kb ) ≪1 . This will be the case if the rates of enzymes in the bulk to reach the cluster and bind the tether are sufficiently small relative to the unbinding rates , effectively compensating for the small ( Kr , Kb ) and leading to . To discuss the robustness of the bacterial chemotaxis system , we note three key considerations . First , we estimate that Kr , Kb≪1 due to the fast rate of the highly localized enzymes binding the modification site ( Text S1 ) . Second , we note that the CheB-P feedback loop is not by itself sufficient to make the steady-state robust to the total enzyme ratio . While the term due to the feedback loop in , , is greater than 1 and therefore confers some degree of robustness , for typical values of activity , a∼0 . 2 or greater , the term is only of order 1 and therefore not sufficient to compensate for small Kb . Robustness therefore likely arises from the slow kinetics of tether binding . The final consideration is that measurements [36] indicate that the number of cytoplasmic and localized enzymes are comparable and therefore that the forward and reverse rates of Eqs . ( 9 ) and ( 10 ) are roughly equal . This condition not only leads to comparable numbers of localized and cytoplasmic enzymes , but also indicates that the rates of tether binding and unbinding fall in the regime in which the steady-state activity is robust to the total number of enzymes . Specifically , for CheR , requiring the forward and backward rates of Eq . ( 9 ) to be comparable yields , leading to for typical values of a ( 0 . 3–0 . 5 ) [43] . The argument for CheB is analogous . Satisfying this constraint therefore leads not only to both comparable numbers of localized and cytoplasmic enzymes , but also to a steady-state activity that is robust to the total enzyme ratio . In this manner , the steady-state of the bacterial chemotaxis system can remain robust even when the localized enzymes operate at saturation .
Chemotactic bacteria are able to navigate chemical gradients with strengths ranging over five orders of magnitude [19] . This remarkable capability results from the capacity of the system to amplify small input signals while adapting to a wide range of concentrations of persistent stimulus . The cooperative receptor-receptor interactions that amplify input signals are facilitated by the formation of large receptor clusters , structures that are strongly conserved across bacterial species [5] . Adaptation to stimulus requires the efficient recruitment of cytoplasmic enzymes to these clusters , which is achieved through the presence of a high-affinity enzyme-tethering site on most receptors . These tethers , together with the dense structure of the receptor lattice , give rise to assistance neighborhoods [27] and possibly enzyme brachiation [37] . These features increase the distributivity of methylation , decreasing the likelihood that enzymes become localized in neighborhoods within which they have low binding affinity and therefore act inefficiently . Building on previous work that showed assistance neighborhoods were necessary for precise adaptation in a single strongly coupled signaling complex [28] , [38] , we found that assistance neighborhoods and enzyme brachiation contributed to precise adaptation to stimulus . We further linked distributive methylation to the presence of signaling noise in the output and showed how high signaling noise may coexist with a mean level of receptor activity that is robust to changes in the ratio of the adaptation enzymes . This ratio is not exactly conserved across populations . Consequently , if the mean activity were not sufficiently robust , the ultrasensitivity of the flagellar motor [59] , [60] would lead to a significant fraction of nonfunctional cells permanently in the running or tumbling state . This robustness to the ratio of adaptation enzymes occurs even though the localized enzymes work in the saturated regime . This scheme is not possible for the simpler covalent modification system studied by Goldbeter and Koshland , in which saturated enzyme kinetics always corresponds to ultrasensitivity to the enzyme ratio . The mechanism described here is not necessarily restricted solely to the bacterial chemotaxis system . The analytical model presented in this study describes generally an extension of the Goldbeter-Koshland [12] motif in which enzymes transition between active and inactive states , whether by localization to the substrate prior to modification , as in the bacterial chemotaxis model , or by chemical activation of the enzyme . This simplified model captures the essential features underlying large fluctuations: slow enzyme activation relative to the modification rate , saturated kinetics between the activated enzyme and the substrate , and distributive modification . While the kinetics of activated enzyme and substrate may be saturated , the robustness of the system to the overall expression levels of the enzymes may be preserved if the enzyme activation ( localization ) rate is sufficiently small relative to the deactivation ( delocalization ) rate . The effects of enzyme localization and the relationship between rapid enzyme rebinding and processivity have been considered in studies of MAP kinase cascades . A recent study of the mating response in yeast [61] discusses a mechanism in which the kinase Fus3 and phosphatase Ptc1 bind a docking site on the substrate Ste5 prior to modification . Since the docked enzymes operate at saturation , the system is ultrasensitive to changes in the number of recruited enzymes , similar to the chemoreceptor-enzyme system discussed in this work . Unlike the chemotaxis system , however , yeast exploits these saturated kinetics to produce a switch-like response in the steady state . The theoretical work of Takahashi et al . [13] also considers the MAP kinase system , using it as a model to explore the role of enzyme diffusion in determining whether substrate modification is processive or distributive . The authors conclude that slow diffusion , which causes the enzyme to bind and phosphorylate the same substrate molecule repeatedly , can effectively convert a distributive mechanism into a processive one , reducing the sensitivity of the system . The same effect figures prominently in our model of the bacterial chemotaxis system but in the opposite regime , in which the brachiation process serves to randomize enzyme positions between methylation events . Future studies of the bacterial chemotaxis system may further clarify the role of enzyme brachiation in adaptation . Different configurations of clustered receptors from that considered here , such as less dense clusters that have been shown to reduce cooperativity [62] , or larger numbers of significantly smaller clusters [63] , could hinder the ability of localized enzymes to visit a large number of unique receptors . In these cases our results suggest that signaling noise would be reduced . Interestingly , brachiation may be particularly important when considering cluster structure within local adaptation models [64] . In these models , receptors of different types respond specifically to different stimuli . Consequently , successful adaptation may depend on the ability of the adaptation enzymes to localize efficiently to responsive receptors . Brachiation may be critical for such efficient localization , particularly when considering the adaptation of low abundance receptors to their specific stimuli . While many systems benefit from minimizing signaling noise , studies of bacterial chemotaxis have shown that noise may increase the performance of the system in sparse environments while introducing only minimal deleterious effects . In empty environments , signaling noise may lead to faster cellular exploration to locate nutrient sources more efficiently [32] , [33] , [39] . Signaling noise has also been shown theoretically to increase tracking performance in shallow gradients [32] , [33] , [35] . These results are consistent with a picture of the chemotaxis system being not purely a signal transduction system , for which minimizing noise would typically be desirable , but also a feedback system in which the output controls the sampling of the input .
Since changes in receptor activity are effectively instantaneous relative to the slow methylation kinetics , activation of the receptor clusters is described by an equilibrium MWC model [22] , [23] . Clusters in the model are composed of N = 6 Tar homodimers . The free energy difference between the active and inactive states of the cluster is decreased by ε1 per methylation level and increased by in the presence of methyl-aspartate attractant L . Then the fraction of active clusters is given by ( 13 ) with m the methylation level . Parameter values were taken from fits to dose response measurements [43] and reproduced in Table S1 . In the stochastic simulation , m is taken to be the methylation level of a single MWC signaling unit and a ( m , L ) is used to calculate the activity of each MWC unit individually . In the analytical model , following Shimizu et al . [43] , m is the average methylation level per receptor cluster and a ( m , L ) is taken to be the average activity of all receptors in the system . We analyze the signaling properties of the model Eqs . ( 8–10 ) by performing a perturbation analysis around the steady state . Small displacements in the numbers of chemical species x evolve according to the linear system of Itô stochastic differential equations ( 14 ) in which A is the Jacobian matrix of the system , B is the diffusion matrix , and W ( t ) is the multidimensional Wiener process . By the linear noise approximation , BTB = S diag ( v ) ST with S the stoichiometry matrix and v the propensity vector [55] , [56] . The system in Eq . ( 14 ) is a multivariate Ornstein-Uhlenbeck process [57] . A has eigenvalues with negative real components , indicating the system relaxes to steady state after perturbation . The steady-state variance in the output of the system is obtained by solving the Lyapunov equation ( 15 ) for the covariance matrix σ . Additional details of the noise calculation are presented in the Supporting Text S1 .
|
To navigate their environments , organisms must remain sensitive to small changes in their surroundings while adapting to persistent conditions . Bacteria travel by performing a random walk biased toward nutrients and away from toxins . The decision of a bacterium to continue in a given direction or to reorient is controlled by the state of its chemoreceptors . Chemoreceptors assemble into large polar clusters , an arrangement required for the amplification of small stimuli . We investigate how this organization affects the kinetics of the enzymatic reactions through which the receptors adapt to persistent stimuli . We show that clustering can lead to large fluctuations in the state of the receptors , which have been observed in Escherichia coli and may aid in the navigation of weak stimulus gradients and the exploration of sparse environments . Additionally , we show that these fluctuations can occur around a mean receptor state robust to changes in the numbers of the adaptation enzymes . Since enzyme expression levels vary across a population , this feature ensures a high proportion of functional cells . Our study clarifies the relation between fluctuations , adaptation , and robustness in bacterial chemotaxis and may inform the study of other biological systems with clustered receptors or similar enzyme-substrate interactions .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[] |
2013
|
Adaptation Dynamics in Densely Clustered Chemoreceptors
|
Repairing broken chromosomes via joint molecule ( JM ) intermediates is hazardous and therefore strictly controlled in most organisms . Also in budding yeast meiosis , where production of enough crossovers via JMs is imperative , only a subset of DNA breaks are repaired via JMs , closely regulated by the ZMM pathway . The other breaks are repaired to non-crossovers , avoiding JM formation , through pathways that require the BLM/Sgs1 helicase . “Rogue” JMs that escape the ZMM pathway and BLM/Sgs1 are eliminated before metaphase by resolvases like Mus81-Mms4 to prevent chromosome nondisjunction . Here , we report the requirement of Smc5/6-Mms21 for antagonizing rogue JMs via two mechanisms; destabilizing early intermediates and resolving JMs . Elimination of the Mms21 SUMO E3-ligase domain leads to transient JM accumulation , depending on Mus81-Mms4 for resolution . Absence of Smc6 leads to persistent rogue JMs accumulation , preventing chromatin separation . We propose that the Smc5/6-Mms21 complex antagonizes toxic JMs by coordinating helicases and resolvases at D-Loops and HJs , respectively .
Sexual reproduction in eukaryotes relies on the generation of haploid gametes from diploid somatic cells by a process called meiosis . Meiosis achieves the required reduction in ploidy by executing a single round of DNA replication followed by two consecutive rounds of chromosome segregation . The correct segregation of homologous chromosomes depends on the formation of chiasmata , which are crossovers ( COs ) held in place by distal sister chromatid cohesion [1] , [2] . Crossovers , and in many organisms also the identification and pairing of homologous chromosomes , require the repair of programmed meiotic DNA double-strand breaks ( DSBs ) in prophase I at sites that have completed pre-meiotic DNA replication [3] . As a consequence of the repair of DSBs , stable recombination intermediates called double Holliday Junctions ( dHJs ) can arise , and can be resolved to generate crossovers ( COs ) or non-crossovers ( NCOs ) . Sophisticated mechanisms controlling CO numbers and distribution ensure that each bivalent ( two paired homologous chromosomes ) receives at least its obligate CO . In budding yeast , the ZMM pathway ( an acronym for the involved proteins Zip1-4 , Mer3 , Msh4/5 [4] ) is a key part of this control . It guides a subset of DSBs to become allelic COs between homologs by allowing these breaks to form dHJs specifically resolved to COs depending on Exo1-Mlh1/3 [5] . A sophisticated machinery , including the Synaptonemal Complex ( SC ) , regulates the progression of recombination intermediates in the ZMM pathway . The SC is a tripartite proteinaceous structure connecting bivalents along their whole length at a distance of 100 nm in the pachytene stage of meiosis . After completion of synapsis , Polo-like kinase Cdc5 activation triggers the resolution of dHJs shortly before cells become committed to enter the first meiotic division [6] , [7] . Non-ZMM DSBs are not destined to become COs and follow another main route , involving fast and minimal-risk repair by Synthesis Dependent Strand Annealing ( SDSA ) . In SDSA the invasion of one broken DNA terminus into an intact template allows DNA repair synthesis beyond the break . Importantly , interaction remains limited by the rapid displacement of the invading strand . This pathway does not produce COs but may nevertheless facilitate homolog recognition . Recently , the RecQ helicase BLM/Sgs1 [8] has been shown to be a central player in both primary pathways of meiotic recombination [5] , [9]–[12] . In the SDSA pathway , Sgs1 promotes strand displacement thereby preventing stabilization of the invasion . In the ZMM pathway , Sgs1 delays repair until the formation of stable Single End Invasion ( SEI ) and dHJ intermediates is appropriate , presumably after the cell has accumulated some information about the correctness of the invaded target . In budding yeast , the ZMM and SDSA pathways together accomplish the large majority of meiotic recombination events [5] . If recombination intermediates escape the two pathways described above , unregulated , mitotic like Joint Molecules ( JMs ) can arise , such as dHJs not associated with the proper ZMM machinery . We will refer to these intermediates as “rogue” , in the sense that they are “unprincipled , unreliable and with potentially destructive properties” which could result in non-allelic COs after escaping the two canonical , safe pathways . Although these rogue intermediates constitute only a minor fraction of recombination events in wild type meiosis , they can block chromosome segregation if unresolved . Resolution of such JMs is mediated by the overlapping activity of the three HJ resolvases Mus81-Mms4 , Slx1-Slx4 , and Yen1 [5] , [7] , [9] . Cdc5/Plk1 also controls the activation of Mus81-Mms4 and Slx1-Slx4 . dHJs represent potential danger for their inability to elicit a DNA damage checkpoint response [10] . Due to their stability , they may cause chromosome nondisjunction or even block segregation if not resolved . Conversely , resolution of HJ intermediates between non-allelic positions can result in deletions or translocations , even in dicentric chromosomes . This explains attempts of the cell to avoid the formation of such stable recombination intermediates right away , mediated by the action of helicases such as Srs2 and BLM/Sgs1 that can destabilize Rad51 filaments and nascent invasions [13]–[17] . Even if dHJs formed , the cell may be able to unwind them conservatively through the action of Sgs1-Rmi1-Top3 [18] , [19] , a process termed dHJ dissolution . Ultimately , once the cell has passed the DNA damage checkpoint and is committed for division , resolvases will be given priority for timely removal of linking Holliday Junctions ( HJs ) to avoid chromosome nondisjunction events . While a key set of DNA metabolizing activities have recently been described , important questions concerning the chromosomal context remain unanswered . Coordination of local recruitment , regulation and orientation of anti-JM helicases and resolvases in the in vivo context remain completely obscure to date . In this study , we provide evidence that the Smc5/6-Mms21 complex mediates such functions . SMC complexes are evolutionarily conserved from gram-negative bacteria to mammals , serving critical functions in chromosome metabolism , thereby helping to preserve chromosomal and genomic integrity . Eukaryotes have three distinct , essential ring- shaped SMC complexes at their disposal; the cohesin complex ( Smc1-Smc3 ) , linking sister chromatids until the metaphase/anaphase transition , condensin ( Smc2-Smc4 ) , thought to regulate higher order chromosome structure and finally the Smc5/6-Mms21 complex , involved in recombinational DNA repair . All SMC ring complexes exhibit the ability to associate with DNA and the hinge region of several SMC complexes , including condensin , has been shown to bind to DNA [20] . At least for cohesin and condensin there is direct evidence that DNA is topologically entrapped inside the ring [21]–[24] . While this has not yet been tested for the Smc5-Smc6 complex , it was shown that both Smc5 and Smc6 individually bind stably to DNA with a strong preference for single stranded DNA [25] , [26] . Taking into consideration the high similarity in structure and size to the other SMCs , it is not unlikely that Smc5/6-Mms21 also encloses DNA strands topologically . The SMC proteins consist of two extended domains that fold back on themselves at their central hinge region into an anti-parallel coiled-coil structure . This brings the two terminal Walker A/B motifs in close proximity to form ABC-like ATPases . Two SMC proteins connect via their hinge regions , while the kleisin subunit closes the ring by linking the terminal SMC ATPase heads . Different SMC complexes contain characteristic non-SMC subunits that form integral components of the functional complex . In budding and fission yeast the Smc5/6-Mms21 complex comprises six Non-SMC Elements ( NSEs ) , Nse1 to Nse6 , of which Nse4 is the kleisin ( Figure 1A ) . Beside Smc5 and 6 , at least Nse1 to Nse4 are conserved from yeast to man [27]–[29] . A unique feature of the Smc5/6-Mms21 complex , compared to cohesin and condensin , is its SUMO E3 ligase subunit Mms21/Nse2 [27] , conferring the ability to post-translationally modify target proteins . The SUMOylation substrates of Mms21 remain poorly defined to date , however , candidates include Smc5 , Scc1 , the kleisin subunit of cohesin , as well as telomeric proteins [30] , [31] . Mms21 is stably bound through an extensive N-terminal coiled-coil interface to the coiled coil domain of Smc5 [32] , [33] while the E3 SPL RING structure that recruits the SUMO E2 ligase Ubc9 resides at its C-terminus [32] . Mms21 is an essential subunit of the Smc5/6-Mms21 complex in Saccharomyces cerevisiae and the mere disruption of its interaction with Smc5 is lethal [32] , [34] . Notably , elimination of Mms21's SUMO E3 ligase activity alone is not lethal , but sensitizes the cell to genotoxic agents [27] . Yeast cells mutated in the Smc5/6-Mms21 complex become hypersensitive to genotoxic agents including hydroxyurea , MMS , ionizing irradiation and UV [27] , [35] , [36] , and DNA regions vulnerable to homologous recombination ( HR ) like rDNA and telomeres are particularly affected [37] . Accordingly , Smc5/6 was also found to accumulate at such sites prone to recombinogenic damage [38] , [39] . Accumulation of X-shaped DNA intermediates upon challenge and aberrant processing of stalled replication forks have been reported in Smc5/6 mutants [40] and roles for the Smc5/6-Mms21 complex in multiple repair pathways , including homologous recombination , have been suggested [41] . Complete elimination of functional Smc5/6-Mms21 complex from vegetative cells leads to heterogeneous defects , including cells arresting in metaphase , chromosome missegregation and eventually lethality [42] . In meiosis , phenotypes of Smc5/6 mutants in S . pombe and S . cerevisiae represent exacerbated manifestations of those observed during mitosis , including catastrophic failures in meiotic divisions [43] , [44] . In a synapsis mutant ( zip1Δ ) of S . cerevisiae , the homologous chromosomes tended to become more attached to each other and chromosomal entanglements seemed to increase . These defects were partly caused during premeiotic S-phase , as they were not fully dependent on initiating meiotic recombination [44] . Another study in S . pombe found accumulation of X-shaped DNA molecules in nse6 mutants , originating from meiotic recombination [43] . Here we study the role of Smc6 in meiotic recombination by eliminating Smc6 after premeiotic S-phase is largely complete . This leads to an accumulation of unresolved JMs at meiotic recombination hotspots and a corresponding uniform block in nuclear divisions . We observe strong overlaps of Smc6-chromatin binding with that of Sgs1 at a substantial number of meiotic DSB-hotspots . In chromosome spreads , Smc5/6 foci co-localize with Rad51/Dmc1 foci side by side . We further show that the E3 ligase deficient mms21-11 allele restores dHJ and CO formation and improves spore viability in the robust ZMM mutant zip3Δ , implying an important role of Mms21 in the prevention of dHJ formation in this background . Resolution of the JMs in the mms21-11 mutant depends on the Mus81-Mms4 resolvase . The Smc5/6-Mms21 complex is also required for HJ resolvase activity responsible for eliminating rogue HJ intermediates . A dramatic accumulation of unresolved JMs and a pronounced reduction in COs and NCOs was observed in the smc6-56 sgs1 double mutant , in which all meiotic recombination follows an aberrant non-ZMM , non-SDSA pathway , while COs form at near normal levels in the smc6-56 SGS1 . We conclude that Smc5/6-Mms21 collaborates with helicases and resolvases to both prevent and eliminate JMs that arise outside the ZMM recombination pathway .
In order to characterize the role of the Smc5/6-Mms21 complex in the context of meiotic recombination , we utilized two previously described mutants with distinctive phenotypes . smc6-56 is a temperature sensitive allele , carrying three missense mutations in the N-terminus proximal coiled-coil region ( A287V , H379R , I421T ) conferring lethality at restrictive temperature [45] . The mms21-11 allele terminates after Thr183 and lacks the C-terminal SPL-RING domain , thus depriving Mms21 of its SUMO E3 ligase activity by abolishing its interaction with the SUMO E2 enzyme Ubc9 [27] , [46] . To distinguish the role of the Smc5/6-Mms21 complex in meiotic DSB repair from that of mitosis and premeiotic S-phase we inactivated the smc6-56 allele by shifting the temperature of the synchronized cultures gradually to restrictive conditions at 33°C ( Figure 1B ) . Cells are allowed to exit mitosis and undergo most of meiotic DNA replication under permissive or semi-permissive conditions ( Figure S1A , B ) . Fully restrictive conditions were applied from 2 . 5 hrs post induction of meiosis , when most cells are in late S-phase and the earliest meiotic DSBs become detectable [47] . The constitutive mms21-11 mutant was analyzed at 30°C . Under these conditions , smc6-56 causes neither a delay in meiotic progression and spore formation , nor does it produce evident defects in meiotic DSB formation , repair , or chromosome axis architecture . Chromosome synapsis is flawless with normal timing ( Figure S1C ) , meiosis I and meiosis II spindles form with normal kinetics , and a mnd1Δ smc6-56 double mutant arrests indistinguishably from mnd1Δ which prevents DSB repair at the strand invasion step [48]–[50] in prophase1 indicating a functional DNA damage checkpoint ( Figure 1F ) , suggesting efficient DSB-turnover in smc6-56 comparable to wild type . However , 92–99% of cells with inactivated Smc6 failed to segregate the chromatin during both meiotic divisions and thus produced a single nucleus outside of four empty spores ( Figure 1C , E ) . Unable to separate the chromatin , anaphase I and II spindles ultimately collapse , resulting in aberrant spindle morphologies from Meiosis II onwards ( Figure 1D ) . The timing of spore formation and the number of cells forming spores was as in wild-type , however , the number of aberrant asci containing one , two or three spores was increased at the expense of complete tetrads ( 21–27% , less than half of wild type ) . In 88–95% of the tetrads , all four spores were empty and only 0 . 5–4 . 5% of tetrads had DNA in all four spores . Most empty spores ultimately collapse and partially lyse by 24 hours . Notably , those few tetrads that managed to receive DNA in all 4 spores and survive zymolyase digestion show a fairly high spore viability ( 53 . 75% , n = 80 spores/20 tetrads ) . To test whether the observed meiotic catastrophe depended on meiotic recombination , DSB formation was abolished by introducing the spo11Δ mutation in the smc6-56 background . Indeed , meiotic divisions were largely restored with 80% of the cells completing at least one division and 60% of the cells finishing both meiotic divisions ( Figure 1G ) . To demonstrate that also the sister chromatids can separate , the triple mutant with spo13Δ was analyzed . We found that almost 60% of the cells underwent the single division with normal kinetics ( Figure 1G ) . Thus , the applied regime of conditional inactivation of Smc6 separates most of the defects related to mitosis and meiotic S-phase from those in meiotic recombination . We conclude that the Smc5/6-Mms21 complex is essential for allowing chromosome disjunction upon initiation of meiotic recombination . In contrast to the smc6-56 mutant , mms21-11 showed almost normal kinetics of meiotic divisions , forming tetra-nucleated cells with only a slight delay ( Figure 1E ) . Synapsis and sporulation are efficient ( Figure S1C ) and the resulting tetrads exhibited high spore viability ( 88 . 75%; n = 400 spores of 100 tetrads ) . Only a small fraction of tetrads ( 3–3 . 25% ) were missing DNA in one or more spores . Consequently , the SUMO E3 ligase activity of the Smc5/6-Mms21 complex is not essential to prevent massive chromosome non-disjunction . To characterize the binding of Smc5/6-Mms21 to meiotic chromatin in the context of meiotic recombination we used the epitope tagged SMC6-myc13 construct and prepared meiotic nuclear spreads from synchronized cultures at various time points for cytology . Smc6-myc13 does not show any obvious defects and co-localizes with Smc5-HA3 on chromosome spreads , suggesting it is a valid representation of the complex ( Figure S2A ) . Smc6-myc13 localizes in individual foci which appear early in meiosis at approximately the same time as Rec8 ( Figure S2B ) and soon accumulate to considerable numbers until earliest prophase 1 ( Figure S3 ) . We used antibody-staining against the synapsis specific Zip1 protein for staging ( from isolated Zip1 foci up to full SCs ) . In nuclei engaged in meiotic recombination ( Zip1 positive ) an average of 118±18 ( n = 21 nuclei scored ) Smc6 foci were scored and this number did not change significantly in different stages of prophase 1 . After prophase 1 , the intensity of foci decreased slightly , but numbers remained high until Metaphase II/Anaphase II transition ( Figure 2C , S3 ) . A particular enrichment was observed for the rDNA region on chromosome 12 , which remains unsynapsed in the course of meiotic recombination ( Figure 2A white arrows , S3 ) . The preferred binding sites of Smc6 in wild-type meiosis differ markedly from those of the related cohesin complex , which binds specifically to chromosome regions that constitute the chromosome axis upon condensation [1] , [51] , [52] . On meiotic chromatin , foci of Smc6-myc13 and Rec8-HA3 ( a tagged version of the meiotic kleisin-subunit of cohesin ) generally exclude each other ( Figure 2A , S2B ) . When chromosomes condense , Smc6 signal often protrudes from the Rec8-axes , suggesting preferential binding of the Smc5/6-Mms21 complex to DNA regions not associated with the axis ( Figure 2A ) . In order to determine the chromosomal binding sites of Smc5/6-Mms21 complex , ChIP-Seq ( Chromatin-Immuno Precipitation followed by next generation sequencing ) of Smc6-myc13 was performed . Synchronous meiotic cultures were crosslinked with formaldehyde , subjected to ChIP and the precipitated DNA deep sequenced on an Illumina platform . Figure 2D , E show a 130 kb representative region of Chromosome V with about 55 medium to weak DSB hotspots mapped previously [53] . Figure 2D , E demonstrate that the majority of Smc6 peaks localize precisely to these hotspots . In the example shown , the 3 . 5 hour Smc6 peaks fall into 45 ( 82% ) hotspots , while 49 ( 89% ) hotspots coincide with a 4 . 5 hour Smc6 peak . The vast majority of peaks are precisely on top of the hotspots ( up to 80% genome wide ) , however , we note that the intensities of the co-localizing Smc6 peaks are often not proportional to the intensity of the break site , resulting in the relatively low genome wide Pearson correlations despite precise co-localization . To address whether Smc6 localization depended on the formation of DSBs by Spo11 , the experiment was repeated in a spo11Δ mutant . A genome wide reduction of Smc6 at DSB sites was observed , however , surprisingly , Smc6 still accumulated at the majority of hotspots ( Figure 2D , E , S4 ) . The genome wide ChIP experiments were accompanied by qChip at a DSB hotspot ( YCR047 , 211k ) , a core site ( 219k ) and a cold region ( ADP1 , 136k ) all on chromosome III ( see Figure S4 ) . qChIP confirmed the results on the corresponding positions of the ChIP-Seq profiles ( Figure 2F ) . The ChIP-Seq experiments were repeated confirming the reported results ( not shown ) . The qPCR of the biological repeat confirmed the Spo11 improved enrichment of Smc6 at the hotspot precisely ( Figure S4C ) . Not all binding sites of Smc6 are DSB-sites . For instance , sharp Smc6-peaks mark all the centromeres ( Figure 2H , S4A , B ) . Furthermore , confirming our cytological observation of abundant rDNA localization of Smc6 foci , Spo11 independent Smc6 signals flank the 35S rDNA transcriptional units ( Figure 2G ) . Binding of Smc6 to the rDNA region is in agreement with the crucial anti-recombination role that Smc5/6-Mms21 plays at this repetitive DNA locus [37] . For the remaining Smc6 signal we often observe overlaps with meiotic chromosome axis sites as defined by Mer2 and Hop1 [51] . Such axis specific enrichment for Smc6-myc13 is , however , rather low ( Figure 2 , S4 ) . Similar conclusions have been drawn in the study of Copsey and coworkers ( accompanying manuscript ) [54] , although more prominent binding of Smc5 at core sites is found than in our study . The quantitative discrepancies in the recovery for Smc5 and Smc6 may reflect different biological properties of the complex . For instance , it was reported for the related cohesin complex , that different cohesin populations exist regarding stability of chromatin binding [55] . Further , we observed previously that the choice of a tag can influence the balance between core and DSB site residency [51] , [56] . Differences can also arise due to the different subunits analyzed , the different platforms and resolutions used and the highly stringent background subtraction that we use . Importantly , both studies observe signals consistent with a role of Smc5/6 at sites of meiotic recombination . Due to observed recruitment of Smc6 to DSB hotspots , we asked whether Rad51 , the eukaryotic strand exchange protein which assembles along the resected DSB-ends and facilitates the strand invasion step in HR repair , co-localizes with Smc6-myc13 . The number of Smc6-myc13 foci far exceeded ( about 4-fold ) the Rad51 foci , consistent with DSB independent loading of Smc6 to chromatin . Strikingly , we found almost no on-top co-localization of Smc6 with Rad51 ( Figure 2B ) . Counting foci on 6 meiotic nuclei confirmed this impression . Only 0–10% of the Rad51 foci directly co-localized with Smc6 ( on average 6 . 5±4% ) . However , a total of 85±4% ( n = 6 nuclei scored ) Rad51 foci were in a side-by-side configuration with one or two Smc6 foci , even on nuclei with strongly spread chromatin ( Figure 2B ) . A similar relation of Smc6-myc13 with the ZMM-DSB marker protein Zip4-myc9 ( Figure S2C ) was observed . In summary , we identify a precise and sensitive association of Smc6 with meiotic recombination hotspots and enrichment of Smc6 to DSB hotspots upon actual DSB formation . The frequent side-by-side co-localization of Smc6 and Rad51 recombinase foci is indicative of a primarily spatial separation with juxtaposed positioning of the two complexes . Unresolved Holliday Junctions ( HJs ) stably connect chromosomes that had engaged in homologous recombination and thus may result in chromosome nondisjunction . A sufficiently large number of such unresolved recombination intermediates prevent nuclear divisions resulting in mitotic or meiotic catastrophe [7] , [10] , [12] . Since previous studies reported an accumulation of repair intermediates in mutants of the Smc5/6-Mms21 complex [37] , [57] , [58] , we asked whether an accumulation of unresolved recombination intermediates in the smc6-56 mutant might explain the Spo11 dependent failure in nuclear divisions . Meiotic recombination was followed in a physical assay at the URA3-arg4 locus under conditions that preserve HJ intermediates [59] . This Southern based assay allows quantification of the key intermediates of meiotic recombination: DSBs , dHJ intermediates , as well as crossover and non-crossover recombination products due to restriction polymorphisms between the parental chromosomes ( Figure 3A ) . Using this recombination reporter system , wild type and mutant strains were analyzed under restrictive conditions for smc6-56 in a time course experiment . Joint molecule levels were measured by probing against ARG4 DNA on genomic XmnI restriction-fragments . No differences were noted in the transient appearance of DSBs . However , while joint molecules appear transiently in wild type with low steady state levels and a clear peak at 5 hours when cells are in the pachytene stage of meiosis , the smc6-56 mutant accumulated high molecular weight recombination intermediates that failed to be resolved ( Figure 3B ) . The observed persistent joint molecules in smc6-56 amount to more than twice that of the JM peak in wild type , exceeding the amount of stable joint molecules reported to block chromosome segregation ( for example in the mms4-mn yen1Δ double mutant [9] ) . In addition , while joint molecules rarely form between sister chromatids in wild type ( IS:IH≤0 . 2 ) , in the smc6-56 mutant a substantial amount of inter sister JMs ( IS-mom:IH 0 . 32–0 . 38 , by two different assays; XmnI and XhoI/EcoRI ) contribute to the overall persistent JMs ( Figure 3B , D ) . Using a HIS4 fragment to probe for COs and NCOs after a different digest ( XhoI/EcoRI ) revealed that the levels of both these recombination products were not significantly altered in smc6-56 ( Figure 3C , D , F , G ) . Consequently , the sum of COs , NCOs and JM-intermediates in the mutant exceeds corresponding numbers in the wild type , uncovering a role of Smc5/6-Mms21 in early destabilization of intermediates to prevent stable JM formation . In summary , the Smc5/6-Mms21 complex is required to prevent the accumulation of aberrant , unresolved recombination intermediates . The results further suggest that most of the stabilized JMs arise from recombination events not resulting in CO or NCO products in wild type , such as inter sister SDSA or dissolution of rogue dHJs . In particular , this suggests that the CO specific ZMM-pathway is not affected by the defect in smc6-56 . If the smc6-56 mutant confers persistent joint molecules to the cell that impede nuclear divisions , then there are two feasible explanations as to why the mms21-11 mutant does not . Either mms21-11 represents a plain Smc5/6-Mms21 hypomorph in meiosis with remaining activity at a level such that formation of aberrant joint molecules is negligible . Alternatively , mms21-11 may represent a separation of function mutant of the complex in which prevention of aberrant JM formation and the resolution of such by the Smc5/6-Mms21 complex had become separated . If the latter is true , it should be possible to detect inappropriate joint molecule formation as well as their removal in mms21-11 . To test if the mms21-11 mutant indeed fails to prevent the formation of additional and atypical joint molecules in considerable amounts , we assayed on surface spread meiotic nuclei for the presence of excess axial associations in the background of the ZMM mutant zip3Δ . In mutants of the ZMM pathway chromosome synapsis is defective , the repair of ZMM-breaks is inhibited through activity of the Sgs1 helicase [11] , and dHJ and CO formation are impaired while Non-ZMM breaks are repaired by SDSA [4] . Axial associations as first described in the zip1Δ mutant are the cytological manifestation of stable recombination intermediates between the homologs , approximating dHJs and COs numbers in wild type [60] . In the SK1 strain background , zip3Δ confers one of the most severe ZMM phenotypes , exhibiting a strong reduction in JMs and a robust prophase I arrest [4] . Accordingly , at 5 hrs in SPM , when in wild type cells most homologs are synapsed , univalents almost bare of axial associations and recognizable pairing dominate the zip3Δ mutant phenotype ( Figure 4A ) . 63% of nuclei have no more than 2 chromosomes per nucleus paired or connected by axial associations ( n = 100; Figure 4B ) . In contrast , pairing and axial associations are frequent in the zip3Δ mms21-11 double mutant , with 5–6 chromosome pairs on average in one experiment and as many as 7 in a biological repeat ( n = 100; Figure 4A , B ) . A similar improvement in pairing can be seen for zip3Δ smc6-56 , but not for mnd1Δ mutants , which are defective in the strand invasion step ( Figure S5A ) [48] , [49] . In addition to the increased number of axial associations , mms21-11 ameliorated the pachytene arrest of zip3Δ . Following the spindle morphology as a marker of meiotic progression , we find 88% of zip3Δ cells at 11 hrs still in prophase I , whereas 60% of the zip3Δ mms21-11 cells had at the same time already progressed beyond prophase I ( n = 200; Figure 4C ) . Additionally , sporulation of zip3Δ was improved from 12 . 75% to 22 . 5% in the double mutant at 24 hours ( n = 400 ) . Thus , ZMM DSBs of zip3Δ are turned over in the mms21-11 background into intermediates that do not elicit a DNA damage checkpoint response . The suppression of the prophase I arrest is not due to a checkpoint defect because the resulting tetrads of the double mutant exhibit a greatly improved spore viability of 65 . 63% , from 23 . 13% in zip3Δ ( n = 160 spores; Figure 4D ) . The opposite would be expected for a checkpoint failure . Impairment of an early anti-recombinogenic function and inappropriate transformation of DSBs to JMs in mutants of the Smc5/6-Mms21 complex is further supported by the observation that neither absence of the three resolvases Mus81-Mms4 , Slx1-Slx4 and Yen1 , responsible for the removal of unregulated JMs [5] , [9] , nor the absence of all four meiosis relevant resolvases , Mlh1/3-Exo1 , Mus81-Mms4 , Slx1-Slx4 and Yen1 which account for ≥90% of the meiotic resolution activity [5] can improve chromosome pairing in zip3Δ ( Figure S5B , C in comparison to nse4-mn and sgs1-mn , meiosis specific null alleles of the Smc5/6 kleisin subunit and the Sgs1/BLM helicase , respectively ) . These results indicate that the SUMO E3 ligase deficient mms21-11 allele fails to antagonize dHJ-formation during DSB repair in the zip3Δ mutant , thereby improving bivalent formation , facilitating meiotic progression and ultimately greatly improving spore viability . This is similar to the effect of sgs1-mn ( zip3Δ sgs1-mn: 71 . 25% , n = 160 spores ) . While restoration of COs in zip3Δ mms21-11 remains to be confirmed by physical analysis , we conclude that Mms21 as part of the Smc5/6 complex is specifically required for an early , antagonistic function of the complex in meiotic recombination , most likely by preventing the formation of illegitimate joint molecules , a role previously reported for the helicase BLM/Sgs1 [10]–[12] , [61] . Three resolvases with overlapping function are responsible for eliminating rogue HJ intermediates in budding yeast: Mus81-Mms4 , Slx1-Slx4 , and Yen1 . JMs arising outside the ZMM pathway depend on these three partially redundant resolvases for resolution , with the XPF family nuclease Mus81-Mms4 showing the biggest contribution [5] , [9] . If , as the Zip3 results suggest , considerable amounts of rogue joint molecules form also in the mms21-11 single mutant , these joint molecules must apparently be efficiently removed prior to nuclear divisions as implied by functional chromatin segregation and high spore viability in mms21-11 . The most important “rogue JM resolvase” Mus81-Mms4 is the most likely candidate to carry out this vital role . If mms21-11 represented a separation of function mutant of the Smc5/6-Mms21 complex , unable to prevent aberrant JMs , additional inactivation of Mus81-Mms4 should block nuclear divisions , thereby recreating the full smc6-56 mutant phenotype . To test this hypothesis Mus81-Mms4 function was eliminated in the mms21-11 background using a characterized meiotic null allele of MMS4 ( mms4-mn ) [10] . Indeed , in the mms21-11 mms4-mn double mutant meiotic segregation of chromosomes is blocked , resulting in meiotic catastrophe , while only rarely a cell of the single mutants shows this defect ( Figure 4G , H ) . Chromosome synapsis and meiotic progression were not different from wild type for the single mutants or the double mutant ( Figure 4F ) . As predicted by these results , persistent JMs accumulated in the mms21-11 mms4-mn double mutant at levels comparable to those found in the smc6-56 mutant while the JMs of both single mutants were successfully resolved ( Figure 4E , I ) . In wild type meiosis , most COs are formed in the ZMM pathway from dHJs resolved via Exo1-Mlh1/3 , while the NCOs arise from non-ZMM DSBs repaired through SDSA [5] , [9] . In the mms21-11 mutant , a fraction of both COs and NCOs become dependent on MMS4 as their formation is reduced in mms21-11 mms4-mn ( Figure 4J , K ) but not in MMS21 mms4-mn . Thus , these recombination products form in mms21-11 from the resolution of rogue dHJs by Mus81-Mms4 , implying that a significant fraction of ZMM and SDSA breaks in mms21-11 switch to a rogue JM fate . In summary , the evidence indicates that the Mms21 SUMO E3 ligase is required to prevent the formation of inappropriate HJ intermediates . However , in contrast to the smc6-56 mutant , the aberrant JMs can still be removed in mms21-11 , but require for this the activity of the Mus81-Mms4 resolvase . Despite the failure to resolve a considerable amount of JMs ( Figure 3B–D ) , the levels and kinetics of meiotic recombination products in the smc6-56 mutant are only marginally affected . Consequently , the Smc5/6-Mms21 complex is not essential for the resolution of all JMs in budding yeast meiosis . However , as the mms21-11 mutant does generate aberrant HJ intermediates that need Mus81-Mms4 for resolution , the Smc5/6-Mms21 complex could be required exclusively for the removal of the ZMM-independent “rogue JMs” . To specifically address the ability of the smc6-56 mutant to remove ZMM-independent JMs , JM resolution and product formation was analyzed in the background of an SGS1 meiotic null allele ( sgs1-mn ) . In sgs1-mn mutants , nearly all ZMM recombination intermediates adopt a rogue JM fate and , consequently , nearly all recombination products , namely COs and NCOs , depend on the “rogue JM resolvases” Mus81-Mms4 , Slx1-Slx4 , and Yen1 [5] , [9] . If the Smc5/6-Mms21 complex were critically required for the removal of rogue JMs via these resolvases , the smc6-56 mutant should strongly inhibit both JM resolution and product formation in the sgs1-mn background . As shown in Figure 5A–E , the sgs1-mn smc6-56 double mutant does indeed dramatically accumulate JMs , accompanied by a substantial loss in recombination products consistent with a failure of the rogue JM resolvases in non-ZMM JM removal . In contrast , the sgs1-mn single mutant efficiently resolves JMs to CO and NCO products ( Figure 5A–E ) , as expected [5] , [9] , [10] . Similarly to the smc6-56 single mutant , nuclear divisions are completely blocked in the double mutant , while meiotic progression based on spindle morphology is unaffected in single and double mutants ( Figure 5F , G ) . In the sgs1-mn background where nearly all DSBs adopt a rogue fate , the amount of persistent JMs in the sgs1-mn smc6-56 mutant is 3 to 4-fold elevated compared to the smc6-56 single mutant , reaching a high level of 6–7% unresolved JMs . The failure to resolve JMs in the sgs1-mn smc6-56 double mutant resulted in corresponding depletion of 60% of the sgs1-mn COs and 45% of the sgs1-mn NCOs ( Figure 5C–E ) . We conclude that the Smc5/6-Mms21 complex is specifically required for the resolution of the unregulated , “rogue” JMs and that the resolution activity lost due to smc6-56 equals the effect of loss of at least the most active rogue JM resolvase Mus81-Mms4 [9] , [10] . To investigate whether JMs might become terminally inaccessible to resolvases due to an early defect in smc6-56 , cells were allowed for 7 h 30 min to complete prophase I until the arrest in late pachytene of ndt80 under restrictive temperature . Cells were then released into permissive temperature to supply functional Smc6 for JM resolution . The release was mediated by addition of estradiol using an estradiol inducible NDT80 allele ( ndt80-IN ) . Ndt80 expression induces pachytene exit and Cdc5 expression which in turn activates the resolvases for resolution of dHJs [6] , [7] , [62] . If JMs had derailed and become unresolvable early , the late addition of Smc6 should not be able to support nuclear divisions . The key result of the experiment is shown in Figure 5I , namely that providing Smc6 after the ndt80 arrest restores nuclear divisions to wild type levels , with only a slight delay , required to resolve the accumulated JMs . Notably , all the controls were as expected , that is cells efficiently resumed meiotic progression upon estradiol addition independent of the temperature ( Figure 5H ) . Also , SMC6 performed nuclear divisions independent of the restrictive conditions , while of course successful nuclear divisions in smc6-56 depended on the termination of restrictive conditions ( Figure 5H , I ) . We conclude that it is sufficient to provide Smc6 function after the ndt80 arrest point to ensure chromosome segregation . The converse experiment , in which cells process DSBs under permissive conditions but are released under restrictive conditions from the ndt80 arrest , shows that cells block ( Figure S6 ) consistent with a critical role of Smc5/6-Mms21 complex in mediating the function of rogue JM resolvases beyond only Mus81-Mms4 . In summary , we conclude that Smc5/6-Mms21 promotes the function of the “rogue JM resolvases” at the time of resolution - and thus rather directly , and that the integrity and accessibility of the aberrantly formed JMs is not affected . The biological function of the Mms21 SUMO E3 ligase domain is presumably mediated through regulation of downstream factors . The most obvious candidates for destabilizing early intermediates are helicases that can unwind recombinogenic structures . The helicase reported to interact with the Smc5/6-Mms21 complex is Mph1 , an anti-recombinogenic , FANCM like helicase , but unlike Sgs1 or Mms21 SUMO ligase activity , Mph1 is dispensable for coping with the bulk of induced lesions [63] , [64] . However , absence of Mph1 activity partially alleviates defects in Mms21 SUMO E3 ligase mutants , suggesting dysregulation of Mph1 activity [63] , [64] . The same study also found that this suppression depends on the activity of Sgs1 , the sgs1Δ mutant being epistatic . These data suggest that Mph1 activity becomes toxic in the absence of Smc5/6-Mms21 and may hamper Sgs1 mediated repair . It is therefore possible that the Mms21 SUMO E3 ligase mediates its anti-JM formation activity by coordinating the two helicases , Sgs1 and Mph1 . However , observations in a previous study in which mms21-11 and sgs1Δ conferred a synthetic growth defect in the double mutant led to the notion that Sgs1 and Smc5/6-Mms21 may not work together [64] . In this study a striking overlap between the biological functions of Mms21 and Sgs1 is apparent . Similarly to Sgs1 , Mms21 prevents the formation of rogue JMs and is required for SDSA and the repair block of early ZMM breaks , although the defect of mms21-11 is clearly less pronounced than that of an sgs1 mutant . This similarity in behavior of mms21-11 and sgs1 mutants is also observed in studies on mitotic cells [40] . In vegetative growth , we found very similar synthetic interactions for mms21-11 as for sgs1Δ and mph1Δ ( Figure S7B , C; and [65] ) , including reduced growth for the mms21-11 sgs1Δ double mutant . For the synthetic interaction of mms21-11 and sgs1-mn in meiosis , we observed a moderate chromosome segregation defect in the double mutant . Specifically , 60% of tetrads were missing DNA in at least one spore and 39% of the tetrads ( n = 200 ) were devoid of DNA in all four spores ( Figure 6A ) . Since Mms21 is an essential subunit of the Smc5/6-Mms21 complex and because the Smc5/6 complex itself is targeted by Mms21 for SUMOylation [27] , [28] we tested whether the mms21-11 allele might confer a mild defect in the Smc5/6-Mms21 rogue JM resolvase activity which could explain a synthetic interaction between mms21-11 and sgs1 . Since the JMs of ZMM mutants also depend on the rogue JM resolvases for their resolution [9] , we tested whether mms21-11 is permissive for DNA segregation in a zip1Δ mutant meiosis . zip1Δ is the most permissive ZMM mutant ( presumably providing no substantial Sgs1 mediated ZMM-DSB repair block ) , forming JMs readily from its DSBs and consequently exhibiting a very short prophase 1 delay [4] . Accordingly , while DSB turnover and meiotic progression in the zip3Δ mutant could be enhanced by mms21-11 , the already swift exit of zip1Δ from prophase 1 in SK1 is not improved further by mms21-11 ( Figure 6B ) . However , chromatin segregation is notably affected in the zip1Δ mms21-11 double mutant with 28 . 5% of the tetrads missing DNA in at least one spore and 13 . 5% of the tetrads being completely devoid of DNA in all four spores ( n = 200 ) . Consistent with a reduced number of crossovers , spore viability is decreased from 65% of zip1Δ to 51 . 25% in the double mutant ( n = 80 spores of 20 tetrads ) . Combination with a hypomorphic smc3-myc6 allele , which decreases the viability of zip1Δ to 20 . 0% ( n = 160 spores of 40 tetrads ) , even exaggerates this segregation defect to 52 . 7% of the tetrads lacking DNA in at least one spore and 24 . 3% of the tetrads ( n = 300 ) devoid of DNA in all spores in the resulting zip1Δ smc3-myc6 mms21-11 triple mutant . We conclude that mms21-11 indeed mediates a defect to the rogue JM resolvase function of Smc5/6-Mms21 . Given that already low amounts of unresolved JMs completely block nuclear divisions and that sgs1-mn imposes a rogue JM fate on basically all recombination events [9] , the defect in the rogue JM resolvase function of mms21-11 , although relevant , must be rather weak . Finally , we tested by ChIP-Seq if the localization pattern of the epitope tagged Sgs1-myc18 would lend support to the idea of cooperation between Sgs1 and the Smc5/6-Mms21 complex . Indeed , 70% of the 1000 strongest Sgs1-myc18 peaks map precisely to sites of meiotic DSB formation , as identified by Pan and coworkers [53] at near single nucleotide resolution where they overlap nearly perfectly with peaks of the Smc6 profiles ( Figure 6C ) . However , the intensity of hotspot-matching Sgs1 peaks is not proportional to the activity of corresponding hotspots but proportional to the corresponding Smc6 peaks . Thus the correlation coefficients between the peaks of Sgs1 profiles and their matching DSB sites is low ( cor = . 27 [Sgs1 , 3 . 5h] , cor = . 41 [Sgs1 , 4 . 5h] , Figure 6D ) . Similarly , Smc6 peaks don't correlate well with DSB intensities ( cor = . 26 [Smc6 , 3 . 5h and Smc6 , 4 . 5h] ) despite localizing precisely at hotspots ( Figure 6D ) for most peaks . This suggests that the proteins are not recruited proportionally to the DSB activity . In contrast , peak intensities match very well between Sgs1 and Smc6 ( cor = ∼ . 8 for different profile comparisons , Figure 6D ) . Smc6 and Sgs1 also match on many non-DSB positions including prominent peaks at the centromeres ( figure 2G , S4 , S7A ) . Sgs1 , and to a lesser extent Smc6 also bind to core sites , but in general these signals are much smaller than their signals at DSB sites . These findings show that Smc6 and Sgs1 populate largely the same chromosomal target sites at high resolution and suggest that they are accumulated by some common characteristics at these sites that is different from actual hotspot activity .
Broken chromosomes pose a considerable threat for cells; unrepaired , they cause potentially lethal loss of genetic information , but their repair is equally risky . Inappropriate recombination intermediates and outcomes are inevitable if repair is not carefully controlled . JMs typically arise as HR intermediates with both ends of the DSB lesion engaged with one or more repair templates . Having eliminated the primary lesion and any major stretches of single stranded DNA , JMs do not trigger a DNA damage checkpoint response . Thus , if JMs arise between different chromosomes they mediate dangerous , stable connections due to reciprocal base pairing and catenation which will cause chromosome non-disjunction if not removed in time before anaphase . JMs generated at non-allelic positions are particularly dangerous because their resolution can lead to deletions or translocations and thus need to be avoided in the first place . In many organisms , HR via JMs is strongly down regulated for much of their life cycle , the exception being meiosis where the generation of crossovers via JMs is imperative . Mechanisms and factors that mediate surveillance of such dangerous intermediates or the coordination of the known JM-antagonists are not well known or understood to date . In meiosis , cells are forced to generate high numbers of crossovers and consequently evolved a specialized recombination pathway on top of the mitotic machinery to do so safely . In budding yeast , the conserved ZMM recombination pathway was first described to perform this task [4] , [66] , [67] . It involves a sophisticated machinery which works on a subset of DSBs ( ZMM-DSBs ) to generate appropriate amounts of stable JMs to achieve at least one obligatory CO per bivalent , and it also ensures that COs are formed between the appropriate partners . The ZMM-dHJs are transformed into COs at pachytene exit by a dedicated ZMM resolution machinery , dependent on Exo1-Mlh1/3 [5] . In contrast , non-ZMM DSBs are repaired fast and safely by SDSA to yield NCOs and are thought to support homology search . Recently , BLM/Sgs1 helicase was identified as being central to SDSA mediated NCO formation , as well as in preserving early ZMM recombination intermediates [10]–[12] , [61] . Unregulated or “rogue” JMs that arise from non-ZMM DSBs by escaping destabilization through Sgs1 , or that escape the ZMM pathway , are resolved by a group of rogue JM resolvases: Mus81-Mms4 , Slx1-Slx4 and Yen1 [5] , [9] . Here we demonstrate that the Smc5/6-Mms21 complex specifically binds to sites of meiotic DSBs and plays a dual role in homologous recombination as an antagonist of unregulated JMs and HJ intermediates . It does so by promoting anti-JM formation activity through its Mms21 SUMO E3 ligase and , if inappropriate JMs have already formed , by mediating their resolution through rogue JM resolvases . Similar conclusions have been drawn in an independent study ( accompanying manuscript ) [54] . In particular , that study showed directly that Exo1-Mlh1/3 dependent resolution of ZMM-dHJs does not require Smc5/6-Mms21 . Phenotypes associated with defects in these functions include the recombination dependent failure to separate chromatin during meiosis upon Smc6 inactivation and the concomitant accumulation of persistent JMs that also include a considerable fraction of unresolved IS-JMs and some three-strand JMs indicative of aberrant JM formation . This phenotype resembles the behavior of sgs1-mn mms4-mn mutants [10] , [12] in which most meiotic recombination events give rise to “rogue” non-ZMM JMs and block nuclear divisions , because they depend on the inactivated “rogue JM resolvase” Mus81-Mms4 for resolution . In analogy , this implies that the Smc5/6-Mms21 complex mediates both functions in the management of unregulated JMs . It supports the prevention of rogue JMs as does Blm/Sgs1 and mediates their efficient removal through the rogue JM resolvases . Early defects of Smc5/6-Mms21 mutants were also noted in the accompanying manuscript by Copsey and coworkers [54] , who observed an increase in IS JMs , as well as an increase of Zip3 foci indicating compensation for derailed ZMM intermediates , and high levels of unresolved JMs . This view is corroborated by the role of the SUMO E3 ligase domain of Mms21 . mms21-11 , which lacks this domain , largely separates the two functions of the complex . This allele is defective in the anti rogue JM formation activity of the complex but only mildly affects Smc5/6-Mms21 resolution function . In the mms21-11 mutant , a significant amount of additional JMs are formed and in the background of the meiotic null allele mms4-mn , unresolved JMs accumulate comparable to the smc6-56 mutant ( Figure 3B , E , 4E , I ) . Thus , the SUMO E3 ligase domain of Mms21 is required for the anti rogue JM formation activity of the complex but is largely dispensable for the resolution activity . On the other hand , inactivation of Smc6 severely compromised JM resolution and recombination product formation in an sgs1-mn mutant background ( Figure 5A–E , see also accompanying manuscript for a mutant in another Smc5/6-Mms21 subunit , Nse4 [54] ) , where JM resolution is almost fully dependent on the rogue JM resolvases [5] , [9] . The resolution deficiency in sgs1-mn smc6-56 equals or exceeds that seen in sgs1-mn mms4-mn [9] , [10] , thus resolution activity equivalent to at least the Mus81-Mms4 resolvase must have been lost ( identical conclusion in the accompanying study [54] ) . Providing Smc5/6-Mms21 function after the Ndt80-IN pachytene arrest proves sufficient to ensure meiotic chromosome segregation . This indicates that the lack of early ( destabilizing ) function produced no irreversible damage and can be compensated for by the late ( resolution ) function . This also suggests that late recruitment of the complex to promote resolution is possible . In this way , the Smc5/6-Mms21 complex supports meiotic recombination pathway choice and safeguards the turnover of non-ZMM intermediates , particularly the otherwise unregulated non-ZMM JMs ( Figure 7A ) . These observations imply that the Smc5/6-Mms21 complex recognizes recombinogenic lesions and locally mediates antagonistic activities . How might the Smc5/6-Mms21 complex mediate this activity ? The Smc5/6-Mms21 complex binds to DSB sites , suggesting it might locally recruit and/or orchestrate the function of anti-recombinogenic helicases and of resolvases . The absence of the complex in meiosis renders the activity of Sgs1 insufficient for normal intermediate destabilization and protection and strongly impairs the function of the rogue JM resolvases , supporting the above model . Notably , the defect of smc6-56 and mms21-11 mutants on rogue JM prevention is weaker than that of sgs1-mn leaving the ZMM pathway largely operative and thus CO formation unaffected by Smc6 inactivation . The phenotype of the SUMO E3 domain deletion suggests that Smc5/6-Mms21 mediates its anti-rogue JM formation activity through SUMOylation dependent regulation and coordination of anti-recombinogenic helicase activity at the site of the lesion . Sgs1 and Mph1 are the most likely targets . Interestingly , it has been reported that the FANCM-like helicase Mph1 becomes toxic in the absence of the Mms21 ligase activity [63] . The observed auto-SUMOylation of the complex could also be critical in this function for recruiting or sequestering regulatory targets [27] . In addition , since DNA damage induced SUMOylation of Sgs1 has been reported [40] , Sgs1 may represent a direct substrate for post translational regulation by Smc5/6-Mms21 . However , it was also reported that Mms21 is not essential for this SUMOylation and that different SUMO E3 ligases may provide redundancy [40] , [68] . Consistent with a regulatory role of the Smc5/6-Mms21 complex for anti-recombinogenic helicase function , we found that the recruitment of Sgs1 to meiotic DSBs does require neither Mms21 SUMO E3 ligase activity nor an intact complex since Sgs1-myc18 enrichment at DSB hotspots comparable to wild-type is detected by qChIP in both mms21-11 and smc6-56 mutants ( Figure S8A , B ) . Beyond its supportive role in the prevention of rogue JMs , we find a profound defect of smc6-56 in resolution of rogue JMs in meiosis , implying a direct function of Smc5/6-Mms21 in promoting rogue JM resolvase activity . There are three rogue JM resolvases identified in S . cerevisiae , Mus81-Mms4 , Slx1-Slx4 , and Yen1 . Mus81-Mms4 is the one resolvase which facilitates the removal of the bulk of rogue JMs arising from homologous recombination [5] , [9] . Comparing our results to published data , we estimate that in the smc6-56 mutant resolution activity equivalent to at least the Mus81-Mms4 resolvase must have been lost [9] , [10] . There is no evidence that Mus81-Mms4 is subject to SUMOylation or direct interaction with Smc5/6 . However , the complex may stabilize the HJ and present the substrate DNA to Mus81-Mms4 , or it may mediate processing steps preceding resolution . With the Smc5/6-Mms21 complex , we could identify the second known factor critical for full activity of the Mus81-Mms4 resolvase after the identification of its activating kinase Cdc5 [7] . In contrast to Mus81-Mms4 , direct interaction of the Slx1-Slx4 resolvase with the Smc5/6-Mms21 binding partner Rtt107 was observed , although it is unclear whether Rtt107 binds Slx1-Slx4 and Smc5/6-Mms21 alternatively or simultaneously [38] . Slx4 is also a SUMO substrate and interacts with Rad1-Rad10 , a ssDNA nuclease involved in DNA processing during repair [69] , [70] . Therefore , direct regulation of Slx1-Slx4 and Slx4-Rad1-Rad10 via Smc5/6-Mms21 is possible . However , during meiosis , Slx1-Slx4 only plays a minor role for rogue JM resolution , and also Rad1 did not appear to be required for JM resolution [5] , [9] . Yen1 is the third rogue JM resolvase in budding yeast . Biochemical studies indicate that HJs are its natural substrate without the need for prior processing [71] . Yen1 becomes active around the onset of metaphase 2 and will resolve most rogue JMs still present , even in the absence of Mus81-Mms4 and Slx1-Slx4 [5] , [7] , [9] but not at elevated JM levels as in an sgs1 background . smc6-56 may represent a strong hypomorph for one , two or all rogue JM resolvases , however , the strong JM accumulation seen in sgs1-mn smc6-56 suggests that more than just the function of Mus81-Mms4 is affected . Intriguingly , we find preloading of Smc6 to break sites . This could be a reasonable safety measure when anticipating the programmed generation of hundreds of DSBs . It was reported recently that Rtt107 promotes recruitment of Smc5/6-Mms21 specifically to the HO site upon DSB formation [38] . However , as Rtt107 is neither essential ( like Smc5/6 ) nor required for Smc5/6-Mms21 function , Rtt107 may rather enhance the recruitment of Smc5/6-Mms21 to specific loci in anticipation of DNA damage . Meiotic DSB-hotspots are located in promoters and nuclease sensitive DNA regions [72]–[74] and are frequently associated with chromatin features such as H3K4me3 [75] that may directly , or via Rtt107 , allow for local enrichment of Smc5/6-Mms21 to protect vulnerable DNA regions . Smc5/6-Mms21 represents a ring shaped SMC complex , highly similar to its closely related brethren , cohesin and condensin [28] for which topological mechanisms of function were successfully demonstrated [23] , [76] . While evidence for topological binding of the Smc5/6-Mms21 complex to DNA is still missing to date , its overall high similarity to cohesin and condensin and also previously proposed DNA damage independent loading through the cohesin loaders Scc2-Scc4 to chromatin [39] , make a topological component in the function of Smc5/6-Mms21 likely . If the Smc5/6-Mms21 complex would serve solely as a platform for JM antagonizing factors and their regulators , it would appear inconceivable why the SMC ring should have been maintained throughout evolution . Instead , it is far more likely that a preceding topological function of the evolutionary ancestor SMC complex fulfilled a function rudimentarily similar to Smc5/6-Mms21 , which was ultimately enhanced in the course of evolution . In addition , preemptive loading of Smc5/6-Mms21 to DNA would appear considerably more effective in preventing dangerous lesions if the Smc5/6-Mms21 complex were not statically bound but rather could slide along longer regions of DNA , as proposed for cohesin [77] . How could the Smc5/6-Mms21 complex mediate its function in antagonizing JMs through an underlying topological association with DNA ? We propose the following model to explain the previously observed functions for the Smc5/6-Mms21 complex . The Smc5/6-Mms21 ring , preemptive of DNA damage , is loaded onto dsDNA , entrapping ( in contrast to cohesin ) only one dsDNA molecule . Such topological loading onto a single dsDNA molecule and the previously reported strong ssDNA binding properties of Smc5 and Smc6 [25] , [26] are sufficient to counteract and stabilize recombination intermediates , and can serve to direct the activity of anti-recombinogenic helicases as well as resolvases at ( and to ) the according site , namely , the junction between the HR partners ( Figure 7C , S9A ) . In such a model , Smc5/6-Mms21 could slide freely along the intact dsDNA until encountering the ssDNA/dsDNA interface of an occurring lesion . Such an interface would be present at a progressing D-Loop , or at the HJ of the mature intermediate . Stable binding of the Smc5/6-Mms21 ring at such ssDNA junction sites may limit the spatial freedom for subsequent D-Loop extension , reduce the chance for second end capture , and impede further branch migration of HJs towards detrimental dHJ extension ( as such expanding branch migration would require to overcome one obstructing strand of the HJ topologically ) ( Figure 7C , S9A ) . We also believe that this model could be extended to mitosis and be applied to recombinogenic structures such as cruciform DNA structures arising from replication fork regression . By these means , the Smc5/6-Mms21 complex could survey for recombinogenic lesions , stabilize them through binding and , ultimately , employ counteractive measures by means of helicases and resolvases . As a corollary , this model postulates that Smc5/6-Mms21 mediates its function from outside the lesion , with involvement from only the intact donor DNA molecule being sufficient , to ultimately associate with the very borders of the recombinogenic lesion . Our observations that Smc5/6-Mms21 specifically enriches to sites of DSB formation after break formation but , nonetheless , fails to reveal significant on-top co-localization with the recombinosome marker Rad51 , are consistent with this supposition . In addition , it is to date unknown how anti-recombinogenic helicases like Sgs1-Rmi1-Top3 or Mph1 are directed to mediate their function in the context of a lesion . Association of a helicase to the wrong strand of the recombination junction will , instead of disassembly , result in extension . This problem is also eminent in the function of the Sgs1-Rmi1-Top3 dHJ dissolvase . While hetero-duplex DNA may provide information about the relative position of the parent molecules , it cannot account for directing the dissolution of JMs between perfectly identical sister-chromatids . Since in our model Smc5/6-Mms21 would inherently mark the parent associated “outsides” of a lesion , it may thus provide the lesion with a polarity and direct the anti-recombinogenic helicases for acting in the appropriate orientation ( Figure S9B ) . By these means , the Smc5/6-Mms21 complex could stably mark recombination intermediates and thus orchestrate the activity of anti-recombinogenic helicases and resolvases at the very site where JM antagonizing effectors are needed . While downstream effectors of the complex may vary to suit the specific needs of the individual organism or cell type , we postulate that the underlying JM restraining activity of Smc5/6-Mms21 complex is conserved throughout organisms .
All strains used in this study are derivatives of SK1 . Detailed genotypes are provided in Table S1 . Strains were constructed by crossing or LiAc transformation using standard procedures . The URA3-arg4 recombination reporter was described in [47] . In the meiotic null mutants mms4-mn , sgs1-mn and nse4-mn the respective promoters are replaced by a CLB2 promoter fragment [78] and the estrogen inducible ndt80-IN system has been described [79] . Yeast strains were grown at 30°C in supplemented YPD ( 1% Difco yeast extract , 2% Difco peptone , 2% dextrose , 75 mg/L Ade , 75 mg/L Ura , 75 mg/L Trp ) with exception of the smc6-56 mutants which were grown at 23 . 5°C permissive temperature . All manipulation of strains followed standard procedures . For synchronous sporulation in liquid culture , strains were grown at 30°C overnight in SPS pre-sporulation media ( 0 . 5% Difco yeast extract , 1% Difco peptone , 0 . 17% Difco yeast nitrogen base w/o AA&AS , 1% potassium acetate , 0 . 5% ammonium sulphate , 0 . 05 M potassium-biphthalate , pH 5 . 5 ) to an OD660 of 1 . 1–1 . 3 ( 4×107 cells/ml ) . Meiosis was induced by a subsequent wash and transfer of the cells into pre-warmed supplemented SPM ( 1% potassium acetate , 0 . 001% PPG2000 , 4 mg/L Ura , 4 mg/L Trp , 4 mg/L His , 4 mg/L Arg , 6 mg/L Leu ) at equal volume . Maximum aeration was provided for efficient meiosis . For synchronous meiosis , smc6-56 mutants were allowed to exit mitosis and proceed through most of pre-meiotic S-phase under ( semi- ) permissive conditions with the following temperature regime applied ( after pre-growth at 23 . 5°C ) : At 0 hr SPM , upshift to 26°C , at 1 hr15 min to 28°C , at 1 hr40 min to 30°C , at 2 hr20 min to 33°C . Every temperature shift is translated to the culture media within 10 minutes . We find that smc6-56 is just able to perform nuclear divisions if undergoing meiosis at 30°C as long as not burdened by any additional number of rogue JMs ( data not shown ) . Therefore , we define 28–30°C as semi-permissive temperature for smc6-56 in meiosis . ndt80-IN expression and exit from ndt80-IN pachytene arrest was induced by the addition of β-estradiol ( ED , 5 mM stock in EtOH , stored at −20°C ) to a final concentration of 1 µM at 7 hr30 min SPM , at which point approximately 80–90% of possible JMs had already formed [9] . The separation of early and late functions of Smc5/6-Mms21 in the respective experiments was performed using smc6-56 and the ndt80-IN system . Elimination of the early , pre pachytene-exit function and providing Smc5/6-Mms21 only after pachytene: Normal S-phase upshift to restrictive 33°C and arrest in pachytene followed at 7 hr30 min by addition of ED for Ndt80 induction and linear downshift to 23 . 5°C over 25 minutes . Elimination of the late , post pachytene function of Smc5/6-Mms21: Cells were transferred into SPM at 23 . 5°C and allowed to arrest in pachytene . The temperature was raised to 25 . 5°C from 6 hr45 min to 7 hrs , to 33°C until 7 hr20 min , and ED for ndt80-IN release was added at 7 hr30 min . Preparation of genomic DNA and Southern blotting of one dimensional 0 . 6% agarose gels for physical detection of recombination intermediates was performed as described in detail [47] , [80] , [81] using the CTAB/CoHex/Mg2+ procedure which stabilizes HJ intermediates . Relevant sets of JM experiments were performed in parallel for comparability . For radioactive Southern hybridization , the ArgD probe was used for detection of both maternal and paternal URA3-arg4EcPal/URA3-ARG4 loci and the HisU probe for allele specific detection of the maternal his4::URA3-arg4EcPal locus . Both probes are described [47] . As size marker , lambda HindIII digested DNA was used and probed against . Storage phosphor screens were used for signal detection and scanned with the Bio-Rad Molecular Imager FX . Quantification of signals was performed using Fuji Image Gauge Ver4 . 0 . Frequencies are given as a percentage of total DNA signal . Nuclear divisions were followed on ethanol fixed and DAPI stained whole cells . For spindle staining on whole yeast cells , 1 ml aliquots of culture were fixed in 3 . 2% formaldehyde overnight at 4°C . Cells were washed and cell walls digested at 37°C using 100 µg Zymolyase 20T ( SEIKAGAKU #120491 ) , 70 mM DTT and 1 M sorbitol , in 200 µl . Cells were mounted on poly-L-lysine coated slides and fixed for 3 minutes in ice cold methanol and 10 seconds in ice cold acetone . Cells were blocked and immunostained in 0 . 5% BSA fraction V and 0 . 2% gelatin in 1×PBS . Microtubuli were detected using rat anti-tubulin-alpha ( Serotec MCA78 , 1∶200 ) and FITC-conjugated rabbit anti-rat ( Sigma F1763 , 1∶100 ) antibodies . Vectashield with DAPI ( Vector Laboratories H-1200 ) was used to stain the DNA and stabilize the label . Chromosome surface spreads and immunostainings were performed as described [82] . For antibodies used in this study see Supplemental Information , Table S2 . Evaluation of synapsis followed the criteria used in [83] . Chromosome pairing in zip3Δ mutant surface spread nuclei was evaluated by counting pairs of Hop1 axes of roughly equal length and in close proximity to each other , strictly parallel to each other , or connected via axial associations , or engaged in ( pseudo- ) synapsis . Images were taken on a Zeiss Axioskop fluorescence microscope with a Photometrics CH250 CCD camera using IPLab Spectrum with magnification and exposure times constant . Flow cytometric quantification of cellular DNA content was performed with a BD Biosciences FACSCanto on ethanol fixed cells in the presence of 20 µg/ml propidium iodide in 50 mM Tris pH 7 . 5 . Prior measurement , cells were treated overnight with 2 mg/ml RNaseA in 50 mM Tris 15 mM NaCl pH 7 . 5 at 40°C and one hour with 350 µg Proteinase K at 50°C in 500 µl volume . After brief sonication , propidium iodide was added 15 minutes before performing measurements . 5000 events were counted for each sample . Analysis was performed with Treestar FlowJo v10 . ChIP from meiotic cultures was performed as described by Panizza and coworkers [51] . In brief , 4×109 cells were collected per time point and incubated with para-formaldehyde ( 1% final concentration ) for 15 minutes at 25°C . Cross-linking was stopped by addition of glycine to 131 mM . 4×109 cells were divided into 8 aliquots and separately opened using a multibead shocker ( YASUI-KIKAI , Osaka ) at 2 , 500 rpm , 15 cycles of 30 sec on and 30 sec off , at 4°C . Extracts were sonicated to an average of 200 bp by a Covaris S2 instrument . After removing the cell debris , the supernatant was used for the chromatin immunoprecipitation . For each sample , 50 µl Dynabeads Pan mouse IgG ( Invitrogen ) were incubated with the 9E11 anti-myc ( mouse ) antibody for 6–15 hr at 4°C . The precipitated DNA was used as a template for quantitative real-time PCR using GoTaq qPCR Master Mix for SYBR assay ( Promega ) ( primer sequences: DSB1: CCGCAGAAGCCAACAAACGG , CTTTCGGTGGAACCTCGACC; DSB2: CGTGCCAGATTGAATTTTGA , GAATGGCCTTGGTAGCAAAT; DSB3: ACTTCCAACTGCAGGACGAC , ATCTGGCGGATGAACTTGAG; DSB4: ACGAACAGAGTCCCGAACCT , GCGGTTAATTCGATGGAAAG; CORE1: TGGATGGCAACTGAAGGAGC , TGGAATACCTATGAGTTGACTGC; ADP1: GGTGATGATTGCTCTCTGCC , CGTCACAATTGATCCCTCCC ) . For genome-wide analysis ( ChIP-seq ) , one-tenth of the ChIP DNA was analyzed by real-time PCR ( qChIP ) while the remaining DNA was concentrated by precipitation in ethanol . For the preparation of libraries for Illumina sequencing , we strictly followed the protocols provided by Illumina ( Illumina ChIP-seq DNA sample prep protocol ) . Briefly , DNA-ends were repaired to convert overhangs to phosphorylated blunt ends with T4 DNA polymerase , E . coli DNA Pol I ( Klenow fragment ) , and T4 polynucleotide kinase ( PNK ) . An ‘A’ nucleotide was added to the 3′ end of the blunt ended fragments using Klenow fragment ( 3′ to 5′ exo minus ) . This prepared the DNA fragments for ligation to the adapters ( Truseq Adaptor1-5 ) which have a single ‘T’ base overhang at their 3′ ends . After ligation , excess adaptors were removed by selecting a certain size range from 200–500 bp with QIAquick Gel Extraction Kit ( Qiagen ) . Purified templates were PCR amplified with 15–18 cycles by KAPA-HiFi Hot Start PCR Kit ( KAPA biosystems ) . Before hybridization to the flow cell , the amount and the size of the DNA library was controlled to be at least 1 µg/ul ( for an average fragment size of 300 bp ) . Sequencing was performed at the CSF NGS Unit ( csf . ac . at ) with Illumina HiSeq 2000 resulting in 50 bp single-end reads in multiplex ( Illumina TruSeq adapters 1–5 ) . 2 . 3–8 . 6M reads per sample could be aligned with high confidence to the S . cerevisiae strain S288C , genome version R64-1-1 ( 20110203; http://downloads . yeastgenome . org/sequence/S288C_reference/genome_releases/ ) . We used NextGenMap ( http://cibiv . github . com/NextGenMap ) for fast and sensitive alignment allowing up to 5000 hits per sequence . Read depth per position was generated in Java , dividing each hit by the number of genome-wide matches in case of multi-mapping reads . All subsequent analyses were performed using R ( version 2 . 15 . 2 ) . After summing up the read depth of both strands , gaps were filled up with zeros followed by smoothing with ksmooth ( Nadaraja-Watson kernel ) with bandwidths as indicated and a resolution of 10 bp . Samples were normalized relative to their negative controls using NCIS [84] . This method estimates the fraction of background signal in each sample . After normalization , the untagged control sample was subtracted from each profile for background removal . The profiles remained robust against changes between different negative controls . Peaks were defined as being flanked on both sides by valleys with a minimum depth . The maximal distance between matching peaks from different profiles was set to be equal to the smoothing bandwidth . Peak heights were subsequently compared by pair-wise Pearson correlation ( R , cor ) . The significance of the number of overlapping peaks between profiles was assessed by a hyper-geometric random model . For matching peaks with DSB hotspots , peaks were required to map precisely between the borders of the DSB hotspots . The significance of the number of peaks matching to hotspot regions was assessed by a binomial random model . Here hotspots are defined according to [53] as groups of Spo11-oligo 5′-ends mapping less than 50 bp from each other . To account for the differences in genome versions , the raw data taken from [53] ( GEO accession GSE26449 ) were re-aligned to the same genome version as the other samples , R64-1-1 . Mapped sequence reads and generated profiles for this study are provided at the GEO repository , accession number GSE51977 .
|
Homologous recombination allows repair of DNA breaks from intact templates of identical sequence by a “copy-and-paste” like mechanism . However , the double Holliday Junction ( dHJ ) is a hazardous intermediate that can form during homologous recombination , if single stranded DNA from both ends of a lesion pair with the template . Once the primary lesion is eliminated , the dHJ connects the chromosomes stably and if unresolved can prevent segregation during cell division . In order to prevent chromosome non-disjunction , resolvases will cut any HJ before division . However , genomes contain many multi-copy DNA sequences as transposons or repetitive elements . If dHJs form between such non-allelic loci , cleavage by resolvases can result in chromosome translocations and deletions . Therefore , stabilization of dHJs is sought to be avoided in the first instance by anti-recombinogenic helicases on early intermediates . Analysis of Smc5/6-Mms21 mutants in meiosis revealed that it antagonizes unregulated dHJs both by prevention and resolution . Elimination of the Mms21 SUMO E3-ligase domain leads to inappropriate dHJ accumulation still resolved by Mus81-Mms4 . Disruption of the whole complex results in persistent dHJ accumulation and dysfunction of resolvases , preventing chromatin segregation . These results provide a first unified view on the function of Smc5/6-Mms21 as an antagonist of dangerous dHJs .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2013
|
Smc5/6-Mms21 Prevents and Eliminates Inappropriate Recombination Intermediates in Meiosis
|
An open problem in the field of computational neuroscience is how to link synaptic plasticity to system-level learning . A promising framework in this context is temporal-difference ( TD ) learning . Experimental evidence that supports the hypothesis that the mammalian brain performs temporal-difference learning includes the resemblance of the phasic activity of the midbrain dopaminergic neurons to the TD error and the discovery that cortico-striatal synaptic plasticity is modulated by dopamine . However , as the phasic dopaminergic signal does not reproduce all the properties of the theoretical TD error , it is unclear whether it is capable of driving behavior adaptation in complex tasks . Here , we present a spiking temporal-difference learning model based on the actor-critic architecture . The model dynamically generates a dopaminergic signal with realistic firing rates and exploits this signal to modulate the plasticity of synapses as a third factor . The predictions of our proposed plasticity dynamics are in good agreement with experimental results with respect to dopamine , pre- and post-synaptic activity . An analytical mapping from the parameters of our proposed plasticity dynamics to those of the classical discrete-time TD algorithm reveals that the biological constraints of the dopaminergic signal entail a modified TD algorithm with self-adapting learning parameters and an adapting offset . We show that the neuronal network is able to learn a task with sparse positive rewards as fast as the corresponding classical discrete-time TD algorithm . However , the performance of the neuronal network is impaired with respect to the traditional algorithm on a task with both positive and negative rewards and breaks down entirely on a task with purely negative rewards . Our model demonstrates that the asymmetry of a realistic dopaminergic signal enables TD learning when learning is driven by positive rewards but not when driven by negative rewards .
In this article we focus on a specific variant of TD learning: the TD algorithm as implemented by the actor-critic architecture [36] . Here , we summarize the basic principles; a thorough introduction can be found in [1] . The goal of a TD learning agent , as for every reinforcement learning agent , is to maximize the accumulated reward it receives over time . The actor-critic architecture ( see Fig . 1 ) achieves this goal by making use of two modules , the actor and the critic . The actor module learns a policy , which gives the probability of selecting an action in a state . A common method of defining a policy is given by the Gibbs softmax distribution:where is known as the preference of action in state and the index runs over all possible actions in state . The critic evaluates the consequences of the actor module's chosen actions with respect to a value function . Once learning has reached equilibrium , the value function is the expected summed discounted future reward when starting from state and following policy . During the learning process only estimates of the actual value function are available . The performance of the agent on a task is improved by making successive updates to the policy and the value function . These updates are usually formulated assuming a discretization of time and space: an error term is calculated based on the difference in estimations of the value function when moving from one discrete state to the next , : ( 1 ) where is the reward the agent receives when moving into state and is a discount factor . This error signal , known as the TD error , is positive if the reward is greater than the expected discounted difference between and , indicating that the estimate of needs to be increased . Conversely , is negative if the reward is less than the expected discounted difference , indicating that the estimate of needs to be decreased . In the simplest version of TD learning , known as the TD ( ) algorithm , the critic improves its estimate of as follows: ( 2 ) where is a small positive step-size parameter . For a given policy and a sufficiently small , the TD learning algorithm converges with probability [37] , [38] . Additionally , the preference of the chosen action in state is adjusted to make the selection of this action correspondingly more or less likely the next time the agent visits that state . One possibility to update the preference in the actor-critic architecture is given by: ( 3 ) where is another small step-size parameter . For the purposes of this manuscript , we shall refer to the calculation of the error signal and the update of value function and policy described above as the classical discrete-time TD ( ) algorithm .
Fig . 2 illustrates the architecture of our actor-critic spiking network model implementing temporal-difference learning ( see Introduction ) . All neurons in the network are represented by current-based integrate-and-fire neurons with alpha shaped post-synaptic currents . A tabular description of our model and its neuronal , synaptic and external stimulation parameters are given in Methods . The network interacts with an environment , which is implemented purely algorithmically for the purpose of this work . The input layer of the neural network represents the cortex; it encodes information about states , each represented by a population of neurons . The environment stimulates the population associated with the current state of the agent with a constant DC input , causing the neurons to fire with a mean rate of ; in the inactivated state the neurons fire on average with . The low background rate in the inactivated state is chosen for the sake of simplicity in developing the synaptic plasticity dynamics , but is not a critical assumption of the model ( see section “Synaptic-plasticity” ) . Each population in the cortex projects to the actor and critic modules . As the focus of our study is the consequences of a realistic dopaminergic signal for temporal-difference learning rather than action selection , we keep the actor model as simple as possible . As in previous models [20] , [34] , [39] , the actor module consists of actor neurons , each corresponding to one action . The synaptic weights between the cortical and the actor neurons represent the policy in our model . Whichever action neuron fires first in response to the activation of the state neurons is interpreted by the environment as the chosen action ( for a review of first-spike coding , see [40] ) . Immediately after an action has been chosen , i . e . after an actor neuron has spiked , the environment deactivates the previous state neurons and activates the neurons representing the new state resulting from the chosen action . At the same time the environment inhibits the actor neurons for a short time period , during which no further action can be chosen , allowing the cortical signal from a newly entered state to build up . For more sophisticated approaches to the action selection problem , see [41] , [42] . Two key experimentally observed features of the activity of the dopaminergic neurons are a constant low background rate with phasic activity with asymmetric amplitude depending on whether a reward is given or withheld [2] . As the basal ganglia dynamics generating this signal is unknown , we select the simplest possible network that generates these features; in general , multiple network configurations can produce the same dynamics [43] . We adapt the circuit model proposed in [18] to perform the role of the critic module , which is responsible for generating a temporal-difference error . The major model assumption here is that the weights of the synapses between the neurons representing a given state and the critic module encode the value of that state . The circuit connects a population of neurons representing the striatum , the input layer of the basal ganglia , to a population of dopaminergic neurons directly and also indirectly via a population of neurons representing the ventral pallidum . The direct and indirect pathways are both inhibitory . Consequently , the synaptic input from the striatum via the indirect pathway has a net excitatory effect , whereas the delayed striatal synaptic input via the direct pathway has an inhibitory effect on the dopamine neurons . This results in a phasic increase if the agent moves from a state with low cortico-striatal synaptic weights to a state with high weights ( see Fig . 3 ) and a phasic decrease if the agent moves from a state with high cortico-striatal synaptic weights to a state with low weights . The length of the phasic activation is determined by the difference in the delays of the direct pathway and the indirect one . We have chosen and which results in a duration of the phasic activation similar to that observed experimentally ( see Fig . 1 in [2] ) . If the agent enters a rewarded state , the dopamine neurons receive an additional DC stimulation from the environment starting after the agent moves and lasting for the duration of the phasic activity , . Assuming the cortico-striatal synaptic weights represent the value function , after each state transition the dopamine neurons integrate information about the current value function with a positive sign , information about the previous value function with a negative sign , and a reward signal . Thus all the information necessary to calculate a form of temporal-difference error is present ( see Eq . ( 1 ) ) . The dopaminergic neurons project back and release dopamine into the extracellular space ( Fig . 2 purple arrows ) which modulates as a third factor the plasticity of the synapses between the cortex and the striatum and between the cortex and the actor neurons . Later in this section we develop synaptic plasticity models using a top-down approach to implement TD learning . In this section we show that our network is able to generate dopaminergic activity with realistic firing rates and discuss its similarities to , and differences from , the classical discrete-time algorithmic definition of the TD error signal given in Eq . ( 1 ) . It has been found that dopamine neurons fire with a low constant baseline activity ( approx . in rats [44] , [45] and in monkeys [46] ) as long as nothing unpredicted happens . This is known as the tonic activity of the dopaminergic neurons . For our model , this implies that the baseline firing rate should be independent of the strength of the cortical-striatal synapses associated with each state . This condition can be fulfilled in our architecture for an infinite number of configurations assuming linear relationships between the firing rates of the neurons in the striatum and the ventral pallidum; for a derivation of these relationships , see Supplementary Text S1 . We select the simplest rate relationship with a linear coefficient of one . This relationship generates a constant baseline activity when and the synaptic weights connecting the striatum to the dopamine neurons are equal in strength to the synaptic weights between the ventral pallidum and the dopamine neurons . For the parameters given in Methods the mean dopaminergic baseline firing rate in our network is approx . , which is close to the experimentally observed stationary dopaminergic firing rate . When the agent transits from one state to another , the dopamine neurons exhibit phasic activity lasting for around in accordance with durations found experimentally [47] , [48] , see Fig . 3 . Fig . 4 shows the amplitude of phasic activity of the dopaminergic neurons after the agent transits from state to state in dependence of the difference in the corresponding cortico-striatal synaptic weights . In accordance with experimental observation [46] the dopamine neurons show a continuum of firing rates lower than the baseline for outcomes that are worse than predicted ( ) and higher than the baseline for outcomes better than expected ( ) . Likewise , entering a state with an unpredicted reward induces a phasic increase of activity . The amplitude of the phasic activity of the dopaminergic neurons therefore has similar properties to the algorithmic TD error signal given in Eq . ( 1 ) . However , the properties of the dopaminergic signal deviate from the TD error in the following points: Point 2 arises due to the nonlinearities inherent in spiking neuronal networks , particularly at low rates ( for a recent account see [49] ) . If a linear rate-based model was assumed , the amplitude of the phasic response would also vary linearly until an amplitude of was reached for some negative value of . Similarly , the addition of the reward signal could only affect the offset of the curve in a linear rate-based model ( point 3 ) . A nonlinear rate-based model may well be able to capture these features , but in order to make the correct non-linear assumptions , the behavior of the system to be abstracted needs to be known first . A nonlinear dependence of the dopaminergic firing rate on the reward prediction error has recently also been observed experimentally [46] . As we show in the next subsection , point 4 can be compensated by introducing a discount factor in the synaptic plasticity dynamics . A -discounted difference can also be obtained if the dopaminergic rate is assumed to depend on the striatal firing rate . As this is not in accordance with experimental findings we do not make this assumption , however , a derivation of the relationship between the firing rates and is derived in Supplementary Text S1 . In order for the network model to realize TD learning , the right synapses have to undergo the right changes in strength at the right time; this is also known as the credit assignment problem [1] . Here , we derive synaptic plasticity dynamics in a top-down fashion for the cortico-striatal synapses and the synapses between the cortical populations and the actor module representing the value function and the policy respectively . In the classical TD algorithm , when the agent transits from state into state , only the value and preference of the most recently exited state are updated ( see Eq . ( 2 ) and Eq . ( 3 ) ) . For a synapse to implement this feature it requires a mechanism that enables plasticity for a short time period after the agent has left the state associated with the pre-synaptic neuron . This situation is characterized by the pre-synaptic rate being initially high and then dropping , as the population of cortical neurons associated with a state is strongly stimulated when the agent is in that state and weakly stimulated otherwise . An appropriate dynamics can be constructed if the synapse maintains two dynamic variables driven by the spikes of the pre-synaptic neuron as originally proposed in [34]: a pre-synaptic activity trace and a pre-synaptic efficacy trace : ( 4 ) ( 5 ) where denotes the th spike of the pre-synaptic neuron . The pre-synaptic activity trace is an approximation of the pre-synaptic firing rate; it is incremented at every pre-synaptic spike and decays to with a time constant ( see top panel of Fig . 5 ) . To restrict the plasticity to the period immediately following a state transition , we assume a value of such that the activity trace decays to zero before the agent performs a further state transition . Efficacy traces as defined in Eq . ( 5 ) have previously been postulated as part of a spike-timing dependent plasticity model that accounts for data obtained from triplet and quadruplet spike protocols [50] . The efficacy trace is set to at every pre-synaptic spike and relaxes exponentially to with a time constant ( Fig . 5 , middle panel ) . This time constant is assumed to be large such that is small in the presence of pre-synaptic activity . When the agent is in the state associated with neuron , is high and is close to zero . When the agent leaves the state , relaxes to and relaxes to . A product of the two traces is therefore close to at all times except for the period shortly after the agent leaves the state associated with neuron ( Fig . 5 , bottom panel ) . Therefore , a synaptic plasticity dynamics proportional to ensures that the right synapses are sensitive to modifications at the right time to implement TD learning . This simple relationship only holds for a very low rate in the inactive state . If the firing rate of cortical neurons in the inactive state were higher , then the product would be non-negligible at all times , resulting in permanent sensitivity of the synapse to irrelevant fluctations in the dopamine signal . Of course , this could be compensated for without altering the functionality by requiring to exceed a threshold , or by adopting a triphasic approach based on successive pre-synaptic activity thresholds as in our earlier work [34] . The low rate therefore does not constitute a requirement for our model . However , to avoid additional factors in the plasticity dynamics , we prefer to keep the rate relationships as simple as possible . In TD learning the value function and the policy are both updated proportionally to the TD error ( see Eq . ( 2 ) and Eq . ( 3 ) ) which in our network model is signalled by the deviation of the dopaminergic firing rate from its baseline . For the sake of simplicity we model the dopamine concentration as the superposition of the activity traces of all dopaminergic neurons: ( 6 ) where is the th dopamine spike and is a time constant . This simplified model captures the experimentally observed feature that the concentration of dopamine is dependent on the firing times of the dopaminergic neurons [51] , [52] . Moreover , we set in agreement with experimental findings on the dopamine uptake time in the striatum [51] . A more sophisticated approach to modelling the extracellular dopamine concentration can be found in [52] . A suitable synaptic plasticity dynamics between a cortical neuron and a striatal neuron to implement value function updates is therefore given by: ( 7 ) where is the baseline concentration of dopamine and is a learning rate parameter . As discussed in the previous section , one difference between the dopaminergic signal as generated by our network model and the TD error is that the dopaminergic firing rate depends on the total value of the current state , rather than the -discounted value ( compare Eq . ( 2 ) ) . However , it is possible to compensate for this discrepancy in the following way . The firing rate of the striatum population expresses the value of the current state , as the value function is encoded by the cortico-striatal synaptic weights . For a given cortico-striatal synapse , the current state value can therefore be approximated by a post-synaptic activity trace as defined in Eq . ( 4 ) with a time constant , which can be chosen quite arbitrarily . We therefore include a term in Eq . ( 7 ) proportional to the post-synaptic activity trace : ( 8 ) where . In our numerical simulations we assume a plasticity dynamics at the cortico-striatal synapses as given by Eq . ( 8 ) . During the short period after a transition from to , the cortico-striatal synapses associated with state are sensitive to modification . As discussed in the previous section , the dopaminergic signal depends nonlinearly on successive reward predictions encoded in the cortico-striatal synaptic weights , whereas the TD error is a linear function on the value function of successive states . Furthermore the slope of the non-linear function depends on the magnitude of any external reward . This means that it is not possible to define a single mapping from the units of synaptic weights to the units of the value function that holds for all values of and all rewards , as in our previous study [34] . However , it is possible to generate a piecewise mapping by approximating the nonlinear function for a given reward signal in a given range of by a linear function . The mapping ( Eq . ( 11 ) ) is derived in detail in the Supplementary Text S2 and consists of two steps . First , the synaptic plasticity dynamics is integrated to calculate the net change in the mean outgoing synaptic weight of the neurons associated with a state when the agent moves from to . Second , the net weight change is converted from units of synaptic weight to units of the value function according to the linear relationships: ( 9 ) ( 10 ) where is a proportionality parameter mapping the mean striatal firing rate to the units of the value function and is a proportionality factor mapping the mean cortico-striatal weights of a state to the mean striatal firing rate . For our choice of parameters ( see Methods ) Eq . ( 10 ) is fulfilled in the allowed range for the cortico-striatal weights with and . Within a given range of , the mean net weight change of the synapses immediately after transition out of corresponds to a slightly modified version of the classical discrete-time value function update with an additional offset : ( 11 ) The learning parameters and of the equivalent TD ( ) algorithm and the offset depend on the synaptic parameters and as defined above . They additionally depend on the slope and intercept of the linear approximation of the dopaminergic signal: ( 12 ) The constants depend on the synaptic time constants; see Supplementary Text S2 for the definitions . Because and are dependent on the range of and the direct current applied to the dopamine neurons , the weight update can be interpreted as a TD learning value function update with self-adapting learning parameters and a self-adapting offset that depend on the current weight change and reward . The greater the difference between the mean synaptic weights of successive states , the higher the learning rate and discount factor . For the parameters used in our simulations , a range of can be realized by a range of . A choice of results in a discount factor . For a specific choice of , the learning rate can be determined by the synaptic parameter . For , the range can be realized by the range . As and can be chosen independently , all possible combinations of and can be realized . If the current state is rewarded , the offset is a -dependent analog to the reward in the TD error Eq . ( 1 ) . Otherwise , for an appropriate choice of parameters ( see Methods ) is always smaller than and has no analog in classical TD learning . Self-adjusting parameters have also been implemented in other three-factor learning rules such as the one in [53] based on the meta-learning algorithm proposed in [54] . In contrast to meta-learning , in our model the values of the parameters do not adjust themselves to optimal parameters for a given task but vary according to the difference between the estimated values of successive states , , and the current reward value . The range of possible learning parameters for a given and reward value depends on the current choice of synaptic parameters and , which can be set arbitrarily . However , meta-learning could be an additional mechanism that adjusts the parameters and to optimal values for a given task . The variable parameters suggest a similarity with value learning , another learning algorithm similar to TD but with a variable discount rate [55] . However , in value learning the discount rate changes over time: it is lowest immediately after an unconditioned stimulus and increases in between them , making the algorithm more sensitive to long term rewards . In our model the learning parameters do not depend on time but on the current reward and the difference in successive reward predictions encoded by . Similarly to the update of the value function , in the classical discrete-time TD algorithm only the policy for the recently vacated state is updated . As described earlier in this section , in the neuronal architecture an action is chosen by the first spike of an action neuron . Therefore an appropriate plasticity dynamics for the synapse between a cortex neuron and an actor neuron is given by ( 13 ) where determines the learning speed , and is a post-synaptic activity trace as defined in Eq . ( 4 ) with time constant . The choice of post-synaptic time constant is not critical , but the activity trace should decay within the typical time an agent spends in a state in order to be selective for the most recently chosen action . Unlike the cortico-striatal synapses described above , the lack of -discounting in the dopamine signal cannot be compensated for by the addition of an additional local term in the synaptic plasticity dynamics . This is due to the fact that the post-synaptic activity here represents whether the encoded action was selected rather than the value function of the next state as in the previous case . Information about the value of the new state could only arrive at the synapse through an additional non-local mechanism . In order to ensure the agent continues to occasionally explore alternative directions to its preferred direction in any given state , we restrict the synaptic weights of the synapses between the cortex and the actor neurons to the range . This results in a maximal probability of and a minimal probability of for any movement direction in any state ( see Supplementary Text S2 for a mapping of synaptic weights to probabilities ) . The parameters for synaptic plasticity models used in our study are summarized in Methods . The proposed cortico-striatal synaptic plasticity dynamics Eq . ( 8 ) depends on three factors: the pre-synaptic firing rate , the post-synaptic firing rate and the dopamine concentration . For cortico-striatal synapses the effect on the plasticity of each of these factors has experimentally been studied in vivo and in vitro ( see [9] for a review ) . The long-term effects found on average across studies are summarized in column six of Table 1 . These results show that in order to induce any long lasting changes in synaptic plasticity , a conjunction of pre- and post-synaptic activity is required . Early studies on the effect of conjoined pre-synaptic and post-synaptic activity on the cortico-striatal plasticity reported exclusively long term depression ( LTD ) . More recent studies have shown that long term potentiation ( LTP ) can also be obtained under some circumstances . The expression of LTP or LTD seems to depend on methodological factors such as the age of the animal , the location of the neuron and the stimulating electrode and the stimulus parameters [9] . Although in these studies it is assumed that dopamine is not involved , it cannot be ruled out as cortico-striatal high frequency stimulation causes dopamine release [56] . The main findings resulting from studies involving all three factors can be summarized in the following three-factor rule [57]: under normal and low dopamine concentrations , the conjunction of pre- and post-synaptic activity leads to LTD , whereas a large phasic increase in dopamine concentration during pre- and post-synaptic activity results in LTP . The predictions of the cortico-striatal synaptic dynamics given by Eq . ( 8 ) for the various permutations of pre- and post-synaptic activity and dopamine concentration are summarized in column ( for , corresponding to ) and column ( for , corresponding to ) of Table 1 . We assume that a value of in the first three columns denotes recent activity; due to the time constants of the activity traces this activation is still perceptible from the point of view of the synapse and can thus be assumed to have an active influence on plasticity . Assuming the baseline dopamine concentration only changes on a long time scale , experiments involving no particular manipulations of the dopamine concentration ( denoted by in Table 1 ) will be characterized by . The plasticity dynamics Eq . ( 8 ) predicts LTD for an active influence of pre- and post-synaptic activity , and in accordance with the majority of the experimental findings; for no change in synaptic strength is predicted . Furthermore , Eq . ( 8 ) predicts that for simultaneous influence of pre- and post-synaptic activity , the direction of the synaptic change depends on the concentration of dopamine . For normal ( ) as well as low dopamine concentration results in LTD ( see Fig . 6 ) , while a large phasic increase in the dopamine concentration results in LTP . For the change from LTD to LTP occurs at , resulting in no change in synaptic strength under normal dopamine concentration in contrast to the experimental findings . The theoretical model makes additional predictions in this case that go beyond the presence or absence of activity and the direction of change . Due to the timing sensitivity of the plasticity dynamics given in Eq . ( 8 ) , a weak synaptic weight change is predicted if the activity of the pre-synaptic neuron overlaps with the activity of the post-synaptic neuron in the presence of dopamine and a strong change if the pre-synaptic activity precedes the post-synaptic activity . Such a dependency on timing involving extended periods of activation have so far not been tested experimentally . However , protocols involving individual spike pairs have revealed comparable effects; for a review , see [58] . The greatest difference between our predictions and the experimental findings is that a simultaneously active influence of pre-synaptic activity and dopamine is sufficient to induce LTD or LTP in the absence of post-synaptic activity . However , this is quite an artificial case as pre-synaptic activity always generates post-synaptic activity in our network model dynamics . The behavior of the model could be brought into better alignment with the experimental data by adding additional complexity . For example , a multiplicative Heaviside function that evaluates to one when the post-synaptic activity exceeds a certain threshold would eliminate the generation of LTP/LTD in the absence of post-synaptic activity without altering the functionality of our model . As the plasticity dynamics was derived to fulfil a particular computational function rather than to provide a phenomenological fit to the experimental data , we prefer to avoid this additional complexity . Apart from this case , our predictions on the direction of cortico-striatal plasticity under the active conjunction of pre- and post-synaptic activity for are in good agreement with experimental findings . As in our previous study [34] , we tested the learning capability of our neuronal network model on a grid-world task , a standard task for TD learning algorithms . In our variant of this task , the grid consists of states arranged in a five by five grid ( see inset of Fig . 7 ) . The agent can choose between four different actions ( south , north , east , west ) represented by actor neurons . If the agent chooses an action which would lead outside the grid world , the action does not lead to a change in its position . Only a single state is rewarded; when the agent enters it a direct current with amplitude is applied to the dopaminergic neurons corresponding to the real-valued reward sent to the critic module in a classical discrete-time TD algorithm ( see Introduction ) . After the agent has found the reward and selected a new action , it is moved to a new starting position that is chosen randomly and independently of the selected action . This is therefore a continuing task rather than an episodic task , as there are no terminal states . To maximize its reward , the agent must find the reward from random starting positions in as few steps as possible . The difficulty of the task is that the agent has to learn a series of several actions starting from each state in which only the last one results in a reward . The grid world task is useful to visualize the behavior of a learning algorithm but is not intended to represent physical navigation task , as spatial information is not taken into consideration ( e . g . exploiting the knowledge of which states are neighbors ) . To evaluate the performance of our model on the grid-world task , we separate the ongoing sequence of states and actions into trials , where a trial is defined as the period between the agent being placed in a starting position and the agent reaching the reward state . We measure the latency for each trial , i . e . the difference between the number of steps the agent takes to reach the reward state and the minimum number of steps required to reach the reward state for the given starting position . To provide a comparison , we also measure the performance of a classical discrete-time TD learning algorithmic implementation with corresponding parameters . The specification of the discrete-time implementation is obtained by mapping the synaptic parameters to the discrete-time parameters for and determining the corresponding reward via a search algorithm ( see Supplementary Text S2 ) . Fig . 7 shows the evolution of latency on the grid-world task for the neuronal network model with optimized parameters and the discrete-time algorithmic implementation with corresponding parameters . Within the first trials the latency of the neuronal network model drops from around steps to steps . After trials the agent has learnt the task; the latency is always below steps . The learning speed and the equilibrium performance of the neuronal network model are as good as those of the corresponding discrete-time algorithmic implementation . The performance of the discrete-time algorithmic implementation does not deteriorate if a discount factor is assumed for the updates to the policy in correspondence with the synaptic plasticity dynamics given by Eq . ( 13 ) ( data not shown ) . As discussed in section “Synaptic-plasticity” , we impose hard bounds on the weights of the synapses between the cortex and the actor to ensure that for a given state , no action becomes either impossible or certain . For this task , it turns out that the lower bound is not necessary; restricting the weights to the range results in a similar learning performance ( data not shown ) . However , the upper bound is necessary for the stability of the system . In the absence of an upper bound , synaptic weights between the cortex and all action neurons other than south increase to unbiological levels . This runaway behavior is detrimental to the learning process; in the agent only locates the rewarded state times , a factor of fewer than for the bounded learning agent . In our model , all cortico-striatal synaptic weights as well as all synaptic weights between the cortex and the actor neurons are initialized with the same value . This corresponds to all states being estimated at the same value and all possible directions of movement from each state being equally preferred . Fig . 8A shows the value function encoded in the mean synaptic cortico-striatal weights associated with each state after the task has been learnt . A gradient towards the rewarded state can be seen , showing that the agent has learnt to correctly evaluate the states with respect to their reward proximity . In order to represent the policy , we mapped the synaptic weights between cortex and actor neurons to the probabilities of choosing each action ( see Supplementary Text S2 ) . Fig . 8B shows the preferred direction in a given state after the task has been learnt indicated by the arrows . The x-component of an arrow in a state gives the difference between the probabilities of choosing east and west , the y-component the difference between the probabilities of choosing north and south:After the task has been learnt the agent tends to choose actions that move it closer to the rewarded state . These results show that not only can our model perform the TD ( ) algorithm , but that its parameters can be successfully mapped to an equivalent classical discrete-time implementation . Despite the inherent noisiness of the neuronal network implementation , it learns as quickly and as well as a traditional algorithmic implementation . In the previous section we demonstrated the ability of the spiking neuronal network model to solve a reinforcement learning problem with sparse positive reward . However , due to the asymmetry of the dopaminergic signal , it is to be expected that differences between the neuronal network model and a standard TD learning algorithm become more apparent in tasks where learning is driven by negative rewards . In this section we study the learning performance of the spiking neuronal network model in tasks with negative rewards and investigate the consequences of the modified TD learning algorithm implemented by the neuronal network . An appropriate task to discriminate between the standard and the modified TD algorithms is the cliff-walk task [1] . In our version of this task , the cliff-walk environment consists of 25 states with five special states: a start state in the lower left , a goal state in the lower right and three cliff states in between the start and the goal state ( see Fig . 9A ) . When the agent moves into a cliff state ( i . e . falls off the cliff ) a negative direct current with amplitude is applied to the dopaminergic neurons , corresponding to a negative reward value in a traditional TD learning algorithm . In the cliff states and the goal state , the agent is sent back to the start state regardless of the next action selected . As before , we treat the task as a continuous one , i . e . the synaptic weights representing the value function and the policy are continuously updated , even when the agent is sent back to the start state . In a first variant of this task , a smaller negative direct current is applied to the dopamine neurons in all non-cliff states except the start and goal states , where the reward is zero . Thus , the agent only receives negative rewards from the environment . Setting and corresponds to setting a negative reward of in the cliff states and in all other states except the start and goal states for the discrete-time algorithmic TD ( ) agent . Fig . 9B shows the total reward received by the neuronal agent and the traditional algorithmic agent , summed in bins of and averaged over runs . All parameters are set as for the grid-world task . The traditional TD learning agent improves its performance rapidly . After approx . the average reward over is always above . The performance continues to improve up to , after which the average reward saturates at around . Unlike the grid-world task , the neuronal agent does not improve its performance even after . During this time the neuronal agent reaches the goal state on average only 30 times . In the same period the traditional agent reaches the goal state on average more than 700 times . Similarly , the average number of times the neuronal agent falls off the cliff is around 660 , whereas the traditional agent makes this mistake on average less than times . These results demonstrate that although the neuronal agent performs as well as the traditional discrete-time agent on the grid-world task , the traditional agent can learn the cliff-walk task with purely negative rewards and the neuronal agent cannot . This is due to the fact that the true underlying optimal value function is negative for this variant of the task , as the expected future rewards are negative . Thus , the synaptic weights representing the value function all reach their minimal allowed values and do not allow the agent to distinguish between states with respect to their reward proximity . In a second variant of this task the agent receives a positive reward in the form of a direct current with amplitude applied to the dopaminergic neurons when it reaches the goal state . The reward in all other states except the cliff and goal states is zero . For the purposes of analysis , the end of a trial is defined by the agent reaching the goal state , regardless of the number of times it falls off the cliff and is sent back to the start state . Fig . 10A shows the development of the latency on the cliff-walk task for the neuronal network model and the discrete-time algorithmic implementation , both with the same parameters used in the grid-world task . The cliff-walk task can be learnt much faster than the grid-world task , as the start state is not randomized , so the agent only needs to learn a good policy for the states around the cliff and the goal . The neuronal network model learns the task more slowly than the discrete-time algorithmic implementation , requiring around trials and trials , respectively . The average latency after learning is slightly higher for the traditional agent ( approx . 3 ) than for the neuronal agent ( approx . 2 . 3 ) . However , this does not mean that the neuronal agent has learned a better strategy for the task , as can be seen in the average total reward per trial shown in Fig . 10B . For the traditional algorithm , the summed reward after learning is equal to the reward of the goal state in almost every trial , demonstrating that the agent has learnt to completely avoid the cliff . The average reward received by the neuronal agent deviates much more frequently from the maximum , which shows that the neuronal agent still selects actions that cause it to fall off the cliff . As for the grid-world task , it turns out that the upper bound on the weights of the synapses between the cortex and the actor neurons is necessary for the stability of the system but the lower bound is not . In the absence of an upper bound , the agent still initially learns the task within about trials . However , the synaptic weights increase to unbiologically high values after approximately trials , which causes the task to be unlearned again . In contrast , the absence of a lower bound on the synaptic weights does not affect the learning performance ( data not shown ) . The differences in the behavior learned by the traditional and neuronal agents are also evident in Fig . 11 , which shows for one run the relative frequencies with which each state is visited after the performance has reached equilibrium . For this purpose , we assume an agent to have reached equilibrium performance once it has visited states . While the traditional agent ( Fig . 11B ) has learnt to avoid the cliff altogether and chooses a safe path one row away from the cliff , the neuronal agent ( Fig . 11A ) typically moves directly along the edge of the cliff and in some trials falls off it . The differences in the strategies learned by the traditional and the neuronal agents account for the finding that the neuronal agent exhibits a shorter average latency but a lower average reward per trial than the traditional discrete-time TD ( ) agent . As discussed in section “Synaptic-plasticity” and derived in detail in the Supplementary Text S2 , the neuronal network implements a modified TD learning algorithm with self-adapting learning parameters and , and a self-adapting additional offset ( see Eq . ( 11 ) and Eq . ( 12 ) ) . Furthermore , a discount factor is only present in the value function update and not in the policy update . Another constraint of the neuronal network is that there is a natural lower bound in the dopaminergic firing rate , so there is a limited representation of negative temporal-difference errors . Similarly , the synaptic weights encoding the value function and the policy have lower bounds and are thus limited in their ability to encode negative values for states . To analyze the consequences of these modifications from the traditional learning method , we implement modified versions of the traditional discrete-time TD learning algorithm incorporating the various modifications present in the neuronal network model . The learned strategies are visualized in Fig . 11C–H . In all variants as well as in the original discrete-time TD learning algorithm , we restrict the maximal and the minimal values for the action preferences to the range . This results in the same maximum probability of choosing an action as given in the neuronal network by the bounds on the synaptic weights representing the policy . In all versions the parameters are set according to our derived mapping; the units of the synaptic weights are mapped into the units of the value function according to Eq . ( 9 ) for and . In the first version , a lower bound is applied to the TD error , thus limiting the system's ability to express that an action led to a much worse state than expected ( Fig . 11C ) . In the second version the values of the value function are bounded to a minimal value function of and a maximal value function of ( Fig . 11D ) . Neither version results in a different strategy on the cliff-walk task from that learned by the traditional algorithm without modifications ( Fig . 11B ) . A minor difference can be seen for the third version ( Fig . 11E ) , which applies a discount factor to the updates of the value function but not to those of the policy . We can therefore conclude that none of these modifications in isolation substantially alters the strategy learned for the cliff-walk task by the traditional TD ( ) algorithm . The fourth version incorporates self-adapting learning parameters and an additional self-adapting offset in the TD error as given by Eq . ( 11 ) and Eq . ( 12 ) . The mapping results in the following parameter sets for different external reward values: , and for the goal state , , and for the cliff states and , and for all other states . This modification results in a strategy that is much more similar to that developed by the neuronal system , in that the agent typically walks directly along the edge of the cliff ( Fig . 11F ) . Unlike the neuronal system , the modified TD ( ) algorithm does not select actions that cause it to fall off the cliff . This can be clearly seen as the cliff states are not visited at all and all the states on the path are equally bright , indicating that the agent is only returned to the start state at the successful end of a trial . The key component of the modification is likely to be the additional offset: a similar strategy is learned by the traditional TD learning agent in an altered version of the cliff-walk task , in which each state other than the goal and the cliff states is associated with a negative reward equivalent to the offset ( data not shown ) . By combining the modifications , the strategy of the neuronal agent is recovered . Fig . 11G shows the strategy learned by a TD learning algorithm with self-adapting learning parameters and offset and with the value function restricted to the range . In this case , the agent mostly chooses the path closest to the edge of the cliff , but occasionally selects actions that cause it to fall off . Additionally enforcing a lower bound on the TD error and applying the -discount to the value function updates only do not cause any further alterations to the learned strategy ( Fig . 11H ) . These results show that whereas the neuronal agent cannot learn a task with purely negative rewards , it can learn a task where external negative rewards are applied when the underlying optimal value function is positive . However , even in this case the neuronal agent learns more slowly than a traditional agent and attains an inferior equilibrium performance . For the cliff-walk task , it is the self-adapting parameters and additional offset which contribute the most to the difference in the strategies learned by the neuronal and traditional agents . The bounds imposed on the value function in the modified TD algorithm contribute second most , whereas the lower bound on the TD error and the absence of -discounting on the policy updates do not play major roles .
The learning speed and performance of the neuronal network on the grid-world task with sparse positive reward are comparable to that of a discrete-time actor-critic TD learning implementation . In some respects this result is not surprising , as the plasticity dynamics were designed to fulfill the main properties of TD learning: value function and policy updates are proportional to the TD error and modifications are applied only with respect to the most recently exited state and the most recently chosen action . However , the dopaminergic signal does not perfectly reproduce the characteristics of the algorithmic TD error signal . The amplitude of the phasic activity is a nonlinear function of the difference in value between two states , and the dynamic range for negative errors is small . Moreover , synapses are not only updated due the presence of an error signal , but also due to small fluctuations of the dopaminergic firing rate around the baseline firing rate . Finally , the timing condition given by the product of the pre-synaptic efficacy and the pre-synaptic activity trace is not as strict as that defined by the discrete-time updates . Consequently , synapses undergo minor changes outside of the desired period of sensitivity . The learning speed of our model is better than that exhibited by an earlier proposed TD learning model on the same task [34] . The major difference between the two models is that in the previously proposed model , each synapse calculates its own approximation of the TD error based on a comparison of two post-synaptic activity traces with different time constants , whereas in the model presented here the TD error is represented as a population signal available to all synapses . This suggests that a population signal is a more reliable method for the brain to represent reward information . Although the grid-world task resembles a navigational task , it has more in common with an abstract association task such as learning associations between pairs of words , as the neuronal agent has no ability to exploit information about the underlying grid-world structure . This is also the reason why the agent requires many more trials to converge to a good performance than a rat requires to reliably find a hidden platform in a watermaze experiment [63] . Considerably faster convergence times have been demonstrated by reinforcement learning methods if the underlying structure of the environment is incorporated into the algorithm , for example by assuming overlapping state representations [29] , [39] . In our model , all states are initialized to the same value , reflecting the assumption that the agent knows nothing about the proximity of the states to the reward position at the outset . After the task has been learnt , a gradient is developed with higher values around the reward state . Clearly , it will take the agent longer to re-learn a new reward position far away from the previous one than it took to learn the original position , as the gradient has to be unlearnt . In contrast , rats re-learn a modified task much faster than they learnt the original task [63] . Faster re-learning has been demonstrated in a non-spiking actor-critic model when the agent learns an abstract internal state representation in addition to the value function and policy [39] . Interestingly , it has been shown that mice with suppressed adult neurogenesis also show highly specific learning deficits , especially in re-learning , which demonstrates the importance of newly generated neurons [64] . In future work we will extend our model to investigate the relationship between neurogenesis , internal state representation and the speed of re-learning a modified task . We have chosen the grid-world task to study the learning behavior of the proposed network model , as the complexity of the task makes it an adequate test case for TD learning algorithms . However , in experimental set-ups the role of dopamine in reward learning is typically studied in conditioning tasks , where a single stimulus is followed by a delayed reward . In order to test our network in such tasks requires an input representation different from the discrete state representation chosen in our model . Typically , in TD learning models such a stimulus is represented as a complete serial compound [2] , [4] . Here , the stimulus is represented by a vector , where the th entry represents the stimulus time steps into the future . Such a representation requires the system to know the number of time steps between the stimulus presentation and the reward delivery . A biologically more plausible representation of stimuli has recently been presented in [65] . Here the complete serial compound is replaced by a series of internal overlapping microstimuli . It has been demonstrated that such a representation results in a TD error in good agreement with most experimental findings on the dopaminergic activity during conditioning experiments [65] . It remains to be investigated in how far such a state representation can be adapted to spiking neuronal networks . Due to its low baseline level , the dopaminergic firing rate has a much smaller dynamic range available for the representation of negative errors than for positive errors . In the literature two main possibilities to represent negative TD errors have been discussed . One possibility is that negative errors are represented by a different neuromodulator such as serotonin [66] . Another possibility is that negative errors are encoded in the duration of the phasic pauses in the dopamine neurons [46] , suggesting that one neurotransmitter is enough to encode negative as well as positive errors . The latter hypothesis is supported in a modeling study demonstrating that dopamine is able to encode the full range of TD errors when the external stimuli are represented by a series of internal microstimuli [65] . Our study on the cliff-walk task with purely negative rewards reveals an additional problem to that of representing negative TD errors: due to their inherent lower bound the cortico-striatal synapses are limited in their ability to store estimates of future negative rewards . A possible hypothesis that would also allow learning to be driven by purely negative rewards is that the absolute values of the estimates of future negative rewards are stored in different synaptic structures from those storing estimates of future positive rewards . This hypothesis is in line with experimental studies in rats and humans showing a functional segregation within the striatum , with anterior regions responding more strongly to positive rewards and posterior regions to negative rewards [67]–[69] . An analogous segregation has also been reported between the amygdala and the ventral striatum , with the former responding only to losses and the latter to gains [70] . Our results support the hypothesis that prediction errors with respect to negative rewards are represented by a different neuromodulator and possibly a different anatomical system , rather than the duration of the phasic pauses in the dopamine neurons . On the other hand , they are compatible with a hybrid strategy in which the brain uses both mechanisms: a neuromodulator other than dopamine to encode negative errors due to punishment , and the phasic pauses in the dopaminergic firing rate to represent disappointment about an omitted reward . These hypotheses could be differentiated by tests on patients with Parkinson's disease or on animal Parkinson's models . In either case , we predict that learning is less impaired when driven by external negative rewards than by positive ones . The extent of the learning impairment in tasks where reward omission plays an important role will further discriminate whether the brain relies on dopamine or some other system to signal such events . We investigated to what extent a top-down derived plasticity model dependent on the dynamics of a dopaminergic signal with realistic firing rates is able to implement the TD ( ) algorithm . For this purpose we assumed a very simplified model of the basal ganglia adapted from [18] . The key feature for our model is that the critic module dynamically generates a realistic error signal in response to the development of the value function encoded in the cortico-striatal synapses and the chosen action , rather than artificially generating a perfect error signal outside of the network . The mechanism by which the dopaminergic error signal is generated by the basal ganglia is as yet unknown , and answering this question is outside the scope of this manuscript . The architecture of the critic module assumed in our model uses an indirect and a delayed direct pathway from the striatum to the dopamine neurons to produce an error signal with activity and temporal features similar to those experimentally . We implement the slowness of the direct pathway by a long synaptic delay; a more biologically realistic realization could be receptors , which are known to mediate slow inhibitory processes . Indeed , high densities of receptors have been found in the substantia nigra [71] . However , there are contradictory findings on whether the inhibitory response of the dopamine neurons is mediated by . Whereas in vitro inhibitory responses in midbrain dopamine neurons can be mediated by and [72] , [73] , in vivo studies in rats have reported that the synaptic connections between the neurons in the striatum and dopamine neurons in the substantia nigra act predominantly or exclusively via the receptors [74] , [75] . However , a recent in vivo study in mice found that after stimulation of the striatum , dopamine neurons in the substantia nigra show a long lasting inhibition mediated by receptors absent in rats [76] . Future experimental studies may reveal whether the dopaminergic signal is indeed generated by a fast indirect path and a slow direct pathway , or by some other mechanism [22] . Some alternative actor-critic models of the basal ganglia are discussed in [23] . Most of the alternative models make assumptions that are experimentally not well supported . For example , several models assume a direct excitatory pathway and an indirect inhibitory pathway between the striatum and the dopamine neurons [4] , [19]–[21] , [77] , whereas in reality the situation is reversed [23] . A model that basically resembles that proposed by Houk et al . [18] but implements several known anatomical structures more accurately than any other model was presented in [78] . However , this model relies on three-factor synaptic plasticity rules for striato-nigral connections , for which there is no experimental evidence . This assumption is also made in [79] . Some of the alternative models also posit a divergent architecture , in which the input arises from two different sources [79] , [80] . Due to the different timing properties along the two divergent pathways , the model proposed in [80] is able to reproduce most of the known experimental data . However , where parallel reciprocal architectures such as those proposed in [4] , [18]–[21] , [77] can be directly related to TD learning , the same is not true for divergent or non-reciprocal architecture [23] . The generation mechanism may also depend on pathways within the basal ganglia that have so far been neglected in modeling studies . For example , input from the lateral habenula to the dopamine neurons has recently been shown to be an important source of negative inputs to the dopamine neurons [81] . The focus of our work is action learning rather than action selection . Consequently , we have kept the actor module as simple as possible . One disadvantage of this choice is its vulnerability: if one actor neuron dies , the action that is represented by that neuron can never be chosen again . Furthermore , the inhibition of the actor neurons after an action has been chosen is applied externally rather than arising naturally through the network dynamics . Candidate action selection mechanisms that would overcome these limitations include attractor networks [82] and competing synfire chains [83]–[85] . Moreover , we have not related the action module to any specific brain region . Imaging experiments have found that the activity in the ventral striatum is correlated with the TD error during a prediction and action selection task , whereas the activity in the dorsal striatum is correlated with the TD error only during the action selection task [10] , [86] . In the context of the actor-critic architecture , this finding implies that the ventral striatum performs the role of the critic and the dorsal striatum performs the role of the actor . Detailed models have been developed that relate the problem of action selection to loops through the basal ganglia [41] , [42] and also loops through the cerebellum and the cerebral cortex [87] , [88] . An overview of different basal ganglia models that especially focuses on the action selection problem can be found in [89] . The error signal in our model is encoded in the difference between the dopaminergic population firing rate from its baseline level . The learning behavior of the model therefore depends on the number of dopamine neurons generating the population signal and the noise of this signal . As learning is driven by fluctuations in the dopaminergic firing rate from the baseline level , a noisier signal will drive the learning process less efficiently . A thorough investigation of the effects of model size and noise is outside the scope of this article , however , it is possible to extrapolate some of these effects from the dynamics of our model . We have shown that even as few as dopamine neurons generate a signal that is sufficiently reliable to learn the tasks investigated here . Increasing the number of neurons , assuming the synaptic baseline reference is correspondingly increased , would have the effect of reducing the noise in the dopamine signal . However , as the neuronal network model already performs as well as the discrete-time algorithm , no performance improvement can be expected . Conversely , decreasing the number of dopaminergic neurons reduces both the amplitude of the phasic signal and the baseline activity and makes the remaining signal noisier and less reliable . Even assuming a perfectly reliable signal , the dynamics developed in our model are such that if the synaptic baseline reference is not reduced accordingly , the lower baseline activity appears in the synaptic plasticity dynamics as a permanent negative error signal . This depresses the synaptic weights that encode the value function and policy until they reach their minimum values . At this point the agent can no longer distinguish between states with respect to their reward proximity and has no preference for any action over any other action . Moreover , decreasing the synaptic weights that encode the policy slows the responses of the actor neurons and therefore leads to slower decision processes . Analogous behavior has been observed in patients with Parkinson's disease , which is characterized by a gradual loss in the number of dopamine neurons , who show movement as well as cognitive deficits [90] . The dynamics of our model predicts that increasing background dopamine concentration after a gradual loss in dopamine neurons maintains any existing memory of state values , as it will restore the amount of available dopamine to the baseline level used as a reference by the synapse . However , learning in new tasks is still impaired , as this is driven by fluctuations in the dopaminergic signal rather than its baseline level . The reduced remaining population of dopaminergic neurons necessarily produces smaller and noisier fluctuations than those generated by an intact population; consequently , they provide a less effective learning signal . This is an equivalent situation to reducing the size of the dopamine population and reducing the baseline reference value in the synapse accordingly . This prediction is consistent with the finding that even fully medicated Parkinson's patients exhibit deficits in a probabilistic classification task [91] . The dynamics of the critic module also predicts that the size of the striatal population should also be critical for the learning behavior , as it determines the amplitude of the phasic dopaminergic signal . This is in agreement with studies showing that a lesion of the dorsal striatum impairs the learning behavior of rats in stimulus-response learning [92] . The plasticity dynamics presented in Eq . ( 8 ) is in some degree similar to the plasticity dynamics derived in our previous investigation of a spiking neuronal network model capable of implementing actor-critic TD learning [34] . The two plasticity dynamics have in common that the dynamics is triggered by biologically plausible measures of the pre-synaptic activity and is dependent on a TD error signal . However , in our earlier model there is no dopaminergic error signal available; each synapse performs its own approximation of an TD error based on the difference in a rapid and a laggard post-synaptic activity trace . The aim was to develop a continuous-time plasticity mechanism that mapped the properties of the discrete-time TD learning algorithm as accurately as possible . Thus , the study can be seen as a proof of principle that a spiking neuronal network model can implement actor-critic TD learning . On the basis of this , in our current study we focus on applying biological constraints to the range of possible plasticity dynamics by combining the previous top-down approach with a bottom-up approach . The biological constraints entailed by our use of a dopaminergic error signal with realistic firing rates to represent the TD error lead to two major differences from the original plasticity mechanism developed in [34] . First , whereas the plasticity dynamics presented in the previous model belongs to the class of differential Hebbian learning rules modulated with a non-local constant reward signal , in the model presented here , the plasticity dynamics belongs to the class of neuromodulated , heterosynaptic plasticity . Second , whereas the earlier synaptic plasticity dynamics can be mapped exactly to the value function update of TD learning , the plasticity dynamics presented here corresponds to a slightly modified TD learning algorithm with self-adapting learning parameters . Our finding that the learning parameters and increase with the difference in successive cortico-striatal synaptic weights could be tested experimentally by fitting TD learning algorithms to behavioral data gathered from animals learning two versions of a task: one with large rewards and one with small rewards . As long as , the task with larger rewards will develop greater differences in the estimation of future rewards of successive states than the task with smaller rewards . We therefore predict that the values of the learning parameters and fitted to the former set of behavioral data will be greater those fitted to the latter set . Additionally , the values calculated by fitting and to different epochs in behavioral data gathered from an animal learning a given task should vary in a systematic fashion . At the very beginning , the animal presumably has no expectations about future rewards and thus estimates all states similarly . During the middle of the learning process , when the animal's performance is improving rapidly , large differences between the estimation of states can be expected . Finally , as the animal approaches its equilibrium performance , differences between the estimations of states should vary smoothly . We therefore predict that fitting and to data gathered from the beginning and end of the learning process will result in lower values than fitting the learning parameters to data gathered whilst the performance on a given learning task is improving rapidly . Is actor-critic TD learning the correct model ? This is outside the scope of the current manuscript , and perhaps out of our remit altogether - this kind of question can only be answered by analyzing behavioral , electrophysiological and anatomical data from carefully designed experiments . There is evidence on the behavioral american [93] as well as on the cellular level american [2] , [9] that mammals implement TD learning strategies . TD learning has been successfully applied to model bee foraging in uncertain environments american [19] , [94] , human decision making american [4] and rat navigation american [39] , but it is unlikely to be the only learning strategy used by the brain [95] . In line with previous studies [10] , [18] , [20] , we have focused on TD learning with the actor-critic architecture instead of other TD learning methods , such as SARSA or Q-learning [1] . However , recent experimental findings also support the interpretation that mammals implement TD learning methods based on action values [17] or an actor-director model [14] . Further research is needed , especially on the theoretical side , in order to understand if these models are compatible with spiking neuronal networks . We have focused on the simplest TD learning algorithm: TD . However , it is likely that the mammalian brain uses more advanced TD learning strategies . TD learning is efficient as long as the number of possible states and actions are restricted to a small to moderate number . To address problems with a large number of states and possible actions , TD learning methods that generalize from a small number of observed states and chosen actions are needed ( see [1] ) . Furthermore , it has been demonstrated that classical TD learning schemes cannot account for behavioral data involving motivation . Modified TD algorithms can explain these data , either by explicitly including a motivational term [96] or by ‘average-reward TD-learning’ , where an average reward acts as a baseline [97] . Here , we have interpreted the phasic dopaminergic signal in the light of TD learning . However , the literature presents a much broader picture of the functional role of the dopaminergic activity . It has been found that only a small subgroup of dopamine neurons show a response consistent with the TD error hypothesis; a much broader group responds with an increase in activity to positive as well as negative reward related signals inconsistent with the hypothesis [98] . There is also evidence that dopamine is involved with signalling ‘desire’ for a reward rather than the reward itself [99] , [100] . Furthermore , the phasic dopaminergic signal responds to a much larger category of events than just to reward related events , including aversive , high intensity or novel stimuli [101] . Alternative interpretations of the phasic signal include the theory that it acts more like a switch than a reward signal , triggering learning at the right point in time [102] , [103] , or that it promotes the discovery of new actions and learning of new action-outcome associations , independent of the economic value of the action [5] . Given the diversity of dopaminergic responses and considering the fact that midbrain dopamine neurons project to many different brain areas , such as the striatum , the orbifrontal cortex and the amygdala [3] , it is also likely that different interpretations are simultaneously valid; the information encoded in the phasic signal being combined with local information in specific areas of the brain to realize a variety of functions .
We investigated our model using numerical simulations . We implemented the model in the simulator NEST [104] and performed the simulations in parallel on two nodes of a cluster of SUN X machines , each with two AMD Opteron quad core processors running Ubuntu Linux . The dopamine modulated plasticity dynamics Eq . ( 8 ) and Eq . ( 13 ) are implemented employing the distributed simulation framework presented in [105] . All neurons in the network are modeled as current-based integrate-and-fire neurons . The dynamics of the membrane potential for each neuron is given by:where is the time constant , the capacity of the membrane and the input current to the neurons [106] . When reaches a threshold , a spike is emitted . The membrane potential is subsequently clamped to for the duration of an absolute refractory period . The synaptic current due to an incoming spike is represented as an -functionwhere is the peak amplitude and the rise time . The neuronal parameters are specified in the following section . The details of the model are summarized in Fig . 12 using the scheme developed by [107] . The parameters used in the numerical simulations are specified in Fig . 13 .
|
What are the physiological changes that take place in the brain when we solve a problem or learn a new skill ? It is commonly assumed that behavior adaptations are realized on the microscopic level by changes in synaptic efficacies . However , this is hard to verify experimentally due to the difficulties of identifying the relevant synapses and monitoring them over long periods during a behavioral task . To address this question computationally , we develop a spiking neuronal network model of actor-critic temporal-difference learning , a variant of reinforcement learning for which neural correlates have already been partially established . The network learns a complex task by means of an internally generated reward signal constrained by recent findings on the dopaminergic system . Our model combines top-down and bottom-up modelling approaches to bridge the gap between synaptic plasticity and system-level learning . It paves the way for further investigations of the dopaminergic system in reward learning in the healthy brain and in pathological conditions such as Parkinson's disease , and can be used as a module in functional models based on brain-scale circuitry .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"neuroscience/behavioral",
"neuroscience",
"neuroscience",
"neuroscience/theoretical",
"neuroscience"
] |
2011
|
An Imperfect Dopaminergic Error Signal Can Drive Temporal-Difference Learning
|
Annual mass drug administration ( MDA ) over five years is the WHO's recommended strategy to eliminate lymphatic filariasis ( LF ) . Some experts , however , consider that longer periods of treatment might be necessary in certain high prevalence and transmission environments based upon past unsuccessful field experience and modelling . To evaluate predictors of success in a LF control program we conducted an ecological study during a pre-existing MDA program . We studied 27 villages in Lihir Island , Papua New Guinea , from two areas with different infection rates before MDA . We undertook surveys to collect information on variables potentially having an influence on the outcome of the program , including epidemiological ( baseline prevalence of infection , immigration rate ) , entomological ( vector density ) and operational ( treatment coverage , vector control strategies ) variables . The success in a village was defined using variables related to the infection ( circulating filarial antigenemia prevalence <1% ) and transmission ( antigenemia prevalence <1 in 1000 children born since start of MDA ) . 8709 people were involved in the MDA program and average coverage rates were around 70% . The overall prevalence of filariasis fell from an initial 17 . 91% to 3 . 76% at round 5 ( p<0 . 001 ) . Viewed on a village by village basis , 12/27 ( 44% ) villages achieved success . In multivariate analysis , low baseline prevalence was the only factor predicting both success in reducing infection rates ( OR 19 , 26; CI 95% 1 , 12 to 331 , 82 ) and success in preventing new infections ( OR 27 , 44; CI 95% 1 , 05 to 719 , 6 ) . Low vector density and the use of an optimal vector control strategy were also associated with success in reducing infection rates , but this did not reach statistical significance . Our results provide the data that supports the recommendation that high endemic areas may require longer duration MDA programs , or alternative control strategies .
Lymphatic filariasis ( LF ) , caused by the mosquito-borne nematode Wuchereria Bancrofti , is a major public-health problem in many tropical and subtropical regions . Papua New Guinea represents the biggest remaining challenge for elimination of the disease . The Global Program to Eliminate Lymphatic Filariasis ( GPELF ) was launched in 1997 . In the Pacific , the World Health Organization ( WHO ) has implemented from 1999 , the Pacific Program to Eliminate Lymphatic Filariasis ( PacELF ) bringing together 22 countries and territories , in a common effort to eliminate the disease [1] , [2] . The PacELF strategy is based on five rounds of mass drug administration ( MDA ) , monitored by a prevalence survey to assess the impact at completion of the last round [3] , [4] . Therefore , the assessment is designed to conclude whether to stop or to continue MDA after round 5 . The rationale of this approach is to suppress microfilaremia ( mf ) in infected populations and bring the infection level down below a threshold that will prevent resurgence of infection and ultimately lead to interruption of transmission [5] . The exact infection level to achieve LF elimination in different endemic regions remains unknown , such that it is difficult to predict or decide when to stop ongoing MDA programs . Previous reports have suggested that residual filarial infections disappear when prevalence rates fall to less than 1% but it may vary depending on specific ecological conditions [6] , [7] . Moreover , some programs which have achieved this threshold have reported evidence of ongoing transmission , as measured by antibody or antigen prevalence in children aged 2–4 years and mosquito infection rates [8] , [9] . The current recommendation of the PacELF is that programs should reach an antigenemia level below 1% and that less than 1 in 1000 children born since start of MDA should become newly infected [1] , [2] . End-points for the GPELF have recently been changed to a level below 2% in areas where the main vector is an anopheline [4] . The optimal duration of MDA programs has also not been established . Mathematical models suggest that 4 to 6 years of treatment should be sufficient [10] . However , several programs have reported evidence of failure to control the infection , as indicated by mf and circulating filarial antigenemia prevalence rates after completing five annual rounds of MDA [11]–[14] . Numerous attempts have been made to establish which variables may influence the outcome of a program . Some variables , such as the coverage of the target population [9] , [15] , [16] , the drug regimen employed [5] , [17]–[20] , and the integration of vector control measures [21]–[23] , are controllable . Other biological and epidemiological variables , such as the initial prevalence of mf in an area and the vectorial abundance of the mosquito , are less amenable to modification . All the above mentioned factors need to be taken into consideration when developing an elimination strategy [24] . The aim of this study was to estimate success rates of the program to eliminate lymphatic filariasis ( PELF ) in villages from different areas and to identify determinants of success affecting a PELF's outcome .
Data entry was undertaken with EpiInfo software ( version 6 ) , with field limits and double data entry . We analysed demographic data ( age and sex ) , migration from other areas , the initial endemicity of infection ( prevalence rate ) , the vector abundance , treatment coverage ( number of tablets distributed ) , the use of adjuvant vector control strategies and the outcome of the elimination program . Univariate analyses of data from the villages were performed using the χ2 test for categorical data , and the t-test or Mann–Whitney U-test for continuous data . Predictors for success of PELF in controlling the infection prevalence were analysed by multivariate logistic regression , which included the following variables: low baseline prevalence , low migration , low vector density , high treatment coverage , and optimal vector control strategy . Predictors of success in stopping the transmission were also analysed by logistic regression analysis , which included the variables: average age <20 years , low endemicity of infection , low vector density and high treatment coverage . Odds ratios and 95% confidence intervals are presented . All multivariate logistic analyses were done using Firth's method to overcome problems of separation in small samples [28] . Data were analysed using SAS 9 . 1 . 3 ( SAS Institute Inc . , Cary , NC , USA ) and SPSS 14 . 0 ( SPSS Inc . , Chicago , IL , USA ) . Reporting of the study has been verified in accordance with the STROBE checklist ( provided as Checklist S1 ) . Ethics approval for this study , including the oral consent process , was obtained from the Papua New Guinea Ministry of Health Medical Research Advisory Committee . The consent sought was verbal because of the high illiteracy rate in rural population , and it was documented on case report forms . Study personnel informed prospective study participants about the study by reading them a consent document in the local language . All subjects provided informed consent at every stage of the study including for collection of samples , interviews , and individuals involved in calculating monthly bite rates . Participation by children required consent from at least one parent and the child's assent .
The annual census in the year 2003 estimated that 8709 people lived in the 27 villages which were part of the study and that were visited and treated annually over a period of 5 years . At the baseline survey , a total of 6037 individuals were registered corresponding to 70 . 0% of the entire population ( as determined by the 2003 census ) . 50% of the villagers were male , and the mean age was 20 . 6 years . Reported coverage in 2004–2007 for MDAs 2–5 was 69 . 8% , 73 . 0% , 74 . 1% and 71 . 5% respectively . Drug coverage remained stable over time and it was similar in all the territories , with an average of 72 , 9% in eastern coast villages and 67 , 4% in western coast villages ( p = 0 . 35 ) . The reason why 26% to 30% of the target population were not treated is that they were not available at the time of the medical team visit ( i . e . working , visiting relatives ) or that they were ineligible persons ( 3 . 0% to 6 . 0% of the de facto population ) , including pregnant women and children younger than 2 years old or of weight less than 10 kg . The demographic and epidemiological data and program operational details of the 27 villages are shown in Figure 1 . Almost half ( 44 , 4% ) of the villages had migration rate over 5% of the total population , 12 ( 80 . 0% ) of them were villages from the eastern coast . The migrants mainly came from low endemic areas for filariasis such as the PNG mainland , and from the small islands surrounding Lihir . The details of treatment coverage were recorded during all the five rounds of treatment and the average calculated . Individually , 55 . 6% of the villages had a high drug coverage , including 60 , 0% ( 9/15 ) in the eastern coast group and 50 , 0% ( 6/12 ) in the western coast group . Mosquito transmission indices varied significantly in different villages . The highest indices of transmission were observed in villages located in the swampier regions of the west coast with a MBR median ( interquartile range ) of 460 . 8 ( 1812 . 6 ) bites/person/month , compared to 57 . 6 ( 180 . 0 ) in the east coast . These data indicate that there is regional micro-variation in the intensity and temporal pattern of filariasis transmission . It is noteworthy that only nine villages ( 33 . 0% ) used indoor residual spraying and therefore achieved an optimal vector control strategy . Table 1 shows data for filarial infection rates before and after the PELF . Overall , all variables showed significant decreases from pre-MDA to round 5 ( p<0 . 001 ) . The analysis of data from all villages shows a significant decrease in circulating filarial antigenemia ( CFA ) over this period from a mean prevalence of 17 . 9% to 3 . 8% ( p<0 . 001 ) . Pre-MDA CFA prevalence rates were much higher in the Western than in the Eastern territories . The mean prevalence of infection was 7 . 7% and 0 . 8% in eastern villages , and 30 . 7% and 7 , 5% in western villages in the pre-MDA and at completion of round 5 , respectively ( table 1 ) . Pre-MDA antigenemia prevalence rates in children under 5 years were not significantly different in the western village schools and in the eastern village schools ( 60 . 1 vs 31 . 0 in 1000; p = 0 . 10 ) . Rates of circulating filarial antigenemia in under 5 y . o . children fell more rapidly in the less heavily infected eastern villages than in the western villages ( 30 . 0 vs 0 . 95 in 1000; p<0 . 01 ) . Twenty-one children with a positive result out of 700 tested were identified in the Western villages after round 5 . As shown in figure 2 , the CFA prevalence in individual villages ranged from 1 . 1% to 58% and 0 to 17% , in the pre-MDA and post-MDA surveys respectively . PELF had a successful outcome on infection prevalence control in 12 of the 27 villages ( 44 . 4% ) , whereas it failed in the remaining 15 ( 55 . 5% ) . Transmission , assessed on the incidence of new infections , was successfully controlled in 19 ( 70 , 3% ) villages . On univariate analysis ( table 2 ) , numerous factors were found to be significantly associated with PELF success on the control of infection status including low baseline prevalence , low vector abundance , and implementation of an optimal vectorial control . A low percentage of migration , unexpectedly , was found to be a risk factor . On multivariate analysis ( table 2 ) , the only independent factor predicting PELF infection control success was low endemicity of infection ( OR 19 . 26; CI 95% 1 . 12–331 . 82 ) . Table 3 shows the univariate and multivariate predictors associated with transmission control success . Low endemicity of infection ( OR 27 . 44; CI 95% 1 . 05–719 . 6 ) was again the only factor independently associated with transmission control success .
In the present study the overall prevalence of circulating filarial antigenemia was reduced by 79 . 0% . Despite undeniable success , the program did not achieve its very ambitious goals based on the PacELF recommendations , with post treatment prevalence remaining at 3 . 8% and 14 new infections per 1000 children born since start of MDA . This overall failure was the result of the intervention specifically failing in some villages . Understanding what factors lead to success or failure when the intervention is applied to a specific setting may help improve the MDA program . Our findings reveal that in Lihir , the baseline infection status was an important factor influencing on the outcome of the PELF . Our discussion will focus on possible explanations for this observation and the influence of other factors on the outcome . The first objective of our study was to test the hypothesis that bancroftian filariasis can be eliminated from communities by yearly cycles of MDA with diethylcarbamazine and albendazole . Over all , success in controlling the infection and in stopping transmission was confirmed in 45% and 70% of the villages , respectively , most of them located in the eastern coast . The eastern coast villages had low baseline levels of filariasis endemicity , whereas the western coast villages had very high baseline rates that were more typical of filariasis endemic islands in the Western Pacific Region [2] . Our data suggest that five rounds of MDA will not have eliminated filariasis in the western study area . However the observed post-treatment antigen prevalence rates are unlikely to sustain transmission as the vector-parasite relationship in Lihir is extremely fragile involving the An . Farauti which is one of the least efficient vectors in the world . The cut-off point for interruption of transmission in the PacELF region was based on the fact that transmission in most of the Pacific island countries is carried out by the highly efficient Culex and Aedes mosquitoes . The second major objective of our study was to evaluate the factors having a positive influence on the markers of success . In our study the most prominent determinant of success was low baseline prevalence of infection . Low vector density appeared to have an association but did not reach statistical significance . The sample size , as it is based on the number of villages , is unfortunately low and may have caused the study to be underpowered when trying to determine the significance of this trend . Other specific factors have previously been described as having a positive influence in the outcome of a PELF . These include high coverage of the targeted population [9] , [15] , [16] , low levels of migration from other areas and integration of the different available control strategies into the program [21] , [22] . In the current study however , these variables did not show an association . Prior to the current study there has been little reliable clinical evidence comparing low prevalence communities with high prevalence populations . Our study followed a large cohort , contained detailed information about risk factors and outcomes and established comparisons among several independent areas of transmission . The current study also had the advantage of being conducted in an area of Papua New Guinea with a relatively high rate of infection and transmission , and a pronounced inter-area variation in prevalence . Moreover , an island population such as Lihir presents more ideal conditions for epidemiological studies and evaluation of control programs than large land areas . This study has the limitations of an observational ecological study , and obviously causality cannot be inferred from it . The observed risk factors need to be considered with caution . Also , data for some factors which may influence success rates ( e . g . number of persons per household , use of bednets ) were not collected . However , the analyses performed controlling for the widely recognized prime predictors , the strong and similar results obtained in the multivariate analyses for two different markers of success , and the use of specific techniques to obtain unbiased risk estimates with small samples allow us to have confidence in the results obtained . The percentage of the population covered is an important factor in determining the success of a PELF that has been previously analysed [9] , [15] , [16] . We developed a timely and coordinated drug delivery strategy that included elements of community information and education in an attempt to achieve widespread acceptance of drug treatment and the DOT ( Directly Observed Therapy ) distribution method . We made an effort to reach those groups of individuals who are recognized to be at risk of systematic non-compliance during MDAs including children , the upper socioeconomic classes , young men and the elderly . We achieved an overall 70% reported population treatment coverage which was probably underestimated , since it was calculated using the number of people in the local census . In the multivariate analysis the coverage was not associated with any of the markers of success . This is likely due to the similar levels of coverage achieved in all the villages in the current program . The lack of influence of population migration in this study may be explained by the demographics of the migrant population . The presence of the gold mining operation at Lihir has attracted a large number of relatively skilled and affluent workers , often from areas such as Port Moresby and other regional centres where LF endemicity is low . Thus migration did not contribute to the reservoir of filariasis in Lihir . The long term impact of MDA is determined by the drugs' effects on microfilaria , and particularly on adult worms [29] . A single dose of DEC and Albendazole rapidly reduces the number of circulating microfilaria , but also has a temporary effect in reducing the production of microfilaria by adult worms , probably due to sterilization . After some months renewed production occurs but at a reduced intensity [20] , [29] . In the absence of macrofilaricidal activity , current programs rely on the interruption of transmission through sustained suppression of microfilaremia over the 5 years of a program . Experimental studies have documented the median fecund lifespan of W . Bancrofti worms to be more than the 5 years typical of an MDA [30] . In addition , single doses of DEC and Albendazole have been shown to have a limited capacity to kill the adult worm . A Brazilian study found a significant proportion of adult worms were insensitive to DEC at doses of 6 mg/kg , with ultrasound studies showing only a 56% mortality for adult worms after 5 years [29] . This has led some physicians to suggest that MDA programs should be of at least the same duration as the lifespan of adult worms . Moreover , it has been demonstrated that individuals in areas of high endemicity will have a higher average adult worm burden , and therefore a higher chance of a fecund worm pair surviving after MDA is complete [14] . A theoretical analysis with field data from 9 villages , in distinct endemic areas , identified the degree of infection aggregation as one of the main factors related to failure of a PELF . Through a simulation procedure , the author estimated that following the current approach , only 50% of programs would achieve parasite elimination [6] . Dunyo et al . in their study showed a higher level of microfilarial resurgence than in other programs , and suggested that this may have been due to a high pre-treatment worm burden [31] . The fact that successful elimination of disease in high prevalence areas may require longer duration of MDA programmes has already been recognized by experts [1]–[4] , [24] , [32]–[34] . This paper provides the data that support such recommendations . It is clear that local data needs to be taken into account when designing MDA programs . Alternative strategies may be needed , including modified drug regimens ( e . g . , biannual MDA ) , vector control measures , or perhaps antibiotic treatment .
|
Large-scale intervention programmes to control filariasis are currently underway worldwide . However , a major unresolved question remains: what is the appropriate duration for these programmes ? Recent theoretical work and clinical field experience has highlighted how the ecological diversity between different endemic regions hinders decision making processes of when to stop ongoing MDA programs . The goal of our study was to identify the factors determining success for a five year LF elimination program . We undertook different types of surveys together with a pre-existing MDA program in villages from two regions that had different infection prevalence rates . Our study shows that the five yearly cycles of MDA could neither eliminate the disease nor stop transmission in the high prevalence villages , such that low baseline lymphatic filariasis prevalence has a positive influence on the outcome of a program . Thus , the study provides data supporting the recommendation that in certain high prevalence and transmission environments more sustained efforts may be necessary .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"infectious",
"diseases",
"neglected",
"tropical",
"diseases",
"lymphatic",
"filariasis"
] |
2011
|
The Impact of a Filariasis Control Program on Lihir Island, Papua New Guinea
|
Photosynthesis is the final determinator for crop yield . To gain insight into genes controlling photosynthetic capacity , we selected from our large T-DNA mutant population a rice stunted growth mutant with decreased carbon assimilate and yield production named photoassimilate defective1 ( phd1 ) . Molecular and biochemical analyses revealed that PHD1 encodes a novel chloroplast-localized UDP-glucose epimerase ( UGE ) , which is conserved in the plant kingdom . The chloroplast localization of PHD1 was confirmed by immunoblots , immunocytochemistry , and UGE activity in isolated chloroplasts , which was approximately 50% lower in the phd1-1 mutant than in the wild type . In addition , the amounts of UDP-glucose and UDP-galactose substrates in chloroplasts were significantly higher and lower , respectively , indicating that PHD1 was responsible for a major part of UGE activity in plastids . The relative amount of monogalactosyldiacylglycerol ( MGDG ) , a major chloroplast membrane galactolipid , was decreased in the mutant , while the digalactosyldiacylglycerol ( DGDG ) amount was not significantly altered , suggesting that PHD1 participates mainly in UDP-galactose supply for MGDG biosynthesis in chloroplasts . The phd1 mutant showed decreased chlorophyll content , photosynthetic activity , and altered chloroplast ultrastructure , suggesting that a correct amount of galactoglycerolipids and the ratio of glycolipids versus phospholipids are necessary for proper chloroplast function . Downregulated expression of starch biosynthesis genes and upregulated expression of sucrose cleavage genes might be a result of reduced photosynthetic activity and account for the decreased starch and sucrose levels seen in phd1 leaves . PHD1 overexpression increased photosynthetic efficiency , biomass , and grain production , suggesting that PHD1 plays an important role in supplying sufficient galactolipids to thylakoid membranes for proper chloroplast biogenesis and photosynthetic activity . These findings will be useful for improving crop yields and for bioenergy crop engineering .
Plants possess a sophisticated sugar biosynthetic machinery comprised of families of nucleotide sugars that can be modified at their glycosyl moieties by nucleotide sugar interconversion enzymes to generate different sugars [1] , [2] . UDP-glucose 4-epimerase ( also UDP-galactose 4-epimerase , UGE; EC 5 . 1 . 3 . 2 ) catalyzes the interconversion of UDP-D-glucose ( UDP-Glc ) and UDP-D-galactose ( UDP-Gal ) [3] , [4] . UGE is essential for de novo biosynthesis of UDP-Gal , a precursor for the biosynthesis of different carbohydrates , glycolipids , and glycosides . Genes encoding UGE have been cloned from a range of different organisms including bacteria , yeast , and human [5]–[7] , and the crystal structures have also been obtained [8]–[10] . The original biochemical and genetic analyses of UGE in plants was described by Dörman and Benning [11] . To date , five UGE isoforms have been identified in Arabidopsis [2] , [12] , three in barley [13] , and a family of four putative UGE isoforms exist in rice . In Arabidopsis , global co-expression analysis revealed that UGE2 , -4 , and -5 preferentially act in the UDP-Glc to UDP-Gal directions , whereas UGE1 and UGE3 might act in the UDP-Gal to UDP-Glc directions [14] . Reverse genetic studies demonstrated that UGE2 and UGE4 influence vegetative growth and cell wall carbohydrate biosynthesis , that UGE3 is specific for pollen development , and that UGE1 and UGE5 act in stress situations [15] , [16] . Compared to 4-day-old seedlings , UGE expression increased 5-fold in roots of 3-week-old pea plants , suggesting that increased UGE expression correlated with the copious secretion of pectinaceous mucigel in older seedling roots [17] . To date , all UGEs identified from plants lack transmembrane motifs and signal peptides and appear to exist as soluble entities in the cytoplasm . Photosynthetic reactions in higher plants depend on the well-developed chloroplast thylakoid membrane system . Chloroplast thylakoid assembly and maintenance require a continuous supply of membrane constituents . Galactose-containing glycerolipids are predominant lipid components of photosynthetic membranes in plants , algae , and cyanobacteria . The two most common galactolipids are mono- and digalactosyldiacylglycerol ( MGDG and DGDG ) , which account for about 50 and 25 mol% of total thylakoid lipids , respectively [18] , [19] . About 80% of all plant lipids are associated with photosynthetic membranes , and MGDG is considered to be the most abundant membrane lipid on earth . Recent studies have demonstrated that galactolipids play an important role in not only the organization of photosynthetic membranes , but also in their photosynthetic activities [20] , [21] . Arabidopsis mutants with a lower amount of these galactolipids have a reduction in chlorophyll content and photosynthetic activity , alterations in chloroplast ultrastructure , and impairment of growth [22]–[25] . In plants , MGDG is synthesized in two unique steps: ( i ) the conversion of UDP-Glc into UDP-Gal by an UGE , and ( ii ) the transfer of a galactosyl residue from UDP-Gal to diacylglycerol ( DAG ) for synthesis of the final product by MGDG synthase ( MGD1 ) [26] , [27] . Although MGD1 has been characterized at both genetic and enzymatic levels , the UDP-Gal supply mechanisms for the MGDG biosynthetic pathway remain obscure . MGD1 is localized in the inner chloroplast envelope membrane [26] , [27] and uses UDP-Gal as a substrate . However , the concentration of UDP-Gal in chloroplasts is considered to be very low [28] , suggesting that the UDP-Gal source is imported from the cytosol or generated inside chloroplasts . To gain insight into genes controlling photosynthetic activity and carbon assimilation in plants , a rice stunted growth mutant ( phd1 ) with decreased photoassimilate and yield production was selected for further study from a large-scale screening of our T-DNA mutant population . Interestingly , PHD1 encoded a chloroplast-localized UDP-Glc epimerase involved in UDP-Gal supply for chloroplast galactolipid biosynthesis during photosynthetic membrane biogenesis . Its homologs are highly conserved in the plant kingdom , and the gene was preferentially expressed in various young meristems where plastid proliferation actively occurred . Most strikingly , overexpression of PHD1 increased photosynthetic activity and enhanced rice growth . The important roles of PHD1 in photosynthetic capability and carbon assimilate homeostasis are discussed .
To identify genes affecting photosynthetic activity and carbon assimilation , a large-scale screening of our rice T-DNA insertion mutant population ( Oryza sativa var . Nipponbare background ) [29] was carried out . Of 480 mutant lines with altered carbohydrate levels in vegetative organs , photoassimilate defective1 ( phd1 ) with both low carbohydrate contents and stunted growth was selected for further characterization . Scanning electron micrograph of culms demonstrated that fewer starch granules were deposited in parenchyma cells of the phd1 mutants ( data not shown ) . During the young seedling stage , both shoots and primary roots of the mutant were shorter and lighter than those of the wild type ( Figure 1A ) . After internode elongation , the phd1 mutant exhibited a semi-dwarf , less grain-filling , retarded vegetative growth , later flowering , and less tillering phenotype ( Figure 1B–1E ) . In addition , although the grain number per panicle was not altered between the mutant and wild type , the seed-setting ratio of the phd1 mutant was significantly decreased , which finally led to a significant reduction of grain yield ( Figure 1F , 1G ) . Compared to wild type , mature leaves of the mutant had somewhat reduced sucrose ( Figure 1H ) and rather low starch levels ( Figure 1I ) at all time-points taken during the light/dark cycle , while hexose levels were a little higher in the mutant ( Figure S1 ) . Genetic analysis indicated that the phd1 phenotype was controlled by a single recessive gene that did not co-segregate with the T-DNA insertion , and hence map-based cloning was carried out . The PHD1 locus was physically delimited to a 72-kb region on the short arm of chromosome 1 . This region contains six annotated genes , and sequencing of these genes from phd1-1 identified a single nucleotide transition ( G-to-T ) in exon 2 of Os01g0367100 , leading to a premature translational termination . The identity of Os01g0367100 as PHD1 was confirmed by analysis of two other phd1 alleles with similar phenotypes isolated from the same genetic screen , for which a single nucleotide substitution ( A-to-T ) in exon 7 in phd1-2 and a 13-bp insertion between exon 3 and exon 4 in phd1-3 were found ( Figure 2A ) . Almost no PHD1 mRNA was detected in any of the three allelic mutants ( Figure S2 ) . The phd1 phenotype was complemented by transgenic expression of wild type Os01g0367100 in the phd1-1 mutant background ( Figure 2B , 2C ) , confirming that the nonsense mutation of Os01g0367100 was responsible for the presumed null mutant phenotype . Database searches revealed that PHD1 has similarity to proteins from Thalassiosira pseudonana ( XP_002290295 ) , Phaeodactylum tricornutum ( XP_002178225 ) , Chlamydomonas reinhardtii ( XP_001699105 ) , Micromonas pusilla ( EEH60780 ) , Ostreococcus tauri ( CAL54696 ) , Physcomitrella patens ( XP_001767242 ) , Ricinus communis ( XP_002516868 ) , Arabidopsis thaliana ( AT2G39080 ) , Populus trichocarpa ( XP_002311843 ) , Vitis vinifera ( XP_002276706 ) , Zea mays ( NP_001131736 ) , and Sorghum bicolor ( XP_002457832 ) , with 27 to 75% amino acid identity ( Figure S3 ) . Phylogenetic analysis between PHD1 and its 16 putative homologs indicated that PHD1 is closely related to Sb03g014730 from sorghum and LOC100193101 from maize ( Figure 3 ) . PHD1 homologs are only found in the plant kingdom , suggesting that these proteins are evolutionally conserved across plant species . However , none of the homologous genes have been functionally characterized . Analysis of the conserved domain demonstrated that PHD1 and its homologs contain the consensus WcaG domain , featured in nucleoside-diphosphate sugar epimerases ( Figure S3 ) . One of the best characterized nucleotide sugar epimerases is UDP-Glc epimerase , which catalyzes the interconversion of UDP-Glc and UDP-Gal . Hence , PHD1 and its homologs may function as novel plant specific UDP-Glc epimerases . To validate PHD1's biochemical function as an UDP-Glc epimerase , the mature PHD1 protein lacking the putative N-terminal 62-aa transit peptide was expressed in E . coli and UGE activity was examined . The result showed that PHD1 could catalyze the conversion of UDP-Gal to UDP-Glc , and curve fitting indicated that UDP-Gal binding followed a simple Michaelis-Menten kinetics with a Km value of 0 . 84 mM at 30°C ( Figure S4A ) . To examine whether PHD1 had UDP-Glc epimerase activity in vivo , the mature PHD1 under the control of the yeast glyceraldehyde-3-phosphate dehydrogenase promoter was used to complement the auxotrophic phenotype of a yeast gal10Δ mutant which cannot grow on a medium containing D-galactose as sole carbon source . The complementation results demonstrated that PHD1 also had UDP-Glc epimerase activity in vivo ( Figure S4B ) . RNA gel blot analysis revealed that PHD1 was present in all green tissues , with highest abundance in leaf blades and leaf sheaths , then flowers and culms , but only at very low levels in roots ( Figure 4A ) . mRNA in situ hybridization using an antisense probe revealed that PHD1 was expressed predominantly in leaf primordia and shoot apical meristems ( Figure 4B ) , the mesophyll cells surrounding the vascular bundles of young leaves ( Figure 4C ) , inflorescence primordia ( Figure 4D ) , and axillary buds ( Figure 4E ) . In contrast , hybridization with a PHD1 sense probe showed no signal ( Figure 4F ) . PHD1 encodes a 340 aa protein with a putative 62-aa chloroplast transit peptide at the N-terminus . To confirm chloroplast localization of PHD1 , the full-length PHD1 was fused to the green fluorescent protein ( GFP ) reporter gene under the control of the cauliflower mosaic virus ( CaMV ) 35S promoter and subsequently transformed into rice shoot protoplasts . Figure 5A shows that GFP fluorescence co-localized with the red chlorophyll autofluorescence , confirming that PHD1 was a chloroplast-localized protein and the predicted transit peptide was functional . To further investigate the subcellular localization of PHD1 , we performed western blot experiments using purified plastid subfractions ( Figure 5B ) . Several antibodies were used as specific markers for the different chloroplast subfractions . Tic 40 was used as a specific envelope marker , and Rubisco , the major stroma protein , as a marker of this chloroplast subfraction . PsbA , one of the components of photosystem II ( PSII ) , was used as a marker to validate the thylakoid membrane fraction , and HSP82 was used as a cytosol specific marker . As shown in Figure 5B , the PHD1 protein was detected mainly in the stroma fraction and was absent from the cytoplasmic compartment , thus confirming that PHD1 was a chloroplast-targeted protein . To complete the subcellular localization study and to obtain additional information about the distribution of PHD1 in different chloroplast subcompartments , we further performed immunocytochemical analysis on ultrathin sections of rice tissues using polyclonal PHD1 antiserum . The positive signal of PHD1 , visualized as black dots , was found specifically in the chloroplasts ( Figure 5C and 5D ) . In contrast , sections treated with a preimmune serum ( Figure 5E and 5F ) showed no signal . The overall data thus strongly indicated that PHD1 is targeted to chloroplasts in rice . Intact chloroplasts were isolated from leaves of wild type and phd1-1 mutant plants , and the UGE activity in isolated chloroplasts was measured ( Figure S5 ) . Compared to the wild type , a severe decrease ( ca . 50% ) in UGE activity was observed in isolated chloroplasts from the phd1-1mutant compared with the wild type , suggesting that PHD1 was responsible for a major part of the UGE activity in chloroplasts . Moreover , levels of the UGE substrates UDP-Glc and UDP-Gal in isolated chloroplasts were also determined ( Figure 6 ) . While compared to wild type and complemented mutant an overabundance of UDP-Glc was found in chloroplasts isolated from the phd1-1 mutant , almost no amount of UDP-Gal was detected in the mutant . The levels of nucleotide sugars in whole leaves were also determined , which showed that the amount of UDP-Gal was slightly higher in phd1-1 than in wild type plants , and the UDP-Glc amount was significantly higher ( Figure S6 ) . Hence , the ratio of UDP-Glc to UDP-Gal in phd1-1 leaves was also higher than in wild type plants . These results suggested that PHD1 dysfunction may trigger an accumulation of substrates and disturb the balance of interconversion between the two sugar nucleotides . Chloroplast membranes contain high levels of glycolipids , and UDP-Gal is a dominant substrate for glycolipid biosynthesis . To examine the effect of PHD1 dysfunction on membrane lipid homeostasis , the composition of total lipids extracted from phd1-1 , wild type , and PHD1-complemented plants was analyzed ( Figure 7 ) . In the phd1-1 mutant , the mol% amount of MGDG was reduced by 19% compared to wild type and the complemented plants , indicating that PHD1 is involved in MGDG biosynthesis . In contrast , only a slight decrease ( 2 . 5% ) in DGDG content was observed in the phd1-1 mutant , demonstrating that PHD1 may not be required for DGDG synthesis and suggesting that the UDP-Gal substrate for DGDG formation was presumably supplied from the cytosol . Reduced abundance of MGDG in phd1-1 was accompanied by an increase in the abundance of other major membrane lipids such as phosphatidylcholine ( PC ) , phosphatidylglycerol ( PG ) , and phosphatidylinositol ( PI ) , while the mol% levels of sulfoquinovosyldiacylglycerol ( SQDG ) and phosphatidylethanolamine ( PE ) were not altered significantly in the phd1-1 mutant ( Figure 7A ) . Because the two galactolipids and SQDG are major components of thylakoid membrane lipids , this result suggests that the mutant had an overall lower amount of chloroplast membrane lipids than wild type plants . Focusing on the exclusive chloroplast lipid MGDG , the fatty acid composition was also investigated ( Figure 7B ) . MGDG of the phd1-1 mutant contained considerably decreased levels of stearic acid ( 18∶0 ) compared with the wild type and elevated levels of linoleic acid ( 18∶2 ) and linolenic acid ( 18∶3 ) . The levels of other fatty acids were similar to those observed in wild type plants . Hexadecatrienoic acid ( 16∶3 ) , which is typically found in the plant prokaryotic pathway , was not detected in all the rice plants , suggesting that rice entirely relies on endoplasmic reticulum ( ER ) -derived lipids for thylakoid galactoglycerolipid biosynthesis . Noninvasive chlorophyll fluorescence measurements indicated that the maximum quantum yields for photosystem II photochemistry ( Fv/Fm ) were similar for phd1-1 and wild type ( Table 1 ) . The effective quantum yield of photochemical energy conversion in photosystem II ( ΦPSII ) was slightly but significantly reduced in the mutant ( Table 1 ) . Pigment content was also reduced in the phd1-1 mutant ( Table 1 ) . Interestingly , chloroplasts of 2-month-old phd1-1 plants were significantly smaller than those of wild type plants ( wild type , 5 . 0±0 . 4 µm; phd1-1 , 3 . 0±0 . 5 µm ) , and starch grains were also either absent or reduced in size and/or number in the mutant ( Figure S7 ) . These data indicated that a reduced amount of galactolipids in chloroplasts and perhaps a smaller size of chloroplasts due to a decrease in membrane lipid content might lead to reduced photosynthetic capability of higher plants . UDP-Gal is the activated form of galactose in biosynthetic reactions , but a galactose salvage pathway exists in eukaryotic organisms . To assess expression of genes involved in the Leloir salvage pathway , the expression levels of three key genes of this pathway , GalM , GalK , and GalT , were analyzed in both phd1-1 and wild type . The expression of all three genes was significantly upregulated in the phd1-1 mutant , suggesting an activation of the whole salvage pathway ( Figure 8A ) . β-Lactase is involved in the generation of free β-D-Gal from polysaccharide breakdown , and UDP-Glc pyrophosphorylase ( UGP ) catalyzes the formation of UDP-Glc from Glc-1-P . The expression levels of genes encoding β-lactase and UGP3 were also upregulated in phd1-1 . More strikingly , the expression levels of OsUGE1 and OsUGE4 encoding for putative cytoplasmic isoforms of UGE in rice were more than two-fold higher in phd1-1 than in wild type plants , indicating an upregulation of de novo UDP-Gal biosynthesis in the cytoplasm . These results suggested that PHD1 may be responsible for a majority of the UGE function in chloroplasts , and appears to be involved in the generation of UDP-Gal from UDP-Glc to supply building blocks for galactolipid biosynthesis required for proper chloroplast membrane composition . Because the phd1-1 mutant exhibited a dramatic decrease of carbon assimilate levels , we determined whether transcript levels of several key genes involved in the synthesis , transport , and cleavage of starch and sucrose were altered in mature leaves of wild type and phd1-1 plants . Interestingly , while the expressions of starch biosynthesis genes such as AGPL2 , SSI , SSIIIa , GBSS , BE , and BT1 , were suppressed in the phd1-1 mutant ( Figure 8B ) , expression levels of genes participating in sucrose cleavage , such as INV1/3 and SuSy1 , were all increased ( Figure 8C ) . Meanwhile , the GPT gene encoding a glucose-6-phosphate/phosphate translocator was upregulated in phd1-1 , implying an enhanced export of hexose-phosphates from chloroplasts to the cytosol . In addition , increased expression level of UGP2 , a gene involved in UDP-Glc synthesis , was correlated with increased UDP-Glc accumulation and a higher UDP-Glc/UDP-Gal ratio in the phd1-1 mutant . Since a mutation in PHD1 affected photosynthesis and growth rate , we further investigated whether biomass and grain yield could be improved by PHD1 overexpression . When grown in paddy fields , transgenic rice plants overexpressing PHD1 showed a significant increase in tillering ( branching ) and photosynthetic rate ( Figure 9A , Table S1 ) in lines that overexpressed the PHD1 protein ( Figure 9B ) . Compared to non-transgenic control plants , grain yield per plant of transgenic lines S3 , S5 , and S8 increased 10 . 7 , 15 . 5 , and 18 . 3% , respectively ( Figure 9C ) . In addition , the growth rate of transgenic plants accelerated at the seedling stage and dry material accumulation was enhanced 12 . 5% to 22 . 4% at the mature stage compared to non-transgenic plants ( Figure 9D , Table S1 ) . These results demonstrated that PHD1 overexpression in rice is positively correlated with an increase in biomass production and grain yield .
To date all UGE genes coding for UDP-Glc epimerases isolated from plants are localized to the cytosol , where their substrates UDP-Glc and UDP-Gal are present at high levels [30] . As a precursor for the synthesis of the galactolipid MGDG in chloroplasts , UDP-Gal is widely assumed to be mobilized from the cytosol , because the UDP-Gal concentration is relatively low within plastids [28] and MGDG synthase ( MGD1 ) is associated with the inner envelope membrane [26] , [27] . However , a labeling experiment in which radioactively labeled UDP-Gal was applied to isolated Arabidopsis chloroplasts revealed that radioactivity was not efficiently incorporated into MGDG [23] , raising the question of how UDP-Gal is transported into the chloroplasts . In this study , we found that a mutation in PHD1 , which encodes a novel rice plastidial UGE involved in the biosynthesis of chloroplast galactolipids , lead to disturbed carbon assimilation homeostasis and impaired photosynthetic efficiency . Our work revealed that PHD1 codes for an active epimerase that is targeted to chloroplasts , and , therefore , that the UDP-Gal substrate for MGDG biosynthesis can be generated in situ in chloroplasts ( Figure 10 ) . The novel finding that this UGE is chloroplast-targeted was supported by three independent lines of evidence ( Figure 5 ) . First , PHD1-GFP fusion products were found exclusively in chloroplasts . Second , Western blot analyses of fractionated chloroplasts showed that PHD1 was highly enriched in the stroma fraction of chloroplasts . And third , immunocytochemistry indicated that PHD1 was concentrated inside the chloroplast stroma , most likely associated with the thylakoid surface . This striking result provides a well-defined genetic and biochemical framework to study the novel functional mechanism of this UGE in plastids , and to evaluate the role of galactolipids in photosynthetic activity of rice . Of MGDG synthases that are primarily important for thylakoid membrane biogenesis , MGD1 is considered to be the major isoform [24] . In Arabidopsis , two more MGDG synthases , MGD2 and MGD3 , are targeted to the outer chloroplast envelope where substrates can be recruited from the cytosol [27] . MGDG generated by them can move from the outer to the inner envelope and to the thylakoids . Here we show that compared to wild type , the relative amount of the major galactolipid MGDG was reduced by 19% in the phd1-1 mutant , whereas that of DGDG was only slightly decreased by 2 . 5% . We observed a slight increase in the mol% amount of the thylakoid lipid phosphatidylglycerol , which may compensate for a fraction of the galactolipids lost in the phd1-1 mutant . Meanwhile , the relative amount of several extraplastid phospholipids was found to be slightly but significantly higher in the phd1-1 mutant , suggesting that compared to extraplastidic membranes , the overall amount of plastid membranes might have decreased . These results are consistent with the hypothesis that the amounts of glycolipids and phospholipids are reciprocally controlled in plants to maintain a proper balance of lipids in the ER and plastid membrane systems [20] , [31] . It has been shown previously that osmotic stress induced variations in membrane fluidity that correlated with the physical properties of membrane lipids [32] . Due to an overabundance of UDP-Glc observed in chloroplasts and entire leaves of the phd1-1 mutant , hyperosmotic stress might occur , and an increased production of 18∶3 could affect hyperosmotic stress tolerance in the mutant chloroplasts . This would be in agreement with earlier observations that transgenic enhancement of fatty acid unsaturation rendered cells and whole plants more tolerant to sorbitol-induced osmotic stress in tobacco [33] . Most galactolipids are restricted to plastid membranes during normal growth and development , however , DGDG can also be found in extraplastidic membranes following phosphate ( Pi ) starvation [34] , [35] . Importantly , x-ray crystallographic analyses of photosynthetic proteins in cyanobacteria revealed that MGDG is associated with the core of the reaction centers of both photosystems I and II ( PSI and PSII ) [36] , [37] , which suggest that these lipids are required not only as bulk constituents of photosynthetic membranes , but also for the photosynthetic reaction itself . Consistent with this , we found that the effective quantum yield of photochemical energy conversion in photosystem II ( ΦPSII ) was reduced in the phd1-1 mutant . Seedlings lacking MGDG were previously shown to have disrupted photosynthetic membranes , leading to a complete impairment of photosynthetic ability and photoautotrophic growth [22] , [24] . In agreement with this , a possible reduction of thylakoid membrane amount and a changed galactolipid to phospholipid ratio in chloroplast membranes in the phd1-1 mutant might have led to the dramatic phenotype of retarded growth , reduced photosynthetic capability , and decreased photoassimilate accumulation . Taken together , this strongly suggests that the stunted growth phenotype of phd1-1 mutants is due to an insufficient provision or slower production of membrane building blocks to support chloroplast proliferation during plant growth , which is also consistent with the reduced numbers of thylakoid stacks and sizes of chloroplasts observed in mutant plants . In plants , starch acts as a depository for reduced carbon produced in leaves during the day , and as a supply of chemical energy and anabolic source molecules during the night [38] . Pyrophosphate ( PPi ) is produced during the upregulation of UGP3 ( Figure 10 ) , and hydrolyzed by very high pyrophosphatase ( PPase ) activity in plastids [39] . Moreover , inorganic phosphate ( Pi ) released during PPi hydrolyzation is an inhibitor for key regulatory starch biosynthesis enzymes such as AGP [40] . In the phd1-1 mutant , expression levels of starch biosynthesis genes such as AGP , SS , GBSS , and BE , were significantly downregulated in source leaves , leading to a sharp decrease of starch content . However , the reduced starch did not result in increased sucrose levels , because activation of sucrose cleavage genes SuSy1 and INV1/3 resulted in reduced sucrose and increased hexose-phosphate and UDP-Glc levels . Therefore , sucrose as the main transport form of photoassimilate produced in source organs was not able to export efficiently to the sink organs . Moreover , a large amount of UDP-Glc catalyzed by SuSy1 or UGP2 would be converted to UDP-Gal by cytosolic OsUGE1/4 and transported into chloroplast as galactosyl donors of chloroplast glycolipids to compensate for the loss of PHD1 activity in the phd1-1 mutant . In contrast , PHD1 overexpression in rice , which enhanced PHD1 activity in chloroplasts ( Figure S5 ) , might increase the relative amount of MGDG and increase the effective quantum yield of photochemical energy conversion in thylakoid membranes , resulting in increased photosynthetic efficiency and growth rate , implicating a key role of PHD1 for the photosynthetic system in rice . These improvements of both biomass production and grain yield have significant economic implications in both traditional crop improvement and bioenergy crop production .
The rice ( Oryza sativa L . ) phd1 mutant is in the Nipponbare ( ssp japonica ) background . F2 mapping populations were generated from a cross between the rice phd1 mutant and MH63 ( ssp indica ) . Rice plants were cultivated in the experimental station of the Institute of Genetics and Developmental Biology ( IGDB ) in Beijing in natural growing seasons . For analysis of diurnal changes of starch and sugars , rice plants were kept in a growth chamber at 28°C and 70% relative humidity under a photoperiod of 12 h light/12 h darkness , with a light intensity of 200 µmol quanta m−2 s−1 . Genomic DNA was isolated from seedlings of the selected plants with the mutant phenotype . For fine mapping of PHD1 , STS markers were generated based on the polymorphisms between Nipponbare and MH63 . The molecular lesion of phd1-1 was identified by PCR amplification of the PHD1 genomic region from wild type and phd1-1 mutant plants and comparison of their sequences . The candidate gene was mapped between the 2 new STS markers S221 ( 5′-AGAGCTAGGGGGTAAAAA-3′ and 5′-GTGCAGAACAGTGGAATG-3′ ) and S246 ( 5′-AACCCTATCCTTCCTCACCA-3′ and 5′-TTGTCCCTCCGCCTGCTTCC-3′ ) . PHD1 homologs were detected by BLASTp using the entire amino acid sequence of PHD1 as a query in the National Center for Biotechnology Information database ( http://www . ncbi . nlm . nih . gov/BLAST ) . Multiple alignment of the homologs was performed by Clustal X version 2 . 0 with the default parameters [41] and manually adjusted . For constructing phylogenetic trees , the neighbor-joining method of the MEGA 4 . 1 software [42] was used , and a bootstrap analysis with 1 000 replicates was performed to test the confidence of topology . The BAC clone BAC53 containing the entire PHD1 fragment was digested with Sac I and Pst I to generate a 7 . 96 kb genomic DNA fragment . The DNA fragment was ligated to the Sac I and Pst I digested pCAMBIA1300 vector ( CAMBIA ) , to generate the pSCL construct for complementation analysis . The full-length PHD1 cDNA was PCR amplified using primers 5′-GATCCGATCCCCTCACCTC-3′ and 5′- TTCTCTGGCCGAAACCATT-3′ , and subcloned into the pCAMBIA2300-35S binary vector , between the cauliflower mosaic virus 35S promoter and nopaline synthase ( nos ) terminator , to generate the pSOL construct for overexpression analysis . Transgenic rice plants were generated according to Agrobacterium tumefaciens-mediated transformation methods [43] , [44] . The transgenic plants were then transferred to the field at the IGDB experimental station for normal growth and seed harvesting . PHD1 cDNA was amplified by primer sets 5′-TGATGATACAGGGGTCAAGATG-3′ and 5′-ACTGTCAAGACCAAGGAATTCT-3′ and cloned into the Xma I and Xho I sites of pGEX-4T-1 ( GE Healthcare Life Sciences ) and expressed in E . coli strain BL21 ( DE3 ) . Recombinant PHD1 protein was affinity-purified through glutathione Sepharose resin ( Amersham Pharmacia Biotech ) and used for antibody production [45] . Total RNA was prepared with an RNeasy kit ( Qiagen ) . In the RNA gel blot analysis , 5 µg of total RNA was electrophoresed on a 1 . 2% ( w/v ) agarose gel and transferred to a nylon membrane , and mRNA was detected by a digoxigenin labeling system ( Roche Diagnostics ) . For quantitative RT-PCR , 15 ng of cDNA and SYBR Green SuperMix ( Bio-Rad ) were used in 15 µL qRT-PCR reactions with a CFX96 96-well real-time PCR detection system ( Bio-Rad ) and CFX96 software to calculate threshold cycle values , and rice 18S ribosome RNA was used as an internal control . Oligonucleotide primers are given in Table S2 . The 2−ΔΔCT method was adopted to calculate the relative expression levels for the phd1 and wild type samples , and a two-tailed t test used to compare the ratios and determine statistical significance [46] . Freshly collected rice tissues were fixed in FAA solution ( 50% ethanol , 5% acetic acid , 3 . 8% formaldehyde ) at 4°C overnight , dehydrated with ethanol solution from 50% to 100% , cleaned by a series of xylene washes from 25% to 100% , and embedded in paraffin ( Paraplast Plus , Sigma-Aldrich ) at 54–56°C as described in [47] . 8 to 12 µm sections were cut with a microtome ( Leica RM2265 ) , and mounted on RNase-free glass slides and photographed . RNA in situ hybridization was performed as described previously with minor modification [48] . Briefly , the 420-bp region of PHD1 was amplified by gene-specific primers with T7 or SP6 promoters 5′-TAATACGACTCACTATAGGGCCCCTTCTCCGTCAACCT-3′ and 5′-AACGAAAGAGCCTTCACCA-3′ or 5′-CCCCTTCTCCGTCAACCT-3′ and 5′-ATTTAGGTGACACTATAGAACGAAAGAGCCTTCACCA-3′ in front of the reverse primer ( for making anti-sense probe ) or forward primer ( for making sense probe ) . Digoxigenin-labeled RNA probes were prepared using a DIG Northern Starter Kit ( Cat . No . 2039672 , Roche ) according to the manufacturer's instructions . The hybridization signals were observed using bright field imaging with a microscope ( Olympus BX51 ) and photographed with a Micro Color CCD camera ( DVC Co . Austin , USA ) . A binary vector containing GFP fused with full-length PHD1 was constructed as follows . The PCR product amplified with primers 5′-ACCTCCGTCCCTGCTTCCTC-3′ and 5′-GGGCTCCCAACCAATCTCA-3′ was subcloned into the CaMV 35S::GFP vector to generate CaMV 35S::PHD1-GFP . The binary vector was transformed into rice protoplasts using the polyethylene glycol method [49] . After overnight incubation in the dark , the protoplasts expressing GFP were imaged by a confocal laser scanning microscope ( LSM510 , Zeiss , Germany ) using 488 nm excitation and 500–530 nm emission pass-filters . Chlorophyll autofluorescence was detected with 570 nm excitation and 640 nm emission pass-filter [50] . Composite figures were prepared using Zeiss LSM Image Browser software . PHD1 and its derivative cDNAs were amplified by PCR using the primers 5′- ATGATACAGGGGTCAAGATGG-3′ and 5′-ACTGTCAAGACCAAGGAATTCT -3′ , and inserted into the vector pDBLeu ( Invitrogen ) . The Euroscarf S . cerevisiae strain BY4742 ( Matα his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0 gal10::kanMX4 ) was transformed using a lithium acetate procedure and tested on 1% ( w/v ) galactose medium ( 1% ( w/v ) yeast extract ( Duchefa ) , 2% ( w/v ) Bacto-peptone ( BD Bio- sciences ) , 1% ( w/v ) galactose ( Sigma ) , 2% ( w/v ) Micro agar ( Duchefa ) ) . Individual samples ( leaves of circa 500 mg fresh weight ) were harvested and frozen rapidly in liquid N2 . The frozen samples were homogenized and extracted with perchloric acid . Glucose , fructose , sucrose , and starch were measured enzymatically for the neutralized supernatant ( sugars ) and the insoluble pellet ( starch ) [51] . Determination of UDP-Glc and UDP-Gal were performed as described [11] . Total lipids were extracted from 2-month-old phd1-1 , wild type , and the PHD1-complemented plants as described [52] . For quantitative analysis , individual lipids were separated by two-dimensional thin-layer chromatography and used to prepare fatty acid methyl esters . The methyl esters were quantified by gas-liquid chromatography as described [53] . A 1 µl sample was applied for GC-MS ( Agilent 7890A GC coupled to 5975C MS ) analysis at a 10∶1 split ratio . The GC-MS program started with 80°C for 1 min , then ramped at 8°C/min to 300°C and held for 5 min; injector and inlet temperatures were set at 250°C and 280°C , respectively . Separation was performed on a HP-5 MS column ( 30 m×0 . 25 mm×0 . 25 µm ) with a constant flow of 1 . 1 ml/min helium . The MS scan range was from 50 to 500 m/z . The quantification of fatty acid methyl esters was performed by the external standard method . UGE activity was measured using a NADH-coupled assay developed by Wilson and Hogness [54] with some minor modifications . The 1 ml assay mixture consisted of 100 mM glycine buffer ( pH 8 . 7 ) , 1 mM β-NAD+ ( Sigma ) , and 0 . 8 mM UDP-Gal ( Sigma ) . The reaction was started by adding 10 µl of epimerase ( 140 µg/ml ) in 50 mM Tris·Cl ( pH 7 . 6 ) , 1% ( w/v ) bovine serum albumin , 1 mM dithiothreitol , 1 mM EDTA , and 1 mM β-NAD+ , and stopped by incubation for 10 min at 100°C . The UDP-glucose produced was determined by addition of 0 . 04 unit of bovine UDP-glucose dehydrogenase ( Calbiochem ) and incubation for 10 min at 30°C , and the increase in absorbance due to NADH formation was then measured at 340 nm . Km values were determined by varying the UDP-Gal concentration between 0 . 4 mM and 3 . 2 mM . The experiment was conducted in triplicate . All isolation procedures were carried out at 4°C . Batches of 50 g rice leaves were cut to little pieces and homogenized in 250 ml of isolation buffer ( 50 mM HEPES/KOH , pH 7 . 8 , 0 . 33 M sorbitol , 2 mM EDTA , 1 mM MgCl2 , 1 mM MnCl2 , 0 . 1 M Na-ascorbate , 0 . 2% ( w/v ) bovine serum albumin ) using a Waring blender . The chloroplast suspension was passed through four layers of Miracloth and centrifuged at 4 000 g for 4 min . The pellet was gently suspended in the isolation buffer and layered onto a discontinuous density gradient consisting of 10 , 40 , and 80% ( v/v ) Percoll in the isolation buffer . The gradient was centrifuged at 8 000 g for 10 min . Intact chloroplasts distributed around the 40/80% Percoll interface were isolated and reapplied to the Percoll gradient centrifugation . Chloroplasts were lysed by resuspension to 0 . 5 mg chlorophyll ml−1 in 10 mM HEPES/KOH ( pH 8 . 0 ) , 5 mM MgCl2 , for 20 min on ice , and the lysate was fractionated into envelope , stroma , and thylakoids by differential centrifugation as described by Skalitzky et al [55] . All solutions contained a cocktail of protease inhibitors . To verify recovery and purity of the sucrose density fractions , several antibodies against specific marker proteins were used: Tic40 was used as an envelope marker , RbcL as a stromal marker , and PsbA as a thylakoid membrane marker . Immunoelectron microscopy experiments were carried out as previously described [56] . Briefly , nickel grids carrying ultrathin leaf sections prepared from two-week-old wild type seedlings were sequentially floated in 0 . 01 M sodium phosphate buffer ( PBS , pH 7 . 2 ) containing 5% ( w/v ) bovine serum albumin ( BSA ) for 5 min , then for 1 h at 37°C in PBS containing diluted anti-PHD1 antibody . After several washes in PBS , ultrathin sections were incubated for 1 h at 37°C in PBS containing goat anti-rabbit IgG antibody conjugated to 10-nm colloidal gold ( 1∶40 , Sigma-Aldrich , St . Louis , MO , USA ) . After 5 washes with PBS , ultrathin sections were washed with distilled water , air dried , counterstained with 2% uranyl acetate , and examined with a FEI Tecnai G2 20 transmission electron microscopy at an accelerating voltage of 120 kV . Negative controls were performed using the same procedure with the exception of substituting the anti-PHD1 antibody with preimmune serum .
|
Photosynthesis is carried out in chloroplast , a plant-specific organelle . Photosynthetic membranes in chloroplasts contain high levels of glycolipids , and UDP-galactose is a dominating donor for glycolipid biosynthesis . Although glycolipid assembly of photosynthetic membranes has been characterized at the genetic and enzymatic level , the mechanism of substrate supply of UDP-galactose for the glycolipid biosynthetic pathway remains obscure . By genetic screening of rice mutants that are impaired in photosynthetic capacity and carbon assimilation , we identified PHD1 as a novel nucleotide sugar epimerase involved in a process of glycolipid biosynthesis and participating in photosynthetic membrane biogenesis . PHD1 was preferentially expressed in green and meristem tissues , and the PHD1 protein was targeted to chloroplasts . We revealed that UDP-galactose for glycolipid biosynthesis catalyzed by the new enzyme was generated inside chloroplasts , and the reduced amounts of glycolipids in the mutant led to decreased chlorophyll content and photosynthetic activity . Overexpression of this gene lead to growth acceleration , enhanced photosynthetic efficiency , and finally improved biomass and grain yield in rice . These results suggest that PHD1 has significant economic implications in both traditional crop improvement and bioenergy crop production .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"agriculture",
"biology"
] |
2011
|
A Rice Plastidial Nucleotide Sugar Epimerase Is Involved in Galactolipid Biosynthesis and Improves Photosynthetic Efficiency
|
We investigated the relationship of treatment regimens for visceral leishmaniasis ( VL ) with post-kala-azar dermal leishmaniasis ( PKDL ) and visceral leishmaniasis relapse ( VLR ) development . Study subjects included cohorts of patients cured of VL since treatment with monotherapy by sodium stibogluconate ( SSG ) , miltefosine ( MF ) , paromomycin intramuscular injection ( PMIM ) , liposomal amphotericin B [AmBisome ( AMB ) ] in a single dose ( SDAMB ) and in multidose ( MDAMB ) , and combination therapies by AMB+PMIM , AMB+MF , and PMIM+MF . Follow up period was 4 years after treatment . Cohorts were prospective except SSG ( retrospective ) and MF ( partially retrospective ) . We compared incidence proportion and rate in 100-person-4year of PKDL and VLR by treatment regimens using univariate and multivariate analysis . 974 of 984 enrolled participants completed the study . Overall incidence proportion for PKDL and VLR was 12 . 3% ( 95% CI , 10 . 4%–14 . 5% ) and 7 . 0% ( 95% CI , 5 . 6%–8 . 8% ) respectively . The incidence rate ( 95% CI ) of PKDL and VLR was 14 . 0 ( 8 . 6–22 . 7 ) and 7 . 6 ( 4 . 1–14 . 7 ) accordingly . SSG cohort had the lowest incidence rate of PKDL at 3 . 0 ( 1 . 3–7 . 3 ) and VLR at 1 . 8 ( 0 . 6–5 . 6 ) , followed by MDAMB at 8 . 2 ( 4 . 3–15 . 7 ) for PKDL and at 2 . 7 ( 0 . 9–8 . 4 ) for VLR . Development of PKDL and VLR is related with treatment regimens for VL . SSG and MDAMB resulted in less incidence of PKDL and VLR compared to other treatment regimens . MDAMB should be considered for VL as a first step for prevention of PKDL and VLR since SSG is highly toxic and not recommended for VL .
Visceral leishmaniasis ( VL ) or kala-azar has been a public health problem in Bangladesh over the centuries [1] . About 150 , 000 people have suffered from VL since 1994 in Bangaldesh [2] . Victims of VL are the poorest of the poor living in the rural areas of the country . The epidemiology of VL in Bangladesh , India , and Nepal is similar . In these three countries , the protozoan parasite Leishmania donovani ( LD ) is the only causative agent of VL , humans are the only reservoir , and the female Phlebotomas argentipes sandfly is the only vector . In 2005 , these three countries signed a memorandum of understanding to eliminate VL as a public health problem by 2015 , targeting to keep VL incidence less than 1 per 10 000 people at upazila ( sub-district ) , block and district levels in Bangladesh , India and Nepal respectively [3] . However , the target was not achieved in 2015 . Current WHO Road Map aims to eliminate Neglected Tropical Diseases including VL by 2020 [4] . Bangladesh and Nepal achieved the target in 2016 and 2013 respectively . VL peaks periodically in the Indian subcontinent; this is known as the natural trend of VL [5] . It is believed that during inter-epidemic periods , cases of post-kala-azar dermal leishmaniasis ( PKDL ) fuel the transmission of LD in endemic communities . PKDL is a sequel of infection by LD parasite; it develops mostly among patients who have been treated for VL [6–7] . PKDL also may develop among individuals with asymptomatic infection with the parasite [7] . Hypopigmented macular , papular , and nodular or combined skin lesion are the key clinical features of PKDL [6] . Skin lesions usually do not itch and have intact skin sensitivity . Patients with PKDL are clinically healthy and usually do not seek medical care unless they are stigmatized by their lesions . Patients with PKDL harbor the parasite in their skin for years and are infective to sandflies [8–9] . In this way , they continue transmission of LD in the community and threaten VL control in long run . Failure to control PKDL , therefore , is a substantial challenge to the success of the South-East Asia Region Kala-azar Elimination Programme ( KAEP ) [10] . Visceral leishmaniasis relapse ( VLR ) is defined as the reappearance of VL after complete cure at 6 months after treatment . VLR patients have higher concentration of the parasite in their peripheral blood compared to that of new VL patients , are infective to sandflies , sustain transmission of the infection , and threaten VL control [9 , 11] . Therefore , VLR also present challenges for controlling VL in the long run . To date , there are no preventive strategies against PKDL and VLR development . Sodium stibogluconate ( SSG ) for 6 months or miltefosine ( MF ) for 84 days are current treatment options for PKDL [10] . Both drugs are toxic , and serious adverse events associated with them are common [6 , 12] . Further , efficacy with either treatment does not exceed 90% and relapse after treatment with MF also has been reported [10] . Therefore , a strategy to prevent PKDL is urgently needed . Incomplete treatment for VL has been shown as a risk factor for PKDL in Nepal [13] . A study in India found that VL patients living in areas with high arsenic exposure are more prone to develop PKDL [14] . Immune gene polymorphism has been associated with the development of PKDL in Sudan [15] . An association between SSG treatment for VL and PKDL development has been speculated [16] . However , subsequent reports showed that PKDL also developed after treatment of VL with MF , paromomycin intramuscular injection ( PMIM ) , and liposomal amphotericin B [AmBisome ( AMB ) ] [17–19] . To date , there has been no study of the relationship between treatment for VL and development of PKDL and VLR . SSG was the only drug for treatment of VL for more than 70 years . Later , MF , PMIM , and AMB either in monotherapy or in combination of any two drugs were introduced for treatment of VL [20–23] . In this study , we aim to investigate the relationship of PKDL and VLR development to VL treatment among cured VL patients treated with different treatment regimens .
The study site was Mymensingh , the most VL-endemic district in Bangladesh . This was a cohort study which included cured VL patients treated with different treatment regimens for new VL ( please see below ) . Duration of the study was October , 2014 to December , 2018 . Study participants were enrolled after a study initiation meeting , investigator training , and training of field staff in diagnosis of suspected PKDL and VLR . Using clinical trial logs and hospital records ( only for MF cohort ) , study participants were actively searched and identified at their current places of residence . They were invited to participate in the study to complete four years of follow-up , as well as provide information regarding VLR and PKDL ( if any ) since treatment for VL . Trained field research assistants followed up study subjects through home visits every 3 months until completion of the 4-year follow-up period . The study physician referred suspected PKDL/VLR cases to the study clinic , Surya Kanta Kala-azar Research Centre , for medical examination , confirmation , and management . A suspected PKDL patient was a person with history of kala-azar and skin lesions; a probable PKDL case was a suspected PKDL case with positive rK39 test; and a confirmed PKDL case was a probable PKDL case with LD parasite documented either by slit skin examination , culture , or polymerase chain reaction . Probable and confirmed PKDL cases were eligible for treatment following the national guideline for kala-azar [24] . A suspected case of VLR was a case of cured VL with fever for more than 2 weeks and splenomegaly . A confirmed case of VLR was a suspected case of VLR with LD bodies in spleen aspirates documented by microscopy/LD DNA in spleen aspirates or peripheral blood buffy coat . A standard data management plan has been developed by the experts at Clinical data management centre , Christian Medical College ( CMC ) , Vellore , India . Oracle Clinical ( OC ) version 4 . 6 . 6 was used to design web-based dual data entry system following the annotated Case Report Form ( CRF ) of this study . External monitors checked and approved all CRFs before data entry . Data management team at CMC further scrutinized dual entered data and generated data clarification forms which were clarified by the data management and field team in icddr , b . Following data quality control checks , the data management team at CMC provided a final clean data set for analysis . We used Chi-square test for comparing proportions of various socio-demographic variables of the study participant between different cohorts . Comparison of mean/median between cohorts was done by parametric or non-parametric test where applicable . We calculated both the incidence proportion and rate ( 100-person-4years ) of PKDL and VLR . Finally we performed a Cox-proportional hazard model using SSG cohort with least incidence of PKDL and VLR as reference to investigate the independent risk factor ( s ) for development of PKDL and VLR . Data were analysis using STATA 13 . The icddr , b Ethical Review Committee , Western Institutional Review Board ( WIRB ) , and PATH Research Ethics Committee approved the study . Informed voluntary written consent from adults and assent from children between 11 and 17 years old were obtained for their participation . In cases of study participants who were less than 11 years old , consent from a parent/legal guardian was also obtained .
974/984 enrolled subjects completed the study , indicating 1% lost to follow-up ( Table 1 ) . Study cohorts differed significantly in terms of age groups , sex , education level , monthly expenditure , family size , house type , ownership of cattle and bed-nets , and bed-net use ( Table 2 ) . These variables were taken as covariates for multivariate analysis later . 121/984 developed PKDL ( mean , 95% CI , 12 . 3% , 10 . 4%–14 . 5% ) with a median time 2 . 6 years ( IQR , 1 . 84–3 . 12 ) . The incidence proportion ( mean , 95% CI ) with PKDL was lowest in the SSG ( 3 . 0% , 1 . 2%–7 . 0% ) , followed by the MDAMB ( 8 . 0% , 4 . 2%–14 . 7% ) , MF ( 9 . 3% , 5 . 6%–15 . 2% ) , AMB+PMIM ( 10 . 7% , 6 . 1%–18 . 1% ) , SDAMB ( 15 . 9% , 10 . 4%–23 . 5% ) , AMB+MF ( 16 . 2% , 10 . 2%–24 . 7% ) , PMIM ( 19 . 1% , 12 . 5%–27 . 9% ) , and PMIM+MF ( 22 . 9% , 15 . 7%–32 . 0% ) ( Table 3 ) . The SSG differed significantly from all other cohorts , except the MDAMB cohort . All 984 participants contributed 1 370 628 days of observation , with a mean of 1393 days ( 95% CI , 13 80–1406 ) when cohorts were looked for PKDL development . The average incidence rate ( 100-person-4years ) of PKDL was 14 . 0 ( 8 . 6–22 . 7 ) . SSG cohort had lowest incidence rate ( 3 . 0 , 1 . 3–7 . 3 ) , followed by the MDAMB ( 8 . 2 , 4 . 3–15 . 7 ) , MF ( 9 . 7 , 5 . 7–16 . 4 ) , AMB+PMIM ( 11 . 3 , 6 . 4–19 . 9 ) , SDAMB ( 16 . 9 , 10 . 9–26 . 2 ) , AMB+MF ( 17 . 1 , 10 . 6–27 . 5 ) , PMIM ( 20 . 1 , 13 . 0–31 . 2 ) , and PMIM+MF ( 25 . 3 , 16 . 9–37 . 7 ) ( Table 3 ) . Incidence rate of PKDL of SSG and MDAMB cohorts did not differ significantly . Assuming SSG cohort as a reference we analyzed Cox Proportional Hazard Ratio for other cohorts for development of PKDL ( Table 4 ) . The average ( 95% CI ) hazard ratio adjusted for confounders for development of PKDL was 2 . 7 ( 0 . 9–8 . 1 ) , 3 . 4 ( 1 . 2–9 . 5 ) , 3 . 5 ( 1 . 2–10 . 0 ) , 5 . 8 ( 2 . 1–15 . 8 ) , 6 . 0 ( 2 . 2–16 . 4 ) , 6 . 2 ( 2 . 3–16 . 7 ) , and 8 . 0 ( 3 . 0–21 . 1 ) for MDAMB , MF , AMB+PMIM , SDAMB , AMB+MF , PMIM , and PMIM+MF , respectively ( Table 4 ) . The MDAMB’s hazard ratio did not differ statistically significantly with the reference cohort . However , hazard ratio of all other cohorts was statistically significant ( Table 4 ) . We did not find any covariates as a statistically significant factor for PKDL development ( Table 4 ) . Further , when compared existing treatment regimens excluding SSG and using MDAMB as a reference , all treatment regimens had higher hazard ratio for PKDL which was statistically significant for AMB+MF , PMIM and PMIM+MF ( Table 5 ) , borderline significant for SDAMB and statistically insignificant MF and AMB+PMIM ( Table 5 ) . The average trend for development of PKDL peaked at year 3 since treatment for VL ( Fig 1 ) . Interestingly , when stratified by cohorts , the PMIM , AMB+MF , and MDAMB cohorts showed upward trends for PKDL development ( Fig 1 ) . Of the 984 participants , 69 had VLR with a median time 1 . 05 years ( IQR , 0 . 77–1 . 53 ) for 1 364 974 person-days of observation ( mean 95% CI , 1387 , 1371–1404 ) ( Table 6 ) . Overall incidence proportion of VLR was 7% ( 95% CI , 7 . 0% , 5 . 6%–8 . 8% ) . The SSG cohort had lowest incidence proportion ( mean , 95% CI ) for VLR ( 1 . 8% , 0 . 6%–5 . 5% ) , followed by the MDAMB ( 2 . 7% , 0 . 8%–8 . 0% ) , AMB+PMIM ( 3 . 6% , 1 . 3%–9 . 3% ) , PMIM+MF ( 4 . 8% , 2 . 0%–11 . 1% ) , AMB+MF ( 6 . 7% , 3 . 2%–13 . 5% ) , SDAMB ( 7 . 9% , 4 . 3%–14 . 2% ) , PMIM ( 14 . 3% , 8 . 7%–22 . 5% ) , and MF arms ( 14 . 7% , 9 . 8%–21 . 4% ) . The overall incidence rate ( 100-person-4years ) of VLR was 7 . 6 ( 95% CI , 4 . 1–14 . 7 ) . SSG cohort had lowest VLR incidence rate ( rate , 95%CI ) ( 1 . 8 , 0 . 6–5 . 6 ) , followed by the MDAMB ( 2 . 7 , 0 . 9–8 . 4 ) , AMB+PMIM ( 3 . 7 , 1 . 4–9 . 8 ) , PMIM+MF ( 5 . 0 , 2 . 1–12 . 0 ) AMB+MF ( 7 . 0 , 3 . 3–14 . 7 ) , SDAMB ( 8 . 4 , 4 . 5–15 . 6 ) , PMIM ( 16 . 0 , 9 . 6–26 . 5 ) , and MF ( 16 . 3 , 10 . 7–24 . 7 ) cohorts ( Table 6 ) . The Cox proportional hazard ratio ( mean , 95% CI ) for VLR incidence adjusted for covariates considering the SSG arm as reference , was lowest for the MDAMB arm ( 1 . 3 , 0 . 3–6 . 3 ) , followed by the AMB+PMIM ( 1 . 9 , 0 . 4–8 . 7 ) , PMIM+MF ( 2 . 2 , 0 . 5–9 . 2 ) , AMB+MF ( 3 . 5 , 0 . 9–13 . 5 ) , SDAMB ( 3 . 5 , 1 . 0–13 . 0 ) , MF ( 7 . 5 , 2 . 2–25 . 6 ) , and PMIM ( 7 . 7 , 2 . 2–27 . 1 ) cohorts ( Table 7 ) . The hazard ratio of the MDAMB , AMB+PMIM , PMIM+MF , AMB+MF , and SDAMB arms for VLR was higher but statistically insignificant , but the higher hazard ratio of the MF and PMIM arms was statistically significant ( Table 7 ) . None of the covariates had any significant association with VLR development . Using MDAMB as a reference we found higher but statistically insignificant hazard ration for VLR of AMB+PMIM , PMIM+MF , AMB+MF and SDAMB whereas hazard ration for VLR of MF and PMIM was six times higher and statistically significant ( Table 8 ) . The overall trend of VLR peaked at 1 year and then declined thereafter . The VLR trend had a similar pattern for all study arms ( Fig 2 ) .
Key findings of the current study are: there was a significant relationship between the treatment regimens for VL and the development of PKDL and VLR; and the socio-demographic factors investigated in this study did not have any relationship with PKDL nor with VLR development . The study is unique and included all treatment regimens so far for VL . SSG was the only treatment for VL for a century . It became less efficacious over time due to parasite resistance to SSG . Further , the World Health Organization ( WHO ) Expert Committee on the Control of Leishmaniases ( WHOECCL ) does not recommend SSG for VL due to its long treatment duration and severe toxicity [25] . Though SSG for VL was protective against PKDL and VLR , we do not recommend this treatment for VL due to its severe toxicity . The SSG cohort , however , served as reference—as a cohort with the lowest burden of PKDL and VLR and facilitated comparison of PKDL and VLR by other treatment regimens . In our study , SSG resulted in 3 . 0% ( 95% CI , 1 . 3–7 . 3 ) incidence of PKDL . A study in Nepal reported a little higher incidence of PKDL after SSG treatment for VL at 5 . 4% . It also found incomplete treatment for VL as a risk factor for PKDL . 13 The difference between the two studies can be explained by the difference in study designs and populations . In our study , all patients had complete treatment with SSG; this was not case in the Nepalese study . A population-based study in Bangladesh found a cumulative incidence of PKDL of 17% for 5 years [7] . The study had 1002 VL patients treated with SSG , MF , and AMB . The study did not stratify PKDL incidence by treatment regimens . MF monotherapy was introduced in the KAEP after successful completion of phase III and phase IV studies [26–27] . Decreased efficacy of MF; high rate of treatment incompliance; adverse reactions , including renal toxicity and hepatotoxicity; and the availability more safe and effective drug ( AMB ) , led the WHOECCL not to recommend MF monotherapy for treatment of VL . 25 The highest rate of VLR after MF monotherapy in this study ( 16 . 3% ) , further justify the recommendation of the WHOECCL [25] . PMIM monotherapy was developed by the One World Health for treatment of VL [23] . We observed very high rates of PKDL ( 20% ) and VLR ( 16% ) with this treatment for VL . Our observation supports the WHOECCL , which did not recommend PMIM monotherapy for VL [25] . Availability of AMB , a highly safe and effective drug for treatment of VL , changes the scenario for VL case management . Currently , this is the drug of choice for VL case management in the KAEP . AMB is expensive; the national programs of Bangladesh , India , and Nepal get it as a donation from the developer Gilead Sciences , Inc . through WHO . The KAEP has three phases: the attack phase , consolidation phase , and maintenance phase [24] . During the attack phase , the VL case burden was 21 times higher than the VL elimination target [28] . Bangladesh and Nepal completed the attack phase and achieved the target . AMB in a single intravenous infusion at a dose of 10 mg/kilogram body weight is the first treatment option for VL in Bangladesh . This treatment regimen was particularly suitable in the attack phase due to its high safety and efficacy for VL , 100% compliance , and 1 to 2 days of patient hospitalization . During the attack phase of the NKEP when VL burden was very high , SDAMB was the most suitable VL treatment option . However , our study highlighted a concern about its continuation during the consolidation and maintenance phases of the program , as it resulted in a very high incidence of PKDL ( 17% ) and VLR ( 8 . 4% ) . Combination therapy with AMB+MF , AMB+PMIM , and PMIM+MF was introduced for treatment of VL by the Drugs for Neglected Diseases initiative . Present study provides important findings for the first time that PKDL and VLR also develop after different combination therapies for VL . The AMB+MF and PMIM+MF combinations resulted in a very high incidence of PKDL , 17% and 25% respectively . The combination of AMB+PMIM gave better results; the rate of PKDL and VLR was 11 . 3% and 3 . 7% respectively . Among the combination regimens , AMB+PMIM had the least incidence of PKDL and VLR . It is interesting that combination therapies showed different patterns in terms of PKDL and VLR development . The MDAMB treatment regimen for VL had the lowest incidence of PKDL and VLR compared to all other treatment regimens except SSG . MDAMB resulted in 8 . 2% PKDL and 2 . 7% VLR for 4 years . The less incidence of PKDL by MDAMB compared to that by SDAMB could be explained by the findings from a recent experimental study . The study found that the skin pharmacokinetics of AmBisome was different when AmBisome was given as a single dose and as a multidose for treatment of murine cutaneous leishmaniasis [29] . AmBisome when it was given in a multidose it resulted in a better accumulation of the drug in the skin , more reduction in skin parasite load and skin lesion size [29] . A study in India reported even less incidence of PKDL after treatment of VL with MDAMB at a dose of 20 mg/kilogram body weight . Our MDAMB cohort received 15 mg/kilogram body weight of AMB for treatment of their VL . The study in India had PKDL cases who passively reported to the health facility . Therefore , there may have been under-reporting; this could be another reason for the lower PKDL incidence in that study [30] . All cohorts showed a similar trend for VLR: VLR peaked in the first year after treatment . However , PKDL development peaked in the third year after treatment for VL , but this was not the case when results were stratified by treatment regimens . The PKDL trend continued upward with MDAMB , PMIM , and AMB+MF . This necessitates follow-up of cured VL patients for at least 3 years by the NKEP for early detection of PKDL and VLR cases . This also demands longer observation MDAMB , PMIM , and AMB+MF cohorts to find the moment of a downward trend . Our study has some limitations . The entire SSG and partial MF arms were retrospective cohorts , whereas all other arms were prospective cohorts . The SSG and MF arms were conducted when the VL burden was comparatively higher in the country . Since PKDL and VLR are consequences of VL , it would bias study results if the study aimed to survey the PKDL/VLR burden in the community . Our study aimed to investigate the PKDL/VLR incidence only in cured VL patients; therefore , results are free from biases related to time . Another limitation of the study is that the mean days of observation differed in the arms due to different sample sizes . Therefore , a cohort with higher mean days of observation should have a higher incidence of PKDL/VLR . Interestingly , the SSG arm had the highest mean days of observation and the least incidence of PKDL and VLR . The treatment regimen therefore dictated the incidence of PKDL and VLR . The study had been carried out using cured VL patients of the clinical trials in Bangladesh . So , external validity of the study is yet to be established and its results may not be generalizable for other countries . In summary , SSG and MDAMB for VL had least incidences of PKDL and VLR . MDAMB had least hazard ratio for PKDL development compare to other treatment regimens . Since SSG is no more recommended for VL , MDAMB should be the choice for VL in the consolidation and maintenance phases of the NKEP in Bangladesh until better molecules than AMB are found . Therefore , we highly recommend MDAMB for treatment of VL for NKEP in Bangladesh during its consolidation and maintenance phases .
|
Post-kala-azar Dermal Leishmaniasis ( PKDL ) , a sequale of visceral leishmaniasis ( VL ) , and reappearance VL ( visceral leishmaniasis relapse , VLR ) are intra-epidemic reservoirs of VL and threats control of VL in long run . Currently there is no strategy for prevention of PKDL and VLR . If a relationship between treatment for VL and development of PKDL and VLR is there , and then selection of proper treatment regimen for VL should prevent PKDL and VLR . So far no study has been carried out to investigate the relationship between treatment regimens for VL and development of PKDL and VLR . The study demonstrated that multi-dose liposomal amphotericin B ( AmBisome ) defined as MDAMB herein for VL results in least PKDL and VLR among all existing and recommended by WHO treatment regimens for VL . We recommend adaptation of MDAMB in the national visceral leishmaniasis elimination program for VL cases management during subsequent phases of the national program when VL burden is low and hospitalization of VL patients for 3-5-days is now feasible .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"kala-azar",
"antimicrobials",
"medicine",
"and",
"health",
"sciences",
"drugs",
"tropical",
"diseases",
"microbiology",
"geographical",
"locations",
"antifungals",
"parasitic",
"diseases",
"data",
"management",
"pharmaceutics",
"neglected",
"tropical",
"diseases",
"pharmacology",
"veterinary",
"science",
"bangladesh",
"veterinary",
"medicine",
"infectious",
"diseases",
"mycology",
"computer",
"and",
"information",
"sciences",
"zoonoses",
"amphotericin",
"livestock",
"care",
"protozoan",
"infections",
"people",
"and",
"places",
"asia",
"leishmaniasis",
"microbial",
"control",
"biology",
"and",
"life",
"sciences",
"drug",
"therapy"
] |
2019
|
Relationship between treatment regimens for visceral leishmaniasis and development of post-kala-azar dermal leishmaniasis and visceral leishmaniasis relapse: A cohort study from Bangladesh
|
Broadly neutralizing antibodies ( bnAbs ) are thought to be a critical component of a protective HIV vaccine . However , designing vaccines immunogens able to elicit bnAbs has proven unsuccessful to date . Understanding the correlates and immunological mechanisms leading to the development of bnAb responses during natural HIV infection is thus critical to the design of a protective vaccine . The IAVI Protocol C program investigates a large longitudinal cohort of primary HIV-1 infection in Eastern and South Africa . Development of neutralization was evaluated in 439 donors using a 6 cross-clade pseudo-virus panel predictive of neutralization breadth on larger panels . About 15% of individuals developed bnAb responses , essentially between year 2 and year 4 of infection . Statistical analyses revealed no influence of gender , age or geographical origin on the development of neutralization breadth . However , cross-clade neutralization strongly correlated with high viral load as well as with low CD4 T cell counts , subtype-C infection and HLA-A*03 ( - ) genotype . A correlation with high overall plasma IgG levels and anti-Env IgG binding titers was also found . The latter appeared not associated with higher affinity , suggesting a greater diversity of the anti-Env responses in broad neutralizers . Broadly neutralizing activity targeting glycan-dependent epitopes , largely the N332-glycan epitope region , was detected in nearly half of the broad neutralizers while CD4bs and gp41-MPER bnAb responses were only detected in very few individuals . Together the findings suggest that both viral and host factors are critical for the development of bnAbs and that the HIV Env N332-glycan supersite may be a favorable target for vaccine design .
The humoral immune response to HIV-1 infection comprises in a subset of individuals broad and potent neutralizing antibodies ( bnAbs ) [1–6] . The elicitation of such Abs prior to infection would presumably protect against infection by most circulating HIV strains and is thus considered one of the highest priorities of the HIV vaccine research field [7–10] . However , thus far , no vaccine candidate has been successful at eliciting bnAbs . Therefore , understanding the development of bnAbs and the clinical , immunological and virological correlates of their elicitation during natural infection is likely to be crucial for the design of a protective vaccine [11 , 12] . Broadly nAb responses usually develop after 2 to 4 years of HIV infection , in 10 to 20% of individuals [13–21] . Development of neutralization breadth has been mainly associated with high viral load and low CD4+T cell counts [17–20 , 22] . An association with greater viral diversity in the env coding region at early time-points after infection has also been reported [13 , 18 , 23] and particular viral sequences or features may favor the emergence of bnAb responses [24] . However , the contribution of parameters such as HIV subtype , host genetic background and immune factors is less documented [25] , mostly due to the small numbers of participants , lack of adequate longitudinal sampling and of geographic and demographic diversity in most cohorts studied to date . Furthermore , while an increasing number of studies have focused on the detailed mapping of broadly neutralizing specificities and shown that bnAbs mainly target 5 regions of Env: the CD4 binding-site ( CD4bs ) , the V3-high mannose patch , the V2 apex , the gp41 MPER and the gp120/gp41 interface [26 , 27] , it still remains to be determined whether these different specificities follow similar developmental pathways in all individuals . To better understand the process leading to the development of bnAbs in natural infection , and identify broadly neutralizers for further in depth longitudinal studies , we studied clinical and immunological correlates of breath development and mapped the specificity of the bnAb responses in the IAVI Protocol C cohort , the largest ( N = 439 ) and most diverse longitudinal primary infection cohort studied to date for heterologous neutralization .
The IAVI Protocol C recruited for longitudinal follow up 613 participants with documented acute and very early HIV-1 infection ( Material and Methods ) , through 9 clinical research centers in Kenya , Rwanda , South Africa , Uganda , and Zambia ( S1A Fig in S1 Text ) . Participants were characterized in terms of demographics , HIV infection risk factors , clinical history , CD4 counts , viral load and disease progression , as well as HLA genotype [28 , 29] ( S1B-D Fig in S1 Text , S1 Table in S1 Text ) . For the present study , 439 eligible participants were evaluated for plasma neutralizing activity starting at month 24 post-estimated date of infection ( EDI ) ( S1E Fig in S1 Text , Material and Methods ) , using a 6 cross-clade pseudovirus panel predictive of neutralization breadth on larger panels [16] , which previously allowed the identification of IAVI Protocol G elite neutralizers and isolation of potent and broad nAbs [1 , 2 , 30] ( Material and Methods ) . Neutralization was measured in 2220 unique samples ( 1–12 samples per donor , mean = 5 . 0 ) representing a mean follow-up of 49 . 4 months post infection ( mpi ) ( range 24–90 months ) . Among the 439 participants , 228 ( 52% ) were tested at least up to 48 months post-EDI ( mean 64 . 3 months , range 48–90 months ) ( S2A-B Fig in S1 Text ) , defining the M48+ subset . A neutralization score taking into account both breadth and potency was assigned to each plasma sample tested , as previously described [16] ( Material and Methods ) , and used to rank each participant based on the peak of breadth ( S3A Fig in S1 Text ) . As expected , the score on the 6-virus panel was significantly correlated with neutralization breadth on a medium-sized in-house panel ( N = 37 ) and on a larger reference virus panel [31] ( N = 105 ) ( S4 Fig in S1 Text ) . A score ≥1 approximately predicted a breadth ≥50% on the large 105-virus panel . Overall , 11% ( 46/439 ) of study participants achieved broad neutralization ( score ≥1 ) at some point during the course of infection , including 7 individuals with a score ≥ 2 . Twenty-five percent ( 111/439 ) of participants also acquired moderate neutralization breadth ( score ≥0 . 5 and <1 ) ( Fig 1A ) . While nearly half of the participants ( 204/439 ) displayed low neutralization breath ( 0 < score < 0 . 5 ) , a small fraction ( 8% , 35/439 ) did not neutralize any virus on the panel . When focusing on the M48+ subset , we found that 15% ( 36/228 ) of participants reached a score ≥1 , while 35% ( 79/228 ) reached a score ≥ 0 . 5 and <1 , as expected from exclusion of participants who might not have reached their best level of neutralization yet . At month 24 post-EDI , when we started assessing neutralization , only a small subset of individuals had developed breadth , with a score ≥1 in 1–2% participants , confirming that early development of broadly neutralizing responses is rare . In the overall cohort , the average neutralization score further increased gradually over time to peak at month 48 post-EDI , and appeared to plateau or only increase incrementally thereafter ( Fig 1B ) . As the participants did not always comply with visit schedule , we were unable to systematically test all of them for the same time points . Nonetheless , limiting the analysis to participants all tested for the same visits gave virtually identical results ( S3B Fig in S1 Text ) . Similarly , at an individual level , high neutralization scores ( ≥1 ) were typically achieved between 2 and 4 years post-infection ( mean 3 . 5 years , range 24–78 months ) ( Fig 1C , S3A Fig in S1 Text ) . In most individuals reaching a score of 1 or greater , further neutralizing activity either plateaued or decreased ( Fig 1C , S3C Fig in S1 Text ) . We identified 7 neutralizers in the elite/sub-elite category ( scores ≥2 ) over the study period ( Fig 1C ) . We then conducted a thorough statistical analysis to identify potential associations between the development of bnAb responses and a number of clinical parameters ( Material and Methods ) . As suggested by the kinetics of breadth development described above , time from EDI and number of time points tested were significantly associated with best neutralization score ( S5A Fig in S1 Text ) . For the M48+ subset , this association was no longer significant ( S5A Fig in S1 Text ) and we therefore restricted our further analysis to these individuals , in order to improve accuracy by excluding as much as possible participants for whom follow up time was not long enough to permit breadth development . A Generalized Linear Model ( using a Gamma distribution with Log link function ) was chosen to model neutralization scores , which are positive-valued . Country was excluded in favor of recruitment site and subtype as they were found to be highly correlated ( country of origin and recruitment center ( ρ = 0 . 85 , p≈0 ) ; country of origin and infectious subtype ( ρ = 0 . 90 , p≈0 ) ) . Given the high number of variables , p-values from bivariate analyses were adjusted for False Discovery Rate ( FDR ) and only parameters with q-values <0 . 1 were selected for further analyses . The bivariate GLM analyses revealed that set-point viral load , HIV-1 subtype and HLA-A*03 genotype were significantly associated with the neutralization score ( Fig 2A , S5B-C Fig in S1 Text ) . In contrast , age at time of infection , sex , mode of transmission , recruitment site , other HLA and KIR types , and CD4+ T cell count at set-point were not significantly associated ( Fig 2A , S5B-C Fig in S1 Text ) . The neutralization score was further significantly associated with viral load at any visit from month 6 to 48 post-EDI and with the area under curve ( AUC ) for viral loads between month 6 and 48 ( S5D Fig in S1 Text ) . In contrast to viral load , CD4 T cell counts were inversely associated with neutralization score , and only past 6 months post-EDI , although there was a trend for an association at setpoint and month 6 post-EDI . An inverse association between CD4_AUC and score was also detected . A multivariable GLM analysis further showed that viral load at setpoint remained strongly correlated with neutralization breadth while HIV subtype C and HLA-A*03 genotype became barely significant ( Fig 2A ) . To evaluate the impact of clinical parameters on the kinetics of bnAb response development , we finally compared the time post-infection necessary to reach various levels of neutralization score across different subgroups , using Kaplan-Meier curves with Log-rank test ( Fig 2 , S6 Fig in S1 Text ) . A significant difference was found only for CD4 T cell count at setpoint , individuals with lower CD4 counts developing neutralization score ≥ 0 . 5 faster than individuals with high CD4 counts ( Fig 2 ) . No difference was observed between subgroups for the time to reach broad ( score ≥ 1 ) neutralization . However , the number of individuals included in the latter analysis was very limited . We recently showed , studying the same cohort , that individuals who develop a bnAb response have significantly higher percentages of circulating PD-1+CXCR3−CXCR5+ memory Tfh cells , suggesting that these individuals may be intrinsically more prone to mount antibody responses of greater quality [32] . Therefore , we studied whether this association may be reflected in greater binding titers to HIV Env . As shown in Figs 3A and S7A Fig in S1 Text , neutralization scores were strongly correlated with plasma anti-gp120 , -gp41 and–p24 IgG binding ELISA titers in samples from time points matching development of bnAb responses ( M24-72 , median = 36mpi ) . No correlation was found between score and anti-gp120 IgG avidity index ( Fig 3B , S7A Figure in S1 Text ) , suggesting that the greater binding titers found in broad neutralizers may not correspond to responses of greater affinity but to a quantitative rather than qualitative difference in the Ab response . We thus looked at potential associations with total plasma Ig titers and found that anti-gp120 and -gp41 titers as well as neutralization score were strongly correlated with total plasma IgG titers in these individuals ( Fig 3C and 3D ) . Correspondingly , the data showed that total plasma IgG titers correlated with VL at set-point ( Fig 3E ) . As previously reported , anti-Gag p24 IgG responses did not correlate with total plasma IgG levels and negatively correlated with viral load ( S7B Fig in S1 Text ) [33] . As anti-gp120/gp41 ELISA binding titers and neutralization scores were correlated with the total IgG concentration , we normalized both values to the latter ( Material and Methods ) . The neutralization scores were still significantly associated with higher anti-Env IgG binding titers when using adjusted values , and also found to be negatively correlated with normalized anti-Gag IgG responses ( Fig 3F , S7C Fig in S1 Text ) . Greater gp120/gp41 binding titers may be explained by a higher concentration of specific anti-gp120/gp41 Abs or by the presence of Abs of greater affinity in broad neutralizers . The avidity index of the anti-gp120/41 responses still did not correlate with the normalized score ( S7D Fig in S1 Text ) , suggesting that the affinity of anti-gp120/gp41 binding Abs is overall not different between strong and weak neutralizers , though this may be confounded by differences in antibody specificities to neutralizing versus non-neutralizing epitopes on gp120 and gp41 . We then aimed to investigate the antibody specificities associated with broad neutralization in plasma with a score ≥1 on the 37v-panel , corresponding to ≥ 50% breadth on the 105v-panel ( n = 42 ) ( S4 Fig in S1 Text , S2 Table in S1 Text ) . We first asked whether the broadly neutralizing activity of the plasma could be adsorbed on recombinant monomeric gp120 ( rgp120 ) . Plasma samples were pre-incubated with rgp120 coated beads or control beads before being tested for neutralization . After verifying by ELISA that all rgp120-binding Abs had been removed , we tested the adsorbed fractions against a cross-clade virus panel ( Fig 4A , S8A Fig in S1 Text ) . About a third of the plasma samples ( 11/40 , 27 . 5% ) were efficiently depleted of broadly neutralizing activity across viruses by adsorption on rgp120 , showing that the neutralization breadth was , in those cases , clearly associated with Abs reactive with monomeric gp120 . Eleven other samples ( 27 . 5% ) were only partially depleted of neutralizing activity on rgp120 , suggesting the presence of multiple neutralizing Ab specificities in these donors or a partial match with the rgp120 used for the depletion . Finally , samples from 18/40 ( 45% ) participants retained almost complete neutralizing activity after depletion by gp120 monomers , including when rgp120s from different viral strains were used for adsorption , suggesting the presence of quaternary gp120-specific bnAbs or of gp41-specific bnAbs in these plasma . To investigate the presence of gp41-MPER specific bnAbs , we tested the selected plasma against HIV-2 chimeric pseudoviruses bearing either the full or partial HIV-1 MPER [13 , 14] . We identified 17 ( 40% ) plasma samples with neutralizing activity against the HIV-2 C1 but not HIV-2 WT , most of them ( 12/17 , 70% ) targeting the N-terminus segment ( HIV-2 C4 ) ( Fig 4A , S9A Fig in S1 Text ) . The MPER specific reactivity was further confirmed by competition of the neutralizing activity with an MPER-peptide . Eleven of the 17 plasma samples were competed by this peptide for neutralization of HIV-2 C1 ( S9A Fig in S1 Text ) but only four ( 9 . 5% ) plasma ( PC048 , PC174 , PC011 , PC031 ) were further competed for neutralization of several cross-clade HIV-1 pseudoviruses ( Fig 4A , S9B Fig in S1 Text ) suggesting the presence of bnAbs targeting the MPER . Accordingly , neutralizing activity was not absorbed by rgp120 in these samples ( S8A Fig in S1 Text ) . Peptide competition was particularly strong for participant PC031 suggesting that most of the bnAb activity was directed against the MPER for this individual . We then assessed in which plasma the presence of CD4 binding site ( CD4bs ) specific bnAbs may explain the broadly neutralizing activity . We first tested plasma binding activity to the Resurfaced Stabilized Core 3 ( RSC3 ) , a probe selective for VRC01-like CD4bs bnAbs , and a mutant ( KO-RSC3 ) with decreased CD4bs bnAbs binding capacity [3 , 34] . Forty-five percent ( 19/42 ) of the plasma specimens tested had differential RCS3/KO-RSC3 reactivity suggesting the possible presence of VRC01-like Abs ( S10A Fig in S1 Text ) . However , RSC3 failed to efficiently compete all but one donor plasma ( PC063 ) for neutralization of a small cross-clade panel of HIV-1 pseudoviruses ( Fig 4A , S10B Fig in S1 Text ) . The results agree with a recent report by Lynch and colleagues showing that although RSC3 binding activity could be found in 47% of HIV-1 infected individuals , RSC3-reactive Abs mediating broad neutralization were only detected in a few individuals [34] . Of note , a competition assay using a different gp120-core molecule , TriMut , that binds , in addition to VRC01-like bnAbs , non-broadly neutralizing CD4bs Abs like F105 or b6 [35] could compete 71% ( 30/42 ) of the plasma for neutralization of the CD4bs sensitive strain HxB2 ( Fig 4A , S10C Fig in S1 Text ) confirming that most HIV-infected individuals develop CD4bs Abs that are not broadly neutralizing . A caveat to the RSC3 competition approach above is that some CD4bs bnAbs may not bind this probe . Therefore , we also performed plasma rgp120 adsorptions in the presence of the non-broadly neutralizing Ab b6 at saturating concentrations [2] . With the exception of three donors , the adsorption on rgp120 of the broad plasma neutralizing activity was not significantly inhibited by the presence of b6 ( S8A Fig in S1 Text ) . In the case of donors PC063 and PC053 , b6 greatly competed ( >75% ) the broad plasma neutralizing activity adsorption for 6/6 and 2/6 viruses , respectively ( Fig 4A , S10D and S8A Figs in S1 Text ) . Interestingly , we also found that neutralization by PC053 and PC063 plasma was enhanced for 5/6 and 3/5 N276A-mutant pseudoviruses , respectively ( S4 Table in S1 Text ) , raising the possibility of the presence of early precursors of VRC01-like CD4bs bnAbs [36] . Together the results confirm that in broad neutralizers , although non-broad CD4bs Abs are common , CD4bs bnAbs are rare . Broadly nAbs of the PG9 class , that recognize a quaternary epitope in the V2 spike apex region , require the presence of an N-linked glycosylation site at residue 160 and viruses treated with the glycosidase inhibitor kifunensine resist neutralization by PG9-like Abs to a large extent [1 , 30 , 37 , 38] . We found 2 participants ( PC064 , PC069 ) with N160K-sensitive plasma neutralizing activity ( Fig 4A , S10A Fig in S1 Text ) . Both PC064 and PC069 plasma also showed a markedly reduced neutralizing activity against kifunensine-treated pseudoviruses and other V2 mutant ( residues 166 , 169 , 171 ) pseudoviruses ( Fig 4A , S11 Fig in S1 Text ) . Additionally , the broadly neutralizing activity of the PC064 plasma was retained following removal of gp120-specific antibodies through adsorption on rgp120s from 4 different strains ( S8A Fig in S1 Text ) . Together these results strongly suggested the presence of PG9-like bnAbs in donor PC64 . In contrast , a small but significant decrease in neutralization by the PC069 plasma was observed after rgp120 absorption , suggesting that in this case the N160 glycan-dependent apex epitope recognized by the bnAbs may be less dependent on quaternary structures than PG9 and displayed on the corresponding rgp120s . Four other samples which neutralizing activity was not depleted by rgp120 but did not depend on the N160 glycan ( PC079 , PC174 , PC035 , PC097 ) , were also sensitive to mutations at position 166 , 169 and 171 , suggesting the presence of bnAbs targeting the apex more similar to the recently described CAP256-VRC26 bnAbs [39] . We found 11 other plasma samples ( 26% ) which neutralizing activity was not depleted by rgp120 not mapping to the apex and affected to various degrees when viruses were treated with kifunensine ( Fig 4A , S11A Fig in S1 Text ) . Four of these samples ( PC041 , PC048 , PC178 , PC050 ) were significantly affected by mutations shown to impact the recently described bnAbs targeting the quaternary gp120/gp41 interface epitope [6 , 40 , 41] ( Fig 4A , S11B Fig in S1 Text ) . Interestingly , broad neutralization of PC023 plasma was also affected by these mutations although the broadly neutralizing activity could efficiently be depleted by rgp120 . Of the five remaining samples with undefined quaternary-specific bnAb responses , one sample corresponded to the participant with the greatest breadth ( PC068 ) ( Fig 4A , S8B Fig in S1 Text ) , neutralizing 97% of the viruses tested ( S4A-B Fig in S1 Text ) . A third class of glycan-dependent bnAbs recognizes the high mannose patch centered around the N332 glycan in the V3 region [2 , 20 , 30 , 42 , 43] . We identified 17 broad neutralizers ( 40% ) whose plasma neutralizing activity was significantly affected by single or double N332 supersite mutations in the context of more than 3 viruses ( Fig 4A , S11A fig in S1 Text ) , often across different subtypes . Interestingly , in half of cases , the N332-glycan dependent bnAb activity could only be slightly depleted by rgp120 adsorption ( Fig 4A ) . Plasma from participants PC076 , PC037 and PC011 displayed both sensitivity to the N332A mutation and to kifunensine treatment [44] of various viruses across different clades . Altogether , glycan-dependent neutralization ( ie affected by removal of PNG and kifunensine-sensitive ) was detected in 60% ( 25/42 ) of top Protocol C neutralizers ( Fig 4A and 4B ) .
In the present study , we investigated the development of bnAb responses against HIV-1 in the largest ( n = 439 ) and most diverse longitudinal primary infection cohort studied to date for neutralization . Overall , about 15% of Protocol C participants who were followed for at least 48 months reached a neutralization score ≥1 , roughly corresponding to more than 50% breadth on a large 105-virus panel ( S4D Fig in S1 Text ) a prevalence equivalent to that observed in previously studied cohorts ( S12 Fig in S1 Text , Amsterdam , total HIV+ cohort size N = 82; Massachusetts General Hospital , N = 17; CAPRISA , N = 40 ) [16 , 19 , 20 , 22] . In addition , a moderate neutralization breadth ( score = 0 . 5–1 ) , corresponding to about 20 to 50% breadth on a large virus panel , developed in another third of the donors . This observation is consistent with studies suggesting that some degree of breadth develops in a large proportion of HIV-infected individuals [21 , 45] . Protocol C participants who had not developed breadth by year 4 were unlikely to do so thereafter as the cohort average neutralization score plateaued at 48 months , with only a slight increase in the proportion of the highest scores afterwards , essentially due to an augmentation in neutralization potency in a few donors . Of all the subjects who developed scores ≥1 , only three did so past month 48 . While the development of breadth usually takes at least 2 years possibly due to the stochastic nature of the bnAbs maturation process and Env evolution , this general lack of development of broadly neutralizing Ab responses later in infection may reflect an increased disruption of CD4 and B cell responses as the infection progresses , as seen with decreased responses to vaccination [46 , 47] . While neutralization breadth in the Amsterdam and MGH cohorts , both predominantly composed of subtype B-infected men having sex with men ( MSM ) participants , was found to emerge between 1–2 years of infection [18 , 19 , 22 , 48] , participants in our study developed breadth on average 3 years post-infection , similar to the CAPRISA cohort which is composed of subtype C-infected high-risk women [20] . Differing parameters such as mode of transmission , HIV-1 subtype , host genetics , general health or other concomitant infections could account for the difference between the Caucasian and African cohorts . In agreement with previous publications , we found the development of neutralization breadth to be most strongly correlated with viral load both at setpoint and all further time points tested [15 , 17 , 18 , 20 , 23 , 45] . Viral load is known to be influenced by the host HLA genotype [49–55] and the nature of the transmitted virus [56–59] . Our statistical analysis identified infection by subtype C viruses and HLA-A*03 genotype to be associated with neutralization breadth . However , the relatively low significance of these associations in our multivariate analysis suggests that the correlation may be essentially due to the impact of these parameters on viral load itself [28 , 60] . Nevertheless , particular phenotypic and genotypic features of subtype C transmitted/founder viruses have been described that could potentially favor the development of neutralization breadth [61–64] . In particular , not all individuals with high viral load developed bnAb responses , suggesting that other factors are at play or , that an earlier disruption of immune responses in some individuals may prevent bnAb development as discussed below . An association between the development of breadth and low CD4 T cell levels at various time points has been described in some studies [18–20 , 23 , 31] . Interestingly , in our cohort , in contrast to the association with high viral load , the development of bnAb responses significantly correlated with low CD4 T cell counts only at later time points , past 6 months of infection ( S5D Fig in S1 Text ) . An intrinsic difference in CD4 levels between donors developing broad neutralization or not cannot be totally ruled out as CD4 levels prior to infection were not available [18] , but it is tempting to hypothesize that the VL drives the association with breadth and that the low CD4 levels are merely the reflection of the high viral replication and disease progression [18 , 22 , 65] . However although CD4 T cell levels were for the most part lower in individuals developing bnAbs , we have shown recently that Protocol C broad neutralizers have significantly higher relative frequencies of a population of blood memory CD4 Tfh cells [32] . This is consistent with observations from other studies [19 , 66] and together with the high level of somatic hyper-mutations ( SHM ) found in most bnAbs , suggests that an intrinsic greater ability to provide help to B cells in certain individuals may favor the generation of highly affinity-matured antibody responses and thus the elicitation of bnAbs . Here we showed that bnAb responses were associated with greater titers of Env ( gp120 and gp41 ) and lower Gag binding Abs in ELISA , even after normalization for total Ig concentration , which as previously shown , also correlated with broad neutralization [67] . The absence of difference in avidity index of anti-Env responses between strong and weak neutralizers suggests that the higher binding titers found in individual with breadth may not be due to the presence of more highly affinity-matured Abs but rather , to a larger diversity of Abs , which may favor the emergence of bnAb responses by exerting cooperative pressure on the virus [68] or by limiting the escape landscape that the virus can explore in response to neutralization . Indeed , a number of escape mutations from nAbs lead to the exposure of epitopes that are usually occluded on the Env trimer and that are the target of Abs elicited by monomeric gp120 and gp160 . In this sense , a greater diversity of such Abs may put a pressure on the virus , limiting pathways of escape in a manner that may favor the selection of nAbs targeting conserved regions exposed on the trimer . Further mapping of anti-Env specificities present in weak and broad neutralizers as well as comparison between SMH levels in anti-Env Abs in these donors will help answer these questions . Taken together , these data suggest a model where a higher level of chronic antigenic stimulation over a prolonged time may lead to the activation of a greater number of naïve B cells , and increase the probability , in a stochastic model , to activate cells bearing a germline BCR more amenable to development into a bnAb lineage . In addition , a high level of antigenic stimulation may further impact the B cell selection process in germinal centers and increase the likelihood of the selection of cells on the path to broad neutralization . Alternatively or concomitantly , a high level of viral replication may lead to a greater or differing stimulation of innate pathways that could favorably impact the humoral response ( different “adjuvant effect” , better antigenic presentation ) . A higher viral load may also contribute to the generation of a greater antigenic diversity ( currently under investigation in Protocol C ) which may favor the selection of Abs with broadly neutralizing activity , as suggested by some studies [23 , 39 , 69 , 70] . Alternatively , diversity may be the consequence of the elicitation of a broad Ab response [48 , 71 , 72] . Both processes could be co-dependent , Env escape mutants being selected in response to neutralizing Abs and a greater diversity of escape mutants leading to the selection of a greater variety of Abs in a cycle increasing the probability of eliciting bnAbs . In accordance with previous mapping studies , we found that for most Protocol C broad neutralizers , the bnAb activity essentially mapped to one or a limited number of Ab specificities in each individual [2 , 19 , 20 , 73 , 74] . However , we were unable to clearly map the neutralizing antibody response in about 12% of broad neutralizers suggesting that Abs of not yet known specificity were responsible for the breadth in these donors or , that the bnAb activity in these plasma was due to the presence of Abs of several different specificities [13 , 68 , 69 , 75 , 76] . The prevalence of each bnAb specificity in our cohort was also in line with other studies [2 , 13 , 18–20 , 22 , 34 , 48 , 74 , 77–79] . Although most Protocol C top neutralizers developed antibodies to the CD4bs , very few developed broad CD4bs Abs . CD4bs bnAbs usually bear an exceptionally high level of SHM and may require more time to develop than bnAbs targeting other epitopes . In addition , structural constraints and the need to use mainly a unique VH family likely further limit the probability of developing such bnAbs . Accordingly , the Protocol C participant who developed broad neutralization targeting the CD4bs ( PC063 ) appeared to do so relatively late , at month 66 post-EDI [2–4 , 19 , 34 , 74 , 76 , 80] . Similar to the CD4bs , only one study participant with a broadly neutralizing response mapped clearly to the gp41-MPER . The proximity to the membrane , the important structural constraints that an antibody needs to circumvent to reach the gp41-MPER and potential self-reactivity issues are likely responsible for the paucity of this type of bnAb response [81] . In contrast , we found that glycan-dependent bnAb specificities ( i . e . N332-glycan supersite , V2 Apex , gp120/41 interface , other kifunensin-sensitive specificities ) largely dominated ( 60% ) the bnAb responses in top neutralizers in our cohort . Within the glycan specificities , we identified 6 donors ( 15% ) with bnAbs targeting the trimer apex like PG9 ( N160-glycan dependency and kifunensine sensitivity ) or CAP256-VRC26 bnAbs ( sensitivity to 166/169/171 mutations , low dependency on N160-glycan , low sensitivity to kifunensine ) . The frequency of this type of responses compared to CD4bs or MPER bnAbs responses suggests that the apex may be an interesting vaccine target , particularly in light of the newly developed soluble native Env trimers properly presenting the apex epitopes [82 , 83] . We also identified several individuals ( 25% ) with bnAb responses that were not adsorbed on rgp120 monomers , with varying levels of kifunensine sensitivity and not mapping to either the MPER nor the trimer apex , similar to the recently described bnAbs PGT151-8 , 8ANC195 and 35O22 targeting discrete epitopes the gp120/gp41 interface [6 , 40 , 41] . However , mutations previously shown to impact binding of these mAbs , did not result in a clear phenotype suggesting that the bnAbs in these donors , including the best Protocol C neutralizers , may target either yet other discrete epitopes at the gp120/gp41 interface or at the apex , or a novel epitope of the Env trimer . On its own , the N332-glycan region accounted for nearly 40% of the broadly neutralizing specificities in Protocol C . Additionally , some participants had a mixed signal for bnAbs targeting this region ( PC002 , PC025 PC008 , PC080 , PC049 ) suggesting the presence of glycan-specific nAbs of narrower breadth and that N332 supersite Abs are significantly more easily elicited than any other broad specificity . A greater accessibility on the Env spike may explain the higher prevalence for this region , with the ability for Abs to reach from various angles and potentially allowing the use of a number of different Ab gene families . Furthermore Abs to the glycan patch have been shown to be promiscuous in their binding to different glycans of this region which may also favor recognition by a greater number of Abs ( although whether this is the cause or the consequence of the elicitation of such antibodies is arguable ) [84 , 85] . Together our findings confirm in a large African cohort with a diverse range of infecting HIV-1 subtypes that a combination of viral and host factors is likely to be necessary for the development of a broadly neutralizing antibody response to HIV-1 , explaining why only a fraction of infected individuals develop high levels of such responses . Although in a few individuals neutralization breadth increases over a relatively short time , less than 12 months , in most cases bnAb responses develop around 3 years post-infection , possibly due to the necessity of prolonged antigenic stimulation , and it remains to be seen whether more favorable kinetics may be elicited through efficient vaccination regimens . The development of neutralization breadth may represent a fine balance between a high viral replication needed for antigenic stimulation but leading to a faster decline of the immune system , and the necessity of having immune responses conserved long enough to efficiently elicit bnAbs . As an optimistic note , in healthy individuals the elicitation of bnAbs may not be limited by the crippled immune system found in HIV-infected individuals and may be successful in a larger fraction of individuals . A detailed analysis of the development of bnAb lineages in top neutralizers will help understand which specificities are most amenable to elicitation through vaccination and whether Env evolution pathways associated with specific lineages suggest particular immunogen designs or vaccine strategies . Studying antibody developmental pathways in various individuals sharing the same broad specificity will also be critical , as finding similarities between donors in Env evolution or in the nature of the Env triggering the broad lineage would strongly suggest a path for immunogen design . Our study suggests that the glycan patch surrounding the N332 glycan is the most favorable target for vaccines and should be a high priority for immunogen design .
The IAVI-sponsored Protocol C cohort participants were selected through rapid screening of adults with a recent history of HIV exposure for HIV antibodies in Uganda , Rwanda , Zambia , Kenya and South Africa [28] . After obtaining written informed consent , blood samples were collected from HIV-1 infected participants quarterly for the first two years and every 6 months thereafter . The study was reviewed and approved by the Republic of Rwanda National Ethics Committee , Emory University Institutional Review Board , University of Zambia Research Ethics Committee , Charing Cross Research Ethics Committee , UVRI Science and Ethics Committee , Kenyatta National Hospital Ethics and Research Committee , KEMRI Scientific and Ethics Review Unit , University of Cape Town Research Ethics Committee , University of Kwazulu-Natal Biomedical Research Ethics Committee , Mahidol University Ethics Committee , Sanford-Burnham Medical Research Institutional Review Board , Veterans Affairs San Diego Institutional Review Board , LIAI Human Subjects Committee , and Scripps Institutional Review Board . Between February 2006 and December 2011 , 613 participants were enrolled in Protocol C and over 7 , 600 time points were sampled for plasma , plasma and PBMCs . Median time from EDI to enrolment was 54 days ( mean 81 . 7 days , range 10 to 396 days ) . Protocol C participants eligible for this study were age 18 or older , with >24 months of follow-up and antiretroviral therapy ( ART ) naïve ( N = 439; 232 ( 52% ) are still on-study ) . Overall the participants included in this study ( 439/613 , 72% ) were highly representative of the entire cohort population regarding gender , age , mode of transmission , infectious subtype , clinical site , viral load and CD4 T cell count ( S1B-D Fig in S1 Text , S1 Table in S1 Text ) . Visits were scheduled and coded based on the number of months post infection ( MPI ) , calculated from the estimated date of infection ( EDI ) . Visits that deviated from schedule kept the original coding . However , we verified that despite these deviations from the scheduled visit calendar , the mean of adjusted MPIs within each group was in good accordance with the VC and that each VC group was statistically distinct from the previous and the next ( S2D Fig in S1 Text ) . Data were collected at every study visit , including HIV risk behavior ( baseline only ) , demographics , symptom-directed examinations including data on comorbidities and opportunistic infections , CD4 T cell count and viral load . Although enrollment closed in 2011 , the longitudinal follow-up continues . Neutralizing activity in longitudinal Protocol C samples was assessed with a recombinant virus assay ( Monogram Biosciences , LabCorp ) using a reference panel of full-length env of viruses previously selected to stratify infected individuals by the breadth and potency of their nAb response [16]: 92TH021 ( CRF0-AE ) , 94UG103 ( Clade A ) , 92BR020 ( Clade B ) , JRCSF ( Clade B ) , IAVIC22 ( Clade C , also named MGRM-C026 ) and 93IN905 ( Clade C ) . Briefly , pseudoviruses capable of a single round of infection were produced by co-transfection of HEK293cells with a sub-genomic plasmid , pHIV-1lucΔu3 , that incorporates a firefly luciferase indicator gene and a second plasmid , PC0XAS that expressed HIV-1 env clones . Following transfection , pseudoviruses were harvested and used to infect U87 cell lines expressing CD4 plus the CCR5 and CXCR4 co-receptors . Serial 3-fold dilutions of plasma , starting 1:100 , were assessed for neutralization against each of the 6 viruses listed above and NL43 , a Tier-1A subtype-B virus , as a positive control . Virus infectivity was determined 72h after inoculation by measuring amount of luciferase activity . Positive neutralization was defined as 50% inhibition of infection of an HIV strain at a 1:100 plasma dilution and when the percent inhibition was at least 1 . 7 fold higher than percent inhibition against the specificity control , aMLV . The level of neutralizing activity of an individual sample was determined by a neutralization score defined as a weighted average of log-transformed 50% neutralization end point dilutions across the reference pseudoviruses neutralization screening panel and excluding the negative and positive controls , aMLV and NL43 respectively: ( Score = Average ( log3 ( dilution/100 ) +1 ) ) . All titers below the limit of detection were assigned a value of 33 for purposes of calculating a neutralization score . Log-transformed values ranged from 0 . 0 to 4 . 0 with 0 . 0 representing a sample with undetectable titers to a given pseudovirus as described previously [16] . Samples were equally tested across clinical sites ( S1B and S2C Figs in S1 Text ) . Statistical analyses were performed using free software R Bioconductor , version 3 . 0 . 1 and GraphPad Prism 6 . The analysis was performed on the best neutralization score ( ranging from 0–2 . 33 ) across all time points tested for all Protocol C participants included in this study ( N = 439 ) and the M48+ subset ( N = 228 ) . Factors analyzed included participant age at EDI , gender , risk group , viral load ( VL , log10-transformed ) and CD4 T cell count at setpoint ( defined as the first measurement between days 70 and 350 from the estimated of infection ) , HLA genotype , KIR genotype , infecting HIV subtype , time post-infection of the last time point tested for neutralization ( Follow-Up time ) . A Bivariate Generalized Linear Model GLM ) using a Gamma distribution and a Log link function was used to investigate associations between the best neutralization score and each individual clinical parameter . A value of 0 . 0001 was added to the best neutralization score to avoid zero values . Spearman’s rank correlation , Mann-Whitney U- or t-test , Fisher’s Exact test , and Kruskal-Wallis test or ANOVA were used to examine the associations between factors studied . P-values from Bivariate analyses were adjusted by False Discovery Rate ( FDR ) . Factors significantly associated with best neutralization score in the Bivariate analyses , with q-values < 0 . 1 were selected for further Multivariable modeling . Kaplan-Meier survival analyses with Log rank test were performed to look at the relationship between time post-infection when certain level of neutralization was achieved and different levels of clinical parameters . Two different versions of this analysis were performed either including all the participants described in this study ( n = 439 ) or including only the individuals having reached the level of neutralization assessed at some point during the study . P-values less than 0 . 05 were considered statistically significant . Plasma collected from the Protocol C cohort eligible participants were heat-inactivated at 56° C for 45min prior to use in neutralization assays . Briefly , WT and mutant pseudoviruses were generated by co-transfection of 293T cells with an Env-expressing plasmid and an Env-deficient genomic backbone plasmid ( pSG3ΔEnv ) , as described previously [86] . Pseudoviruses were harvested 72h post transfection for use in neutralization assays . Neutralizing activity was assessed in absence of DEAE-dextran using single-round replication in TZM-bl target cells by measuring luciferase activity after 72h . Pseudoviruses incorporating single amino acid mutations were generated by Quickchange mutagenesis ( Stratagene ) . Kifunensine-treated pseudoviruses were produced by treating 293T cells with 25 μM kifunensine ( α-mannosidases inhibitor , preventing the trimming of Man8/9 glycan to Man5 ) on the day of transfection [87] . Chimeric HIV-2 clones containing the partial or full length MPER of HIV-1 were derived from the parental HIV-2 7312A clone in which the HIV-2 Env MPER sequence ( QKLN- SWDVFGNWFDLASWVKYIQ ) was replaced by the complete ( HIV-2 C1 ) , the N-terminal segment ( HIV-2 C3 , 2F5 epitope ) or the C-terminal segment ( HIV-2 C4 , 4E10/10E8 epitopes ) of HIV-1 YU2 MPER sequence LALDKWASLWNWFDITKWLWYIK , as described [14] . To determine ID50 values , serial dilutions of plasma were incubated with virus and the dose-response curves were fitted using nonlinear regression . For competition assays , plasma dilutions were pre-incubated 30 minutes at room temperature with 25μg/mL of gp120 cores ( RCS3/KO-RSC3 , TriMut/KO-TriMut ) or 10μg/mL of MPER peptide . Both KO cores bore the combined D368R+E370A+D474A mutations essentially affecting binding of CD4bs-specific monoclonal bnAbs . An effect of a particular mutation , competitor or virus treatment on was called positive when it resulted in a > 2-fold or >20% decrease in neutralization of ID50 compared to WT virus , KO-competitor or untreated virus . Neutralization score on the 37-virus panel ( 37v-panel ) was calculated using the same formula used for the 6-virus panel ( 6v-panel ) as detailed above . HIV-1 MN ENVgp41 ( E . Coli , #12027 ) and HIV-1 IIIB GAGp24 ( Baculo , #12028 ) recombinant protein were obtained through the NIH AIDS Reagent Program , Division of AIDS , NIAID ) ( DAIDS Immunodiagnostics , Inc ) . All gp120 monomers and core gp120 proteins were expressed by transfecting 293F cells in plasma-free medium ( OptiMEM , Invitrogen , Carlsbad , CA ) . In brief , cell culture supernatants of 293-transfected cells were harvested 4 days post-transfection , cleared , filtered and two protease inhibitor tablets ( Roche ) per liter of supernatant were added to limit proteolysis . Gp120 proteins were purified on Galanthus nivalis lectin-bound agarose columns ( Vector Laboratories ) . The columns were then washed sequentially with 10 column volumes of phosphate-buffered saline ( PBS ) ( pH 7 . 4 ) containing 0 . 5 M NaCl , followed by 10 column volumes of PBS ( pH 7 . 4 ) . The lectin-bound glycoproteins were eluted with a total of 10 column volumes of elution buffer ( PBS buffer [pH 7 . 4] with 0 . 5 M methyl-D-mannopyranoside and 10mM imidazole ) . The mannoside-eluted glycoproteins were pooled , dialyzed against phosphate-buffered saline ( PBS ) pH 7 . 4 before being size excluded on a Superose 6 . Fractions containing monomers were concentrated with Amicon Ultra 30 , 000 MWCO centrifugal filter devices ( Millipore , Bedford , MA ) . Finally , the purified proteins were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis and ELISA analysis , and protein purity was verified . Plasma adsorptions with gp120-coupled beads were performed using tosyl-activated magnetic beads , as described previously [88] . Bead coupling was performed at a ratio of 1mg gp120 of a single strain per 25mg of beads . Three to four rounds of adsorption were performed to ensure complete removal of antigen-specific antibodies as verified by ELISA . Gp120 from multiple strains were used individually in independent experiments ( S3 Table in S1 Text ) and chosen based on an ENV pseudotyped virus neutralization by a given plasma sample . For plasma adsorptions performed in the presence of b6 , gp120- coupled beads were pre-incubated with 500 μg/ml IgG b6 for 30min at room temperature before adding plasma . The mAb b6 was procured by the IAVI Neutralizing Antibody Consortium . An effect of gp120 absorption was called positive when it resulted in a >2 fold decrease in neutralization of ID50 compared to the untreated plasma . Half-area 96-well ELISA plates were coated overnight at 4C with 50 μL PBS containing 50 to 250 ng of RCS3 , KO-RSC3 , ENV-gp120 , ENV-gp41 , GAG-p24 or anti-human IgG Fc per well . The wells were washed four times with PBS containing 0 . 05% Tween 20 and blocked with 3% BSA at room temperature for 1 h . Serial dilutions of plasma were then added to the wells , and the plates were incubated at room temperature for 1 hour . After washing four times , goat anti-human IgG F ( ab’ ) 2 conjugated to alkaline phosphatase ( Pierce ) , diluted 1:1000 in PBS containing 1% BSA and 0 . 025% Tween 20 , was added to the wells . The plates were incubated at room temperature for 1 h , washed four times , and developed by adding alkaline phosphatase substrate ( Sigma ) diluted in alkaline phosphatase staining buffer ( pH 9 . 8 ) , according to the manufacturer’s instructions . For avidity assessment , washes were performed in presence of 1 . 5M NaSCN or 3M NaSCN . Optical density at 405 nm was read on a microplate reader ( Molecular Devices ) . Endpoint titers of the plasma antibodies were defined as the last reciprocal plasma dilution at which the background-corrected OD signal was greater than or equal to 0 . 1 and EC50 values were calculated using Prism6 ( GraphPad ) .
|
Understanding how HIV-1-broadly neutralizing antibodies ( bnAbs ) develop during natural infection is essential to the design of an efficient HIV vaccine . We studied kinetics and correlates of neutralization breadth in a large sub-Saharan African longitudinal cohort of 439 participants with primary HIV-1 infection . Broadly nAb responses developed in 15% of individuals , on average three years after infection . Broad neutralization was associated with high viral load , low CD4+ T cell counts , virus subtype C infection and HLA*A3 ( - ) genotype . A correlation with high overall plasma IgG levels and anti-Env binding titers was also found . Specificity mapping of the bnAb responses showed that glycan-dependent epitopes , in particular the N332 region , were most commonly targeted , in contrast to other bnAb epitopes , suggesting that the HIV Env N332-glycan epitope region may be a favorable target for vaccine design .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2016
|
Broadly Neutralizing Antibody Responses in a Large Longitudinal Sub-Saharan HIV Primary Infection Cohort
|
Plasmodium vivax remains an important cause of malaria in South America and the Asia-Pacific . Naturally acquired antibody responses against multiple P . vivax proteins have been described in numerous countries , however , direct comparison of these responses has been difficult with different methodologies employed . We measured antibody responses against 307 P . vivax proteins at the time of P . vivax infection , and at 2–3 later time-points in three countries . We observed that seropositivity rates at the time of infection were highest in Thailand , followed by Brazil then PNG , reflecting the level of antigenic input . The majority of sero-reactive antigens in all sites induced short-lived antibody responses with estimated half-lives of less than 6 months , although there was a trend towards longer-lived responses in PNG children . Despite these differences , IgG seropositivity rates , magnitude and longevity were highly and significantly rank-correlated between the different regions , suggesting such features are reflective of the individual protein .
Plasmodium vivax is the most geographically widespread of the Plasmodium spp . causing malaria in humans , and accounts for the majority of cases outside Africa [1] . The leaders of Central American and East Asian countries have declared their intention for elimination of malaria within their regions by 2025 and 2030 [2 , 3] , respectively . Interrupting P . vivax transmission may require the development of an effective vaccine and improved surveillance systems ( due to the high proportion of sub-microscopic P . vivax infections [4] and the presence of currently undetectable hypnozoites ) [5] . Both objectives require a better understanding of naturally induced immune responses , in particular antibody responses , following infection . Protective immunity against clinical episodes of P . vivax malaria is acquired more rapidly than immunity against P . falciparum [6 , 7] . As a consequence , the burden of clinical P . vivax disease falls predominantly on very young children , whilst infected adults are often asymptomatic [8–10] . Immunity against clinical symptoms is thought to be dependent on the development of P . vivax-specific antibody responses [11 , 12] . The prevalence of P . vivax-specific antibody responses within endemic populations increases with age [13] , and antibody levels are generally higher in individuals with current P . vivax infections [14] . The parasite expresses more than 5000 proteins throughout its lifecycle [15] , and through the use of large-scale screening platforms such as protein microarrays [16] , an increasing number of these proteins have been assessed for their immunogenicity in naturally exposed populations . These studies , assessing antibody responses to over 1900 P . vivax proteins , have indicated that up to 50% of such proteins are sero-reactive in individuals from malaria endemic areas [11] . Furthermore , they have demonstrated that asymptomatic individuals have a greater breadth of response [11] , that reactive proteins are more likely to be encoded by genes with high single nucleotide polymorphism diversity ( potentially signifying positive immune selection ) [17] , and that the majority of highly reactive proteins have predicted transmembrane domains and signal peptides ( SP ) ( signifying secreted or membrane-bound proteins ) [18] . Whilst naturally acquired antibody responses have been assessed against multiple P . vivax proteins in individuals from numerous geographic regions , it has been difficult to directly compare these responses due to the different protein expression and antibody measurement systems used . For example , whilst Finney and colleagues reported a greater breadth of antibody response in asymptomatic individuals in PNG compared to febrile patients [11] , as noted above , the opposite was later reported in Thailand [13] and no difference was identified in India [19] . This could be due to the different geographic region and transmission intensity of the study sites , or due to the use of different P . vivax protein microarrays encompassing different sets of proteins . In this study we performed a direct comparison of naturally acquired antibody responses in three different geographic regions: Brazil ( very low transmission ) , Thailand ( low transmission ) and PNG ( moderate transmission ) . In these settings , we measured antibodies at 2–3 time-points following P . vivax infection ( in the absence of detectable recurrent infections ) , to a panel of 307 P . vivax proteins . All samples were measured using the same platform ( AlphaScreen assay ) , with the same data analysis pipeline applied . We found that antibody positivity , magnitude and longevity were highly and significantly correlated between the three study sites; however , in general , antibody longevity was better maintained in PNG where individuals have had greater past exposure to the parasite . Our study provides important new insights into the generation and maintenance of P . vivax-specific antibody responses and how these factors differ in regions of varying transmission intensity .
The relevant local ethics committees approved all field studies , and the Human Research Ethics Committee at WEHI approved samples for use in Melbourne ( #14/02 ) . The PNG study received ethical clearance from the PNG Institute of Medical Research Institutional Review Board ( 0908 ) , the PNG Medical Advisory Committee ( 09 . 11 ) , and the Ethics Committee of Basel ( 237/11 ) . This study was also retrospectively registered at ClinicalTrials . gov ( NCT02143934 ) . The Thai study was approved by the Ethics Committee of the Faculty of Tropical Medicine , Mahidol University , Thailand ( MUTM 2014-025-01 and 02 ) . The Brazilian study was approved by the Ethics Review Board of the Fundação de Medicina Tropical Dr . Heitor Vieira Dourado ( FMT-HVD ) ( 957 . 875/2014 ) . All patients gave informed written consent or assent . For children , informed written consent was provided by their parents . Three different study sites were used: Maprik District , East Sepik Province , Papua New Guinea ( PNG ) ; Tha Song Yang District , Tak Province , Thailand; and Manaus , Brazil . A sub-set of enrolled volunteers , where the probability of reinfection during the study period was low ( in order to accurately determine antibody longevity ) , was included in the present study as described below . The longitudinal study in PNG was conducted from August 2009 to May 2010 , as described [20] . In brief , 524 children aged 5–10 years were enrolled and block randomised to receive 20 days of directly observed therapy ( DOT ) of chloroquine , arthemeter-lumefrantrine with either a placebo or primaquine . Children were subsequently monitored every 2 weeks for 8 months to detect any signs or symptoms of infection , with finger-prick blood samples taken every 2 weeks for 12 weeks , then every 4 weeks for the remaining period . Blood samples were analysed by light microscopy and qPCR for the presence of blood-stage parasites and plasma stored for antibody measurements . A sub-set of 31 children with P . vivax infections at enrolment and no evidence of reinfection during follow-up , with not more than one missed sample , were selected for inclusion in the present study ( all 31 received primaquine treatment ) . Plasma samples from 0 , 3 and 5 months were used . The longitudinal study in Thailand was conducted from April 2014 to September 2015 , as described [21] . Briefly , 57 symptomatic P . vivax patients were enrolled from either the Tha Song Yang malaria clinic or hospital . Patients were treated with chloroquine ( 25 mg base/kg body weight , administered over 3 days ) and primaquine ( 15 mg daily , for 14 days ) under DOT . Volunteers were followed for 9-months after enrolment , with finger-prick blood samples collected at enrolment and week 1 , then every 2 weeks for 6 months , then every month until the end of the study . All blood samples were analysed by both light microscopy and qPCR for the presence of blood-stage parasites as per the PNG cohort , and plasma separated and stored . A sub-set of volunteers , n = 32 , were selected for use in the current study . These volunteers had no detectable recurrent infections during 9-months follow-up , and were the first to complete follow-up . Plasma samples from 0 , 3 , 6 and 9 months were used . The longitudinal study in Brazil followed the same format as in Thailand . The study was conducted from May 2014 to May 2015 . 91 malaria patients at Fundação de Medicina Tropical Doutor Heitor Vieira Dourado in Manaus aged between 7 and 70 years were enrolled . Individuals with G6PD deficiency or chronic diseases were not enrolled . Patients were treated according to the guidelines of the Brazilian Ministry of Health ( 3 days chloroquine , 7 days primaquine ) . Follow-up intervals with finger-prick blood sample collection were as in the Thai study . A sub-set of volunteers , n = 33 , who had no detectable recurrent infections during 9-months follow-up , were selected for use in the antigen discovery project . Plasma samples from 0 , 3 , 6 and 9 months were used . This study utilized a panel of 307 P . vivax protein fragments produced using a wheat germ cell-free protein expression ( WGCF ) system . This panel of protein fragments includes well-known P . vivax proteins such as potential vaccine candidates ( i . e . merozoite surface protein 1 ( MSP1 ) , apical membrane antigen 1 ( AMA1 ) , circumsporozoite protein ( CSP ) ) , orthologs of immunogenic P . falciparum proteins and proteins with a predicted SP and/or 1–3 transmembrane domains ( TM ) , and expands upon our previously published panel [22] . Fig 1 demonstrates key features of these P . vivax proteins ( annotation and expression-stage ) . Approximately 70% contained a predicted SP , 50% at least one TM and 10% a GPI-anchor . All protein-specific information was obtained from the Plasmodium Genomics Resource ( PlasmoDB: http://plasmodb . org/plasmo/ ) release 25 . The recombinant proteins were expressed without codon optimization using the WGCF system as previously described [23] with slight modifications . Briefly , the genes were amplified by PCR and cloned into the pEU_E01 expression vector with N-terminal His-bls tag ( CellFree Sciences , Matsuyama , Japan ) . P . vivax genes were obtained either from parent clones [22] , using SAL-1 cDNA , or commercially synthesized from Genscript ( Japan ) . DNA template for WGCF was prepared from the plasmids by PCR with a forward primer ( spu; 5-GCGTAGCATTTAGGTGACACT-3 ) , and a reverse primer ( pbsa1143; 5-GCTCACATGTTCTTTCCTGC-3 ) . WGCF synthesis of the P . vivax protein library was based on the previously described bilayer diffusion system [24] . For biotinylation of proteins , 500 nM D-biotin ( Nacalai Tesque , Kyoto , Japan ) was added to both the translation and substrate layers . Crude WGCF expressed BirA ( 1 μl ) was added to the translation layer . In vitro transcription and cell-free protein synthesis for the P . vivax protein library were carried out using the GenDecoder 1000 robotic synthesizer ( CellFree Sciences ) [25 , 26] . Based on previous studies , the WGCF system can synthesize biotinylated plasmodial protein ranging from 0 . 3 μg/mL to 26 . 5 μg/mL [27 , 28] . Expression of the proteins was confirmed by western blotting using HRP-conjugated streptavidin . Proteins were preferably expressed as full-length proteins , to ensure that any possible antibody recognition site was covered; however , for very large proteins , multiple fragments were expressed that together cover the entire protein . The panel included 263 unique proteins . The AlphaScreen assay was performed following the manufacturer’s instructions ( PerkinElmer Life and Analytical Sciences , Boston , MA ) as previously reported [27 , 28] , with slight modifications . The protocol was automated by use of the JANUS Automated Workstation ( PerkinElmer Life and Analytical Science , Boston , MA ) . Reactions were carried out in 25 μl of reaction volume per well in 384-well OptiPlate microtiter plates ( PerkinElmer ) . First , 0 . 1 μl of the translation mixture containing a recombinant P . vivax biotinylated protein was diluted 50-fold ( 5 μl ) , mixed with 10 μl of 4000-fold diluted plasma in reaction buffer ( 100 mM Tris-HCL [pH 8 . 0] , 0 . 01% [v/v] Tween-20 and 0 . 1% [w/v] bovine serum albumin ) , and incubated for 30 min at 26°C to form an antigen-antibody complex . Subsequently , a 10 μl suspension of streptavidin-coated donor-beads and acceptor-beads ( PerkinElmer ) conjugated with protein G ( Thermo Scientific , Waltham , MA ) in the reaction buffer was added to a final concentration of 12 μg/ml of both beads . The translation mixture was diluted 250 times , thus final concentration of biotinylated proteins ranged between approximately 1 . 2 ng/mL and 106 ng/mL . The mixture was incubated at 26°C for one hour in the dark to allow the donor and acceptor-beads to optimally bind to biotin and human IgG , respectively . Upon illumination of this complex , a luminescence signal at 620 nm was detected by the EnVision plate reader ( PerkinElmer ) and the result was expressed as AlphaScreen counts . Each assay plate contained a standard curve of total biotinylated rabbit IgG . This enabled standardisation between plates using a 5-paramater logistic standard curve . All samples from Thailand and Brazil were run in triplicate , whilst those from PNG were run only once . Reading the plates was conducted in a randomized manner to avoid biases . The immunogenicity of each protein was assessed in two ways . Firstly by calculating the geometric mean titre ( GMT ) , and secondly through seropositivity–the proportion of samples above a given threshold . The seropositivity cut-off was set at the limit of detection of the assay , taken as half the lowest non-negative value from each of the three sites [24] . Proteins were defined as reactive if more than 10% of the volunteers had levels above the seropositivity cut-off at the baseline measurement ( time of P . vivax infection ) . All data manipulation and statistical analyses were performed in R version 3 . 2 . 3 [29] . Antibody longevity was estimated using data from three or four time-points ( 0 , 3 and 5 months in PNG; 0 , 3 , 6 and 9 months in Thailand and Brazil ) . A linear mixed effects model was used to estimate the antibody half-life; using the lme4 [30] package in R . We assume that antibody titres decay exponentially , equivalent to a linear reduction in log antibody titre over time . Denote Aijk to be the antibody titre to antigen j in participant i at time tk which can therefore be described by the following linear model: log ( Aijk ) ∼ ( log ( αj0 ) +log ( αij ) ) + ( rj0+rij ) tk+εjlog ( αij ) ∼N ( 0 , σA , j ) log ( rij ) ∼N ( 0 , σr , j ) εj∼N ( 0 , σm , j ) where αj0 is the geometric mean titre ( GMT ) at the time of infection; log ( αij ) is a random effect accounting for the difference between participant i's initial antibody titre and the population-level GMT; rj0 is the average rate of decay of antigen j in the population; rij is a random effect for the difference between the decay rate of individual j with the population-level average; and εj is a Normally distributed error term . The model was only fitted to individuals who were seropositive at baseline . This model also generated an estimate of the total variation in the data ( arising from initial antibody level measured , the rate of antibody decay and the measurement error ) . Correlations between the three cohorts were performed using the Spearman’s rank correlation coefficient .
Plasma samples were utilised from three studies in different geographic regions , with slightly different study designs . PNG included only children aged 5–10 years , whilst Brazil and Thailand included volunteers of all ages . Furthermore , all Thai and Brazilian volunteers were enrolled following symptomatic infection , whilst all included PNG children had asymptomatic infections . This resulted in significantly higher antigenic inputs in the Thai and Brazilian volunteers compared to PNG children ( p<0 . 0001 , Kruskal-Wallis test with Dunn’s multiple comparisons test ) ( Table 1 ) . Reactivity was defined as more than 10% of the volunteers above the seropositivity cut-off , with the results at baseline ( time of P . vivax infection ) used for this analysis . Of the 307 P . vivax proteins assessed , 302 were considered reactive in Thailand , 273 in Brazil and 236 in PNG , with the population level breadth therefore highest in Thailand . All proteins considered reactive in Brazil and PNG were also considered reactive in Thailand . There were 3 proteins considered reactive in PNG and not Brazil ( all with low levels of seropositivity , less than 30% ) , 29 proteins reactive in Thailand and not Brazil ( seropositivity less than 35% ) , 40 proteins reactive in Brazil and not PNG ( less than 50% seropositivity ) , and 65 proteins reactive in Thailand and not PNG . The majority of the reactive proteins identified in Thailand that were not observed in PNG had less than 50% seropositivity at the time of infection , however there were 6 exceptions: PVX_082690 ( MSP7 ) , PVX_092995 ( Pv-fam-a ) , PVX_079980 ( hypothetical protein ) , PVX_110945 ( hypothetical protein ) and PVX_099900 ( unspecified product ) . Overall , there was a strong correlation between seropositivity rates at baseline for the overlapping reactive antigens between the three cohorts ( Thailand v Brazil , r = 0 . 95; Thailand v PNG , r = 0 . 89; Brazil v PNG , r = 0 . 88; all p<0 . 0001 ) ( Fig 2 ) . There were two outliers when comparing Thailand and PNG: MSP1-19 ( PVX_099980 , fragment covering 4918–5187 base pairs ) and MSP5 ( PVX_003770 ) ( Fig 2B ) . Both these proteins induced high levels of seropositivity at the time of P . vivax infection in Thailand ( 92% and 94% , respectively ) , but low levels in PNG ( 16% and 32% , respectively ) . We also measured IgG responses against another construct of MSP1-19 , base pairs 4863–5187 , and responses were similarly higher in Thailand ( 95% ) compared to PNG ( 65% ) , although the difference was not as large . There were also two proteins where seropositivity levels were high in PNG ( more than 80% ) , but relatively low in Brazil ( less than 30% ) : PVX_100670 ( aspartyl protease , putative ) and PVX_097930 ( a hypothetical protein ) ( Fig 2C ) . On an individual level , the breadth of the antibody response ( that is , the number of antigens an individual was seropositive for ) was slightly higher in Thailand and PNG compared to Brazil , at the time of P . vivax infection ( Fig 3 ) . The median number of antigens an individual was seropositive for was 202 in Thailand , 173 in PNG and 151 in Brazil . There was a statistically significant difference between the three sites ( p = 0 . 0004 , Kruskal-Wallis test ) . In the AlphaScreen assay , the amount of protein used is not consistent between different proteins . Therefore , the absolute magnitude of the antibody response measured cannot be compared across proteins . However , the magnitude can be compared between individuals or populations for the same proteins . The geometric mean antibody level was determined for each protein for each study site , at the time of P . vivax infection . This value was then correlated between different sites; for the reactive antigens in common at baseline , the geometric mean was significantly correlated between all three cohorts ( Thailand v Brazil , r = 0 . 96; Thailand v PNG , r = 0 . 90; Brazil v PNG , r = 0 . 90; all p<0 . 0001 ) ( Fig 4 ) . Overall , slightly higher mean antibody levels were measured in PNG compared to both Brazil and Thailand . Antibody responses to the 307 P . vivax proteins were measured at multiple time-points; this made it possible to estimate the antibody longevity for the reactive proteins in each cohort ( Fig 5 ) . Antibody responses with an estimated half-life of less than 6 months ( 180 days ) were considered short-lived , whereas those more than 6 months were considered long-lived . The majority of reactive antigens in all cohorts were considered short-lived , with the greatest proportion in Brazil ( 88% ) , followed by Thailand ( 79% ) , then PNG ( 59% ) . The range of estimated antibody half-lives in each cohort were as follows: 12–445 days in Brazil , 18–713 days in Thailand and 10–70 , 000 days in PNG . With 5 to 9 months of longitudinal data , there is limited statistical power to estimate very long half-lives , and hence a half-life of greater than 1 , 000 days is usually indistinguishable from no change in antibody titres over time . Overall , there was a strong correlation between estimated antibody longevity estimates for reactive proteins , when proteins with negative antibody half-life estimates were excluded ( Thailand v Brazil , r = 0 . 92 , n = 273 proteins; Thailand v PNG , r = 0 . 87 , n = 219 proteins; Brazil v PNG , r = 0 . 86 , n = 216 proteins; all p<0 . 0001 ) ( Fig 6 ) . Concerning a number of well-recognised P . vivax antigens mentioned above , IgG responses to MSP1-19 and MSP5 were estimated to be exceptionally short-lived in all three cohorts , with half-lives of 16–41 days and 20–57 days , respectively . Responses to MSP1-42 were similarly short-lived in Brazil and Thailand ( half-lives of 58–97 days ) but were long-lived in PNG ( estimated IgG half-life of almost 3 years ) .
In this study we have compared naturally acquired IgG antibody responses to over 300 P . vivax protein constructs in three geographically different locations with varying levels of P . vivax transmission . Manaus , Brazil and Tak , Thailand are both low-transmission regions ( with approximately 0–5% and 1–11% of individuals infected in recent surveys , respectively [13 , 21 , 31] ) . The participants from PNG were part of a larger study where a PCR prevalence of 47% was measured at baseline [20] . At the time of P . vivax infection we observed that a high proportion of the 307 P . vivax proteins were reactive ( 77–98% ) . This is higher than previous reports using large-scale screening platforms , where sero-reactivity has ranged from 14–50% in P . vivax patients [11 , 17 , 22] , likely due both to our down-selection of proteins predicted to be immunogenic and our definition of seropositivity ( based on the limit of detection of the assay , rather than unexposed controls ) . We observed higher numbers of sero-reactive proteins in Thailand and Brazil compared to PNG; the volunteers in Thailand and Brazil who had symptomatic infections also had significantly higher parasite densities than the asymptomatic children in PNG , signifying that antigenic input has a significant effect on the breadth of the antibody response . Similarly , on an individual level , Thai volunteers showed a trend towards a greater breadth of response compared to Brazil and PNG . This pattern of greater breadth of response in individuals with higher parasite densities [11] , and following symptomatic versus asymptomatic infections [13] , has been reported previously . As asymptomatic infections are becoming increasingly recognised in low-transmission regions such as Thailand [13] , it would be valuable to compare the breadth of response at the time of asymptomatic P . vivax infection in Thailand , or a similar low-transmission region , to that in PNG . The additional proteins identified as sero-reactive in Thailand and Brazil compared to PNG had only low levels of seropositivity amongst the 32–33 individuals at time of infection ( i . e . between 10–50% ) , except for a small number of proteins in the Thai individuals including one Pv-fam-a protein ( PVX_092995 ) and MSP7 ( PVX_082690 ) . In general , there was good correlation in levels of seropositivity to all antigens considered reactive in multiple cohorts . Two key exceptions were observed: high levels of seropositivity in Thai individuals to MSP1-19 ( fragment covering base pairs 4918–5187 ) and MSP5 ( 92% and 94% , respectively ) , with only low-levels of seropositivity observed in the PNG children ( 16% and 32% , respectively ) . MSP1 and MSP5 are both blood-stage antigens considered potential vaccine candidates , based primarily on their location on the merozoite surface and pre-clinical studies of the P . falciparum orthologs [32 , 33] . MSP1 first exists as a 42 kDa protein attached to the surface of merozoites via a GPI anchor; following invasion of RBCs the protein is cleaved into two products , of 33 kDa and 19 kDa , with the 19 kDa fragment remaining on the merozoite surface [34] . We observed 100% seropositivity in all cohorts to the MSP1-42 fragment . Not a lot is known about IgG responses to MSP5 [33] , but responses to MSP1-19 are frequently observed in endemic regions [35] , and are known to be short-lived [36] . We observed short-lived responses to both MSP1-19 and MSP5 in all three study sites . In agreement with our current finding , higher IgG titres to MSP1-19 have been observed in sporadically rather than chronically exposed volunteers in an endemic region of Brazil [37]; together this suggests that in chronically exposed individuals , such as PNG children , MSP1-19 and MSP5 may be able to escape recognition of the humoral immune system or that such low-density infections are unable to adequately boost the IgG response . However , it is important to note that we observed moderate levels of seropositivity in PNG children to a second construct of MSP1-19 ( covering base pairs 4863–5187 ) , and another study has also found prevalent IgG responses to MSP1-19 in younger PNG children [38] . Our current finding therefore warrants further investigation with a larger sample size to clarify these results . The majority of previous studies that have assessed IgG antibody responses to a large panel of P . vivax proteins have been cross-sectional in design; whilst this provides useful information on antibody acquisition and immunogenicity , it fails to provide detailed information regarding antibody longevity . In this study we assessed antibody responses at 2–3 further time points following P . vivax infection , in the absence of detectable recurrent infections , to investigate antibody decay . We observed that in our two low-transmission study sites , Thailand and Brazil , most antigens had relatively short-lived IgG responses , with estimated antibody half-lives of less than 6 months . It is important to note that this was not observed for all proteins , with some exhibiting exceptionally long-lived IgG profiles with half-lives up to 700 days , and we are currently investigating the mechanisms behind such differential responses . Conversely , in the PNG children who had a high level of past , lifetime exposure [6] , 41% of proteins induced long-lived IgG responses with estimated half-lives of more than 6 months , and there was a trend towards longer-lived responses to many proteins in these children compared to individuals from Thailand and Brazil . This suggests that whilst the breadth of response may be dependent upon the level of antigenic input , IgG responses can be long-lived to many proteins if there has been sufficient past levels of exposure within a region . Together , this provides further evidence that IgG responses to P . vivax proteins can be long-lived , dependent on the transmission setting and specific protein , which is in contrast to most reported studies of IgG longevity against P . falciparum proteins , where responses are generally short-lived in the absence of ongoing exposure [35 , 39] . Despite these differences between study sites in terms of the breadth and longevity of the IgG response , there was a highly significant rank-correlation between sites in terms of the proportion of individuals seropositive to each antigen , the magnitude of the IgG response to each antigen , and the estimated antibody half-life to each antigen . This suggests that despite the differences due to parasite density and past exposure levels , immunogenicity and longevity are ultimately characteristics specific to individual proteins . This feature is therefore highly promising for the use of IgG antibody responses as markers of exposure in multiple geographic regions , if the transmission level and current infection status is taken into consideration . An important consideration for future research is the impact of age on the IgG responses observed; our current work has attributed differences in the PNG cohort to the higher level of transmission within this region and lower antigenic input due to asymptomatic infection , however the age of the volunteers could also have an effect . In addition , longitudinal studies of individuals from low-transmission regions following asymptomatic P . vivax infections would also be valuable to further support our hypotheses . Furthermore , the potential for cross-reactivity of these P . vivax proteins with P . falciparum will become increasingly important in regions where these species are co-endemic . The potential for cross-reactivity will be tested by measuring responses in cohorts which are known to have P . falciparum but no P . vivax [40] . In summary , our study has utilised a panel of more than 300 P . vivax proteins , the majority of which were immunogenic , to provide evidence for a number of key features relating to IgG acquisition and longevity: that the individual and population-level breadth of IgG seropositivity is a function of antigenic input , whilst longevity is a function of the level of transmission ( and hence lifetime exposure to the parasite ) .
|
In the pursuit of eliminating all species of malaria , Plasmodium vivax presents one of the most substantial challenges , particularly in countries in Asia , the Western-Pacific and South America . This is primarily due to the ability of P . vivax to cause relapse infections months to years after the initial infectious bite . In areas with low levels of malaria transmission , serology has become an increasingly useful tool for surveillance , as anti-Plasmodium antibodies can be detected in individuals long after blood-stage parasites have cleared . In this study , we provide a detailed characterisation of the antibody response generated following P . vivax infection by measuring antibodies to over 300 P . vivax antigens in three different populations in Thailand , Brazil and Papua New Guinea . The individuals in these populations were followed for up to nine months allowing us to estimate the rate at which antibodies decay over time . This improved understanding of the magnitude and dynamics of the antibody response , validated in multiple populations , will contribute to the development of serological surveillance tools needed for enhanced control and elimination of P . vivax .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"parasite",
"groups",
"medicine",
"and",
"health",
"sciences",
"immune",
"physiology",
"plasmodium",
"immunology",
"geographical",
"locations",
"tropical",
"diseases",
"parasitic",
"diseases",
"parasitology",
"apicomplexa",
"regional",
"geography",
"antibodies",
"antibody",
"response",
"immune",
"system",
"proteins",
"geography",
"south",
"america",
"proteins",
"antigens",
"brazil",
"immune",
"response",
"people",
"and",
"places",
"biochemistry",
"asia",
"physiology",
"earth",
"sciences",
"biology",
"and",
"life",
"sciences",
"malaria",
"thailand"
] |
2017
|
Naturally acquired antibody responses to more than 300 Plasmodium vivax proteins in three geographic regions
|
CD1d-mediated presentation of glycolipid antigens to T cells is capable of initiating powerful immune responses that can have a beneficial impact on many diseases . Molecular analyses have recently detailed the lipid antigen recognition strategies utilized by the invariant Vα24-Jα18 TCR rearrangements of iNKT cells , which comprise a subset of the human CD1d-restricted T cell population . In contrast , little is known about how lipid antigens are recognized by functionally distinct CD1d-restricted T cells bearing different TCRα chain rearrangements . Here we present crystallographic and biophysical analyses of α-galactosylceramide ( α-GalCer ) recognition by a human CD1d-restricted TCR that utilizes a Vα3 . 1-Jα18 rearrangement and displays a more restricted specificity for α-linked glycolipids than that of iNKT TCRs . Despite having sequence divergence in the CDR1α and CDR2α loops , this TCR employs a convergent recognition strategy to engage CD1d/αGalCer , with a binding affinity ( ∼2 µM ) almost identical to that of an iNKT TCR used in this study . The CDR3α loop , similar in sequence to iNKT-TCRs , engages CD1d/αGalCer in a similar position as that seen with iNKT-TCRs , however fewer actual contacts are made . Instead , the CDR1α loop contributes important contacts to CD1d/αGalCer , with an emphasis on the 4′OH of the galactose headgroup . This is consistent with the inability of Vα24− T cells to respond to α-glucosylceramide , which differs from αGalCer in the position of the 4′OH . These data illustrate how fine specificity for a lipid containing α-linked galactose is achieved by a TCR structurally distinct from that of iNKT cells .
Natural killer T ( NKT ) cells are a highly conserved lineage of T lymphocytes found in both human and mice that are involved in the modulation of the immune response in autoimmunity , infection , and tumor development [1] . Unlike conventional CD4+ and CD8+ αβ T cells that recognize peptides presented by MHC molecules , NKT cells are reactive to a broad range of self and foreign lipids displayed by the MHC class I–like molecule CD1d [2] , [3] . This reactivity is initiated by the recognition of the CD1d-lipid complex via the NKT T cell receptor ( NKT-TCR ) followed by Th1 and/or Th2 biased cytokine secretion that can regulate the activity of other immune cells such as conventional αβ T cells , B cells , and Natural Killer ( NK ) cells [4] . The most extensively studied NKT cells in humans and mice are invariant ( iNKT ) or type I NKT cells that express TCRs composed of a highly conserved α chain encoded by a Vα24-Jα18 rearranged gene segment in humans and Vα14-Jα18 in mice . This invariant α chain is covalently paired with a β chain in which the variable region is encoded in humans by the Vβ11 gene and can be Vβ8 , Vβ7 , or Vβ2 in mice [1] . NKT cells expressing these TCRs have a pre-activated phenotype that is due to the expression of the transcription factor pro-myelocytic leukemia zinc finger ( PLZF ) [5] , [6] and are also characterized by high reactivity towards the potent stimulatory lipid antigen α-galactosylceramide ( αGalCer ) [7] . In both humans and mice there are additional classes of T cells that respond to CD1d , one that expresses diverse TCRs but do not respond to αGalCer; these are generally called Type II or non-invariant NKT cells [8] . These NKT cells are typically reactive to lipid antigens such as sulfatide and use an entirely different molecular strategy for recognizing the CD1d/lipid complex [9] , [10] . A third group of T cells exist that do respond to CD1d presenting αGalCer and also express TCRs different from that of the iNKT-TCR . In mice these NKT cells express a TCR comprised of a Vα10-Jα50/Vβ8 pair [11] . These cells are called Vα10 NKT cells and show a preference for α-glucosylceramide ( αGlcCer ) over αGalCer; indeed , Vα10 NKT cells can produce a several magnitudes greater cytokine response relative to iNKT cells when stimulated by the related α-glucuronosyldiacylglycerl ( α-GlcA-DAG ) [11] . In humans this third group of CD1d reactive T cells express TCRs with many different Vα domains joined with Jα18 , paired with the Vβ11 domain [12] , [13] . In contrast to both Type I and Type II NKT cells , these T cells do not typically express CD161 , a Natural Killer cell marker found on NKT cells [13] . They have been called Vα24− NKT cells or CD1d-restricted , Vα24− T cells due to their use of alternative Vα domains rearranged to Jα18 , paired with the Vβ11 domain in their TCRs . These cells are found in all individuals sampled [13] at appreciable frequency ( ∼10−5 ) [14] and express either the CD8αβ or CD4 co-receptors , can be cytotoxic , and can secrete IL-2 , IFN-γ , and IL-13 ( and in some cases IL-4 ) [13] . In contrast to human iNKT cells , they express low to intermediate levels of PLZF and have a naïve phenotype [14] . Importantly , these NKT cells have shifted lipid specificities from that of iNKT cells with an inability to recognize and respond to αGlcCer [12] . The distinctive difference in reactivity between αGalCer and αGlcCer suggests that this population of NKT cells focuses on a different repertoire of lipid antigens than those of iNKT cells . Despite the variability that exists in NKT cell populations , most of our current knowledge of NKT cell recognition of antigen derives from structural studies that have focused on self and foreign lipid antigen recognition by Type I iNKT TCRs [15] . iNKT-TCRs recognize , through their complementary determining regions ( CDR ) loops , a composite surface composed of the α-helices of CD1d and the solvent exposed head group of the CD1d-presented lipid antigens . The CDR3α loop plays a prominent , conserved role in CD1d-lipid recognition , predominantly via residues encoded by the Jα18 segment , which is found in all iNKT TCRs . There are also important contributions from the CDR1α and CDR2β loops , which explain the restricted use of specific Vα and Vβ domains ( which encode the CDR1 and CDR2 loops ) [16] , [17] . For each Vβ chain used in mouse , the docking of iNKT-TCRs on the CD1d/lipid antigen surface is remarkably conserved [18] , [19] , indeed variation of the lipid antigen is accommodated mainly through structural modifications of the lipid antigen as opposed to changes in the iNKT TCR footprint [20]–[24] . The number of human iNKT TCR complex structures are fewer yet reflect some flexibility in docking of the iNKT TCR depending on the lipid antigen [16] , [19] , [23] , [25] , yet appear to be similarly anchored via conserved positioning of the CDR3α loop . The crystal structure of a murine Vα10 NKT TCR in complex with murine CD1d-αGlcCer [11] has shed light onto the molecular mechanisms that murine non-canonical NKT TCRs use to recognize CD1d . Despite significant sequence divergence in the α chain amino acid sequence ( 40% sequence identity ) , the Vα10 NKT TCR assumes a very similar docking mode to that of the iNKT TCR on CD1d . However , unlike the iNKT TCR , all CDR loops of the Vα10 NKT TCR contribute to CD1d/αGlcCer recognition , with seemingly important contacts being contributed by the CDR2β and CDR3β loops . Thus the two Vα chains of these divergent murine NKT cell populations ( iNKT and Vα10 ) have convergently evolved a similar molecular strategy for recognizing CD1d . Recently , crystal structures of the Type II NKT TCR recognition of CD1d presenting sulfatide [9] and lysosulfatide [10] provided an interesting contrast to the conserved recognition of CD1d by the iNKT and murine Vα10 TCRs . The Type II TCRs use all six CDR loops in CD1d/ligand engagement and dock on a separate site on CD1d , concentrating on residues surrounding the A′ pocket . Thus , NKT cells have a range of docking modes used in CD1d/ligand engagement . Structural data on NKT cell recognition in humans remains limited , and information of how Vα24− T cells recognize CD1d/lipid is , to our knowledge , absent . To better understand how this functionally distinct human T cell population recognizes CD1d/lipid , we have co-crystallized a Vα24− TCR with CD1d/αGalCer and present here the structure of this complex resolved to 2 . 5 Å resolution . This structure provides an excellent model by which to understand how functionally distinct human T cells , via their TCR , can recognize CD1d with a shifted specificity from that found in the iNKT cell population .
In order to understand the molecular basis of Vα24− TCR recognition of CD1d , we expressed a soluble , heterodimeric version of the extracellular domains of the J24 . N22 TCR [12] , which uses the Vα3 . 1 ( TRAV17 ) gene segment rearranged with Jα18 complexed with Vβ11 , in insect cells . The purified TCR was co-crystalized with recombinant , soluble CD1d loaded with αGalCer; X-ray data were collected to 2 . 5 Å , and the structure was solved via molecular replacement . Data collection and refinement statistics are listed in Table 1 . One TCR/CD1d/αGalCer ternary complex was identified in the asymmetric unit . All components of this complex were well resolved in the electron-density , enabling unambiguous assignment of TCR-CD1d/lipid antigen contacts . Table 2 presents a comparison between the amino acid sequences of the α and β CDR loops of the Vα24− ( Vα3 . 1+ ) TCR studied here and an iNKT Vα24+ TCR studied previously [25] . Vα3 . 1 and Vα24 share 46% amino acid identity overall , with only 33% ( 2/6 ) identity at the CDR1α and 15% ( 1/7 ) at the CDR2α loop . However , the shared usage between these TCRs of the Jα18 segment and the canonical DRGSTLGR motif that it encodes gives high sequence identity to the CDR3α loops of these TCRs with different residues encoded only at the Vα-Jα junction , with ATY and VVS motifs in the Vα24− and Vα24+ TCRs , respectively . The Vβ11 domain is also shared between these TCRs; therefore , the CDR1β and CDR2β sequences are identical . However , the rearranged CDR3β loops differ due to differences introduced during the rearrangement process . Overall , the Vα24− TCR recognizes CD1d/αGalCer with the α and β chains oriented on CD1d in a parallel fashion unlike the typical diagonal mode of MHC-I peptide-TCR complexes and similar to that of iNKT-TCR and Vα10 NKT-TCR in complex with CD1d/αGalCer ( Figure 1A and 1B ) [11] , [16] , [19] . However , the binding angle of the Vα24− TCR in relation to the CD1d/αGalCer surface is more acute than the almost perpendicular orientation observed with the Vα24+ iNKT TCR-CD1d/αGalCer structure ( Figure 1A ) [16] , [19] . The CDRα loops adopt a similar yet slightly shifted footprint for the α-chain , yet the β-chain CDR loop positioning is counter-clockwise rotated compared with the Vα24+ TCR complexed with αGalCer [16] , [19] , which is even more extreme than rotations observed in structures of human NKT-TCRs complexed with CD1d presenting LPC or βGalCer ( Figure 1B ) [23] , [25] . The TCR-CD1d-lipid contacts mostly fall in the F′ pocket area of the CD1d molecule ( Figure 1C ) , where there are slight differences in TCR contact surface between the Vα24− and Vα24+ . The total buried surface area ( BSA ) between the Vα24− TCR and the CD1d-αGalCer complex was 747 A2 , which is slightly smaller than the previously reported interface area for the Vα24+ TCR , ∼910 A2 . This difference is more pronounced in the β-chain loops with ∼37% less contribution in the Vα24− complex ( 205 . 7 A2 versus 325 . 3 A2 for the Vα24− and Vα24+ , respectively ) . The conformation and positioning of αGalCer presented by CD1d is almost identical in both complexes with the Vα24+ and the Vα24− TCRs . The sphinosine base and acyl chain of αGalCer fall in the F′ and A′ pockets , respectively ( Figure 1D ) . The αGalCer headgroup also adopts a very similar conformation , with solvent exposed with the sugar oxygens displayed for recognition by the TCR . The conformation of the α helical side chains of CD1d were also highly conserved between the Vα24+ and Vα24− complex structures , with only a few exceptions that are noted later in the text . In all three human iNKT TCR-CD1d/lipid complexes that have been resolved to date , the CDR1α loop makes important contacts with the lipid headgroup [16] , [19] , [23] , [25] . In recognition of αGalCer and βGalCer the Oγ of Ser30 and the mainchain carbonyl oxygen of Phe29 make hydrogen-bonds ( some water-mediated ) with the 3′OH of αGalCer and βGalCer , and in the case of LPC , the Oγ Ser27 and the mainchain carbonyl oxygen of Phe29 establish hydrogen bonds with the phosphate oxygens of the phosphorylcholine headgroup . Pro28 establishes van der Waals ( VDW ) contacts with the galactose headgroup; mutagenesis of this residue has a marked effect on recognition but is likely due to global structural changes in the conformation of the TCR as this mutation also disrupted binding of a conformational-specific antibody [17] . In our structure the Vα24− CDR1α loop is slightly shifted from the Vα24+ CDR1α loop ( Figure 1B ) ; therefore , the equivalent structural positions to the Vα24+ S27P28F29S30 motif are T26S27I28N29 in Vα24− . Despite the chemical and structural differences of the CDR1α loops between these TCRs , specific side-chain-mediated hydrogen bonds are still formed in the Vα24− CDR1α loop , both with the galactose headgroup of αGalCer and through VDW contacts with CD1d's Val72 ( Figure 2A and Table 3 ) . The shifted position of Ser27 in this complex enables a hydrogen bond between its Oγ with the 6′OH of αGalCer , whereas the Nδ2 of Asn29 hydrogen bonds with the 3′OH and 4′OH of αGalCer and Asn29 also forms VDW contacts with the galactose headgroup . Therefore , alternative residues in the CDR1α loop are effectively used in recognition of αGalCer with a focus on the 4′OH of the galactose ring , with a novel contact with CD1d also noted . We have also noted residues in the CDR2α loop that make water-mediated contacts with the αGalCer galactose headgroup: Ser50 and Asn51 both establish water-mediated hydrogen bonds with the 4′OH of αGalCer ( Figure 2B ) . In the other human complexes , Phe51 of the Vα24+ CDR2α loop makes VDW contacts with both βGalCer and LPC , however hydrogen bonds have not been noted for the CDR2α loop of Vα24+ TCRs . In contrast to the sequence and contact differences at the CDR1α and CDR2α loops , the residues of the CDR3α loop in the Vα24− TCRs adopt a similar conformation to that of the Vα24+ iNKT TCRs ( Figure 2C ) . Yet despite the similarity in footprint , the Vα24− CDR3α loop establishes fewer contacts with CD1d and αGalCer than does the CDR3α loop of the iNKT TCR ( Table 3 ) ( 25 instead of 32 , respectively , for CD1d and eight instead of 19 , respectively , for αGalCer ) . There are fewer hydrogen bonds ( two versus eight with CD1d and one versus four with αGalCer ) and , in the case of αGalCer , fewer than half ( seven versus 15 ) VDW contacts of those observed in the Vα24+ complex . The residues of the Vα24+ CDR3α were previously shown to be energetically critical for CD1d/αGalCer recognition [17] , a finding recapitulated in our data ( discussed further below ) despite the lower contact number . While the CDR3α loop serves to anchor human iNKT TCRs on the CD1d/lipid platforms with highly similar conformations [16] , [19] , [23] , [25] , the remaining loops have demonstrated rotational flexibility in how they are positioned over the CD1d/lipid surface , in particular at the CDR2β , which establishes energetically critical contacts with CD1d [17] . A similar rotation is seen in the Vα24− TCR docking on the CD1d/αGalCer platform in the complex structure presented here ( Figure 1B and Figure 3A ) . As in the Vα24+ complexes , the involvement of the CDR2β loop in CD1d binding is predominantly mediated by Tyr48 and Tyr50 . Despite an average shift of 4 . 6 Å between the Vα24− and Vα24+ CDR2β CA backbones , the rotationally flexible tyrosine side chains maintain highly similar contacts between the two complexes ( Figure 3A ) . Glu83 on CD1d takes a central role in contact with the CDR2β in both complexes , establishing a hydrogen-bonded network with both Tyr48 and Tyr50 hydroxyls . Met87 also contributes VDW contacts with Tyr50 in both complexes . However , in contrast to the Vα24+ complex , where Glu56 of the CDR2β establishes a robust salt-bridge with Lys86 of CD1d ( 3 . 7 Å distance ) , in the Vα24− complex Lys86 has shifted such that is it 4 . 6 Å from Glu56 ( Figure 3A ) . Thus , the critical contacts of the CDR2β loop are maintained in the Vα24− complex despite large main chain shifts of the CDR2β backbone . The highly variable CDR3β loop has been demonstrated to confer reactivity to specific lipids presented by CD1d by both human [26] and mouse [27] iNKT cells . In the Vα24− complex , the CDR3β loop is well resolved in the electron density and establishes only one weak hydrogen bond and a VDW contact with Gln150 on CD1d's α2-helix via Ser97 ( Figure 3B ) . Thus , unlike the murine Vα10 NKT TCRs , which have CDR3β sequence specificity and use this loop in CD1d binding , this Vα24− TCR does not appear to rely heavily on its CDR3β loop for binding . The availability of a Vα24− TCR also expressing a Vα3 . 1 domain ( named 5B ) [28] in the unliganded state allows a direct comparison between the loop structures between the TCR examined here ( bound to CD1d ) and a Vα24− , Vα3 . 1+ , TCR in its unbound state . Due to the use of different Jβ gene segments that results in global domain orientation shifts , the TCRs are not perfectly superimposable ( Figure 4A ) and there are two amino acid differences in the CDR3α sequences of these TCRs due to junctional diversity ( Figure 4B ) . Alignment of the two Vα3 . 1 domains shows the CDR1 and CDR2 loops are essentially identical structurally ( Figure 4B ) , yet examination of the CDR3α loops ( Figure 4B ) shows significant structural differences . While the unliganded structure of J24 . N22 is not known , modeling of the 5B TCR onto our complex structure suggests a large shift in loop conformation would need to occur in the CDR3α loop for it to dock onto CD1d/αGalCer in a similar fashion . Because of the similarities between these TCRs in all other loops save the CDR3β , it is very likely that the 5B TCR would dock in a similar fashion as seen here . Thus in contrast to the Vα24+ NKT TCRs' recognition of CD1/αGalCer , where loop conformation was highly conserved in the liganded and unliganded state , we suggest that the CDR3α loop can be flexibile in Vα3 . 1+ , Vα24− TCRs , similar to what was previously seen in the iNKT TCR recognition of CD1d/LPC [25] . To evaluate the kinetics involved in binding of our Vα24− TCR with CD1d/αGalCer , we used surface plasmon resonance to measure the association ( kon ) and dissociation rates ( koff ) of this interaction and determine the dissociation constant ( KD ) ( Figure 5A ) . We also used this to calculate KD by equilibrium analysis ( Figure 5A , insets ) . We included an iNKT ( Vα24+ ) TCR in our kinetic measurements such that we could compare these values to a representative of the iNKT population . The affinity of the Vα24− TCR used in this study for CD1d/αGalCer ( 2 . 1 µM kinetic , 2 . 5 µM equilibrium ) was similar to the affinity we measured for the iNKT TCR ( 2 . 1 µM kinetic , 1 . 9 µM equilibrium ) as well as affinities from previous measurements with Vα24− TCRs ( using Vα3 . 1 and Vα10 . 3 domains ) [28] . Stronger affinities ( 0 . 5 µM ) have been noted for other human iNKT TCRs [17] . We sought to further evaluate the residues contributing most to Vα24− TCR binding to CD1d/αGalCer . We chose key TCR residues identified as interacting with CD1d/αGalCer in our complex and evaluated their contribution to binding via alanine-scanning mutagenesis and SPR . We first evaluated the CDR1α loop residues Ser27 and Asn29 , as these appeared to mediate the side-chain-specific contacts that differed most from the Vα24+ TCRs . While mutation of Ser27 to Ala ( S27A ) did not drastically change Vα24− TCR binding kinetics , mutating Asn29 to Ala ( N29A ) resulted in a significant disruption to binding with changes in both the association and dissociation rates and an increase in the KD by an order of magnitude ( Figure 5B ) . Thus the CDR1α loop provides a clear contribution to Vα24− TCR binding to CD1d/αGalCer . Previous mutational analysis of the CDR1α loop of a Vα24+ TCR [17] of Pro28 to Alanine disrupted binding , however this was assumed to be due to changes in the TCR architecture as conformational-specific antibodies failed to bind this mutant . Mutation of the CDR2α side chains Ser50 and Asn51 had subtle effects on kon and koff ( Figure 5B ) yet did not appear to have a substantial effect on the overall affinity of CD1d/αGalCer binding , similar to what we observed with mutation of Ser97 in the CDR3β loop sequence . Because of the similarities in CDR3α loop contacts between Vα24− and Vα24+ TCRs , we included a mutation of Arg95 of the CDR3α as a positive control; this side chain has been shown to be central to iNKT TCR binding to CD1d/αGalCer [17] . We also observed that mutation of this side chain to Ala ( R95A ) abrogated binding of the Vα24− TCR and thus supports the importance of the CDR3α loop to Vα24− TCR docking .
Our complex structure of a Vα24− TCR with CD1d/αGalCer provides a model by which to understand how this diverse population of CD1d-restricted human T cells recognize antigen . These cells differ from iNKT cells in their specificity , effector function , and the markers expressed on their cell surface; these factors combined argue that these cells provide another arm of T-cell-mediated lipid recognition in humans . Here we provide a structural and biophysical foundation upon which to understand the molecular basis of differential reactivity observed at the cellular level in this NKT cell population . Despite the divergent amino acid sequences encoded by the Vα3 . 1 domain for the CDR1α and CDR2α loops , the Vα24− TCR adopts a similar footprint to that of Vα24+ iNKT TCRs . This docking orientation is primarily dictated by the conserved docking of the CDR3α loop , containing the highly similar sequence encoded by the Jα18 segment of iNKT TCRs . The contacts mediated by the other loops , while not identical to those of iNKT TCRs , were very similar , suggesting that despite sequence differences in the Vα loops they could establish contacts with similar regions of the CD1d/αGalCer surface . The αGalCer headgroup position was almost identical to that observed in the iNKT complex structures [16] , [19] . This docking mode , also shared with that of the murine Vα10 NKT TCR [11] , is strikingly different from that of the recently resolved type II NKT TCR structures [9] , [10] , where the TCRs dock on an entirely different surface of CD1d ( the A′ pocket ) and use all six of the TCR's CDR loops in recognition ( similar to what is observed in conventional αβ TCR recognition of MHC/peptide ) . These structures demonstrate that CD1d-restricted T cells can use at least two divergent ways to recognize their antigens [29] . Our complex structure provides a useful model to compare other Vα24− TCRs' structures , notably the structure of a highly related unliganded TCR called 5B [28] . If we assume the 5B TCR would dock similarly to the Vα24− TCR examined in our study , a significant conformational change would have to occur in 5B's CDR3α loop . This conformational flexibility was a feature we also observed in human iNKT TCR binding to CD1d/LPC [25] . In contrast to what was observed with the iNKT TCR complex structure with CD1d/αGalCer [16] , [19] , this suggests that not all CD1d-TCR interactions are “lock and key” and that changes to CDR3α loop conformation may contribute to differences in binding kinetics and thermodynamics . A similar phenomenon of loop movement was observed in the murine Vα10 NKT TCR upon binding [11] . The CDR3α loop footprint on CD1d/αGalCer is conserved in all the iNKT-TCR/CD1d structures noted to date as it is here . However , the number of contacts in this complex structure were less than that observed in the iNKT-TCR CD1d/αGalCer complex structure , yet the binding affinities measured for the Vα24+ and Vα24− TCRs in this study did not differ substantially ( ∼2 µM for both TCRs ) . The alanine-scanning mutagenesis revealed important contributions from the CDR1α loop ( in particular , residue N29 ) in the Vα24− TCR binding that were not noted in Vα24+ TCR binding ( mutation of the equivalent position , S30 in the Vα24+ TCR , showed little effect [17] ) . This shift of importance toward the CDR1α likely compensates for fewer CDR3α loop contacts and would explain the altered reactivity patterns of Vα24− TCRs for lipids that are recognized similarly by Vα24+ TCRs ( such as αGlcCer and αGalCer , discussed more below ) . We cannot rule out that contributions from other loops , such as the CDR2α and CDR3β , contribute as well; while individual mutagenesis of these residues had small effects upon TCR binding , in combination they may have a cumulative effect in binding CD1d/lipid , evident only when they are mutated in concert . Extensive studies in the mouse iNKT cell system have revealed how lipid ligands are structurally modified during recognition by the iNKT TCR . Even though extensive structural variability exists in the glycolipid headgroups , each carbohydrate structure adopts a similar orientation when bound by the TCR [20]–[24] . Therefore , contributions of the CDR1α in recognition of alternative lipids , both α- and β-linked glycolipids , could be an important factor in Vα24− T cell reactivity towards different lipids . Directly relevant to this point is the clear distinction between Vα24− T cells and Vα24+ iNKT cells in their differential reactivity to the α-linked glycolipids αGlcCer and αGalCer . Vα24+ iNKT cells respond well to both lipids , whereas Vα24− T cells do not respond to αGlcCer . The only difference present between these two lipids is the orientation of the 4′OH group on the sugar ring ( glucose versus galactose ) . Our structural and biophysical data provide an explanation for this difference in reactivity . Asn29 , a residue in the Vα24− CDR1α , establishes both VDW and hydrogen bonds with the 3′OH and 4′OH . Mutation of this residue to alanine results in an order of magnitude decrease in binding of the Vα24− TCR , presumably due to disruption of these contacts . Furthermore , the CDR2α loop residues Ser50 and Asn51 establish water-mediated hydrogen bonds with the 4′OH that may help to stabilize the interaction despite lacking clear energetic contributions ( as assessed in our alanine-mutagenesis studies ) . We therefore propose that modification to the 4′OH between the galactose ( αGalCer ) and glucose ( αGlcCer ) structure is the primary molecular factor mediating the differences in reactivity of the Vα24− population of CD1d-restricted T cells . The alternative contacts with the carbohydrate headgroup in the iNKT TCR/CD1d/αGalCer structure may explain why iNKT cells can respond to both lipids; the main contacts with the 4′OH are mediated by Ser30 , which when mutated to alanine only had a minimal effect on binding [17] . The greater number of contacts and BSA of the Vα24+ TCR CDR3α loop on CD1d/αGalCer may make these T cells relatively insensitive to variation in the glycolipid headgroup at other positions . The difference in 4′OH recognition may translate to alternative reactivity to other glycolipid and non-glycolipid lipid structures both in development of these T cells in the thymus and their effector functions in the periphery . Despite their shared use of Jα18 and Vβ11 , the Vα24− T cells are differentiated from iNKT cells in their development and activation state; presumably altered TCR recognition of a selecting antigen during thymic development plays a role in these differences . Our structure provides a model by which to understand the molecular basis of this altered reactivity . Our results , which focus much of the differences in reactivity to αGlcCer on the CDR1α loop and its interaction with the 4′OH , contrast with the murine Vα10 NKT cell preferred reactivity to αGlcCer [11] , where preference in binding appears due to many factors . The highly convergent recognition of αGlcCer by these TCRs distributes the binding contacts over much of the CDR loop surfaces [11] . While mutagenesis data for these residues are not available , it is clear there are differences in the nature of the contacts between the Vα10 and iNKT TCRs with CD1d ( VDW versus hydrogen bonds ) , that many new contacts are established with CD1d , and therefore modification to the sugar ring may have more of a distributed effect over the Vα10 NKT interaction than what we observe in our Vα24− TCR complex structure . Both structures , however , provide molecular models for the observed differences in lipid reactivity and demonstrate how divergent NKT TCR structures can convergently recognize similar CD1d/lipid antigen structures . The molecular basis of the differences in recognition we have described here are the first clues into understanding why Vα24− cells are developmentally and functionally distinct from the iNKT population .
The ectodomain region of human CD1d and human β2microglobulin ( β2m ) were co-expressed in insect cells and purified as described [25] . The cDNAs corresponding to the α and β chains of the Vα24+ NKT TCR clone J24L . 17 and the α and β chains of Vα24− TCR clone J24N . 22 were separately cloned into different versions of the pAcGP67A vector each containing a 3C protease site followed by either acidic or basic zippers and a 6xHis tag . Both chains were co-expressed in Hi5 cells via baculovirus transduction . The heterodimeric TCRs was captured with Nickel NTA Agarose ( Qiagen ) and further purified by anion exchange and size-exclusion chromatography . Mutants of the Vα24− TCR ( S27A , N29A , S50A , N51A , R95A for the alpha chain , and S97A for the beta chain ) were generated through overlapping PCR with specific primers containing the desired mutation . Mutant heterodimeric TCR was expressed in insect cells as described above . Purified human CD1d was used for loading with αGalCer at room temperature with a three molar excess of lipid for 16 h . The excess of lipid was then removed with a Superdex 200 ( 10/30 ) column ( GE Healthcare ) . A human CD1d construct bearing a 3C protease site + 6X-Histidine tag at the C-terminus was expressed in Hi5 cells and purified as described [25] . All interaction experiments were performed in a BIAcore 3000 Instrument ( GE Healthcare ) . Three hundred RUs of wild-type Vα24− NKT TCR or a mutant version of it were captured in a flow channel of an Ni-NTA sensor chip ( GE Healthcare ) previously treated with NiCl2 . Insect-cell-derived recombinant IgFc was used to block unbound sensor chip surface to minimize nonspecific binding events . Increasing concentrations ( 0 , 0 . 037 , 0 . 111 , 0 . 333 , 1 , 3 , 9 , and 27 µM ) of CD1d–αGalCer were injected at a flow rate of 30 µl/min in 10 mM Hepes pH 7 . 4 , 150 mM NaCl , and 0 . 005% Tween-20 . Both kinetic and equilibrium parameters were calculated off of these curves using BIAevaluation software 3 . 2RC1 ( GE Healthcare ) and GraphPad Prism . Nickel agarose-purified Vα24− TCR was digested with 3C protease for 16 h at 4°C to remove the zippers and His tags and purified by anion exchange chromatography in a MonoQ column ( GE Healthcare ) . Endoglycosidase F3 ( EndoF3 ) was used next at a 1∶10 enzyme-to-protein ratio for 2 h at 37°C in order to minimize the sugar content present in the protein . The digested protein was purified by a new round of anion exchange followed by size-exclusion chromatography . Both αGalCer-loaded CD1d and EndoF3-treated Vα24− TCR protein samples were mixed in HBS at 1∶1 molar ratio and concentrated in Nanosep Centrifugal Devices ( Pall Life Sciences ) to 10 mg/ml . Initial hits were found in 0 . 1 M sodium acetate , 20% PEG 4000 , and were optimized to birefringent crystals that grew in 0 . 1 M sodium acetate pH 5 . 0 , 17% PEG 4000 , and 0 . 1 M ammonium acetate . Crystals were cryo-cooled in mother liquor supplemented with 20% glycerol prior to data collection . All data sets were collected on a MarMosaic 300 CCD at the LS-CAT Beamline 21-ID-G at the Advanced Photon Source ( APS ) at Argonne National Laboratory and processed with HKL2000 [30] . The structure of the ternary complex was solved by molecular replacement with the program Phaser [31] using the human CD1d–β2m ( Protein Data Bank ( PDB ) accession number 1ZT4 ) and an iNKT Vα24+ TCR ( 2EYS ) as search models . Refinement with Phenix software suite [32] was initiated through rigid body and followed with XYZ coordinates and individual B-factor refinement . These first steps of refinement yielded clear unbiased and continuous density for αGalCer . Next , extensive cycles of manual building in Coot [33] and refinement were carried out and ligands such as αGalCer or covalently bound sugars were introduced guided by Fo−Fc positive electron density . Ligands structures and chemical parameters were defined with C . C . P . 4 . 's Sketcher [34] and included in subsequent refinement and manual building steps . Translation/libration/screw ( TLS ) partitions were calculated and incorporated at later refinement stages . All the refinement procedures were performed taking a random 5% of reflections and excluding them for statistical validation purposes ( Rfree ) . Intermolecular contacts and distances were calculated using the program Contacts from the CCP4 software package [34] , interface surface areas were calculated using the PISA server ( http://www . ebi . ac . uk/msd-srv/prot_int/pistart . html ) , and all structural figures were generated using the program Pymol ( Schrödinger , LLC ) . Coordinates and structure factors for the J24 . N22 Vα24− TCR/CD1d/αGalCer complex have been deposited in the Protein Data Bank under the accession code 4EN3 .
|
Certain lineages of T cells can recognize lipids as stimulatory antigens when presented in the context of CD1 molecules . We know how most Natural Killer T ( NKT ) cells react with this unusual ligand because they use a single invariant T cell receptor ( TCR ) alpha chain to do the job . NKT cells place particular emphasis on their CDR3α and CDR2β loops in recognition of antigen—these complementarity determining regions ( CDRs ) are the hypervariable parts of the TCR that “complement” an antigen's shape . How do these other T cells recognize closely related yet distinct lipid antigens ? Here we show that human CD1d-restricted T cells , typically called Vα24− T cells due to their use of diverse Vα domains in their TCRs , use similar molecular strategies to respond to lipid antigens presented by CD1d . To this end we present a 2 . 5 Å complex structure of a Vα24− TCR complexed with CD1d presenting the protypical lipid , α-galactosylceramide ( αGalCer ) . The TCR examined in this study notably shifts its binding slightly , placing more emphasis on the interaction with the CDR1α loop as revealed through alanine scanning mutagenesis . This shift explains the inability of these T cells to respond to lipids that vary at this site of contact ( the 4'OH ) , like the related α-linked glucosylceramide . These results provide a molecular basis for the fine-specificity of different CD1d-restricted T cell lineages .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"immunology",
"biology"
] |
2012
|
The Molecular Basis for Recognition of CD1d/α-Galactosylceramide by a Human Non-Vα24 T Cell Receptor
|
The metabolic symbiosis with photosynthetic algae allows corals to thrive in the oligotrophic environments of tropical seas . Different aspects of this relationship have been investigated using the emerging model organism Aiptasia . However , many fundamental questions , such as the nature of the symbiotic relationship and the interactions of nutrients between the partners remain highly debated . Using a meta-analysis approach , we identified a core set of 731 high-confidence symbiosis-associated genes that revealed host-dependent recycling of waste ammonium and amino acid synthesis as central processes in this relationship . Subsequent validation via metabolomic analyses confirmed that symbiont-derived carbon enables host recycling of ammonium into nonessential amino acids . We propose that this provides a regulatory mechanism to control symbiont growth through a carbon-dependent negative feedback of nitrogen availability to the symbiont . The dependence of this mechanism on symbiont-derived carbon highlights the susceptibility of this symbiosis to changes in carbon translocation , as imposed by environmental stress .
The symbiotic relationship between photosynthetic dinoflagellates in the family Symbiodiniaceae [1] and corals is the foundation of the coral reef ecosystem . This metabolic symbiosis is thought to enable corals to thrive in the oligotrophic environment of tropical oceans by allowing efficient recycling of nitrogenous waste products in return for photosynthates from the symbionts [2] . Despite the general acceptance of this assumption , cumulative studies have raised discussions about the molecular mechanisms underlying host-symbiont metabolic interactions . In particular , the role of nitrogen recycling from waste ammonium is still under debate . While it is generally assumed that ammonium assimilation is predominantly performed by the symbiont , some studies indicate that symbiont-growth is nitrogen limited in hospite [3–5] , suggesting that the host might be able to control nitrogen availability . Moreover , it has been suggested that the host might be able to utilize organic carbon [5] , in the form of glucose provided by the symbiont [6] , to promote ammonium assimilation by itself , while suppressing ammonium production from deamination reactions [7] . Consequently , it has been proposed that recycling of ammonium waste by the host might serve as a mechanism to control symbiont densities [5] . Although this potential mechanism to control symbiont densities through nitrogen conservation by the host has been proposed for decades , it still remains highly contentious . Consequently , the coral research field still recognizes nitrogen recycling as a main function of the symbiont [8 , 9] . To better understand these metabolomic interactions , the sea anemone Aiptasia ( sensu Exaiptasia pallida ) [10]—an anthozoan as corals—has emerged as a powerful model system because of the similar symbiotic relationship it forms with Symbiodiniaceae . Multiple symbiosis-centered transcriptomic studies have provided invaluable information on the interactions between Aiptasia and Symbiodiniaceae [11–13] . To generate a more concise set of high-confidence symbiosis-related genes , we adapted a meta-analysis approach , which is a statistical method developed from evidence-based medical research [14] . Because of its statistical power in integrating results from multiple sources , it has been recently applied to transcriptomic studies from both animals [15] and plants [16] , and allows for the identification of high-confidence genes associated with certain biological processes . By carrying out a meta-analysis on available symbiosis-centered RNA-seq datasets , we identified a core set of high-confidence genes and pathways involved in symbiosis establishment and maintenance . To further verify our conclusions made from expression changes of these core genes , we subsequently analyzed metabolomic profiles of symbiotic and non-symbiotic Aiptasia using 13C bicarbonate labeling . Through the integration of these two layers of omics-level information , we identified the pathways associated with nitrogen conservation in the host animal , and further revealed competition for nitrogen as a central mechanism in this relationship that is generally believed to be entirely mutualistic . Based on these findings , we propose a glucose-dependent nitrogen competition model that highlights the sensitivity of the symbiotic relationship to environmental stresses .
The initial PCA performed on samples from individual studies showed a clear separation of samples by symbiotic condition ( S1 Fig ) . This implied that the symbiotic state was the main driver of expression changes in each of the individual studies . To further investigate the relationship between samples from different studies , we performed a principal component analysis ( PCA ) and a rank correlation analysis ( RCA ) on inter-sample normalized transcripts per million ( TPM ) values across all studies . Both the PCA ( Fig 1A ) and RCA ( Fig 1B ) showed clear grouping of samples by experiment rather than symbiotic state . This indicated that technical and/or experimental effects from each study exerted stronger effects on gene expression profiles than the actual symbiotic state of the animals . Although the four datasets grouped distinctly in the PCA analysis ( Fig 1 ) , there was still a clear separation of symbiotic and aposymbiotic replicates within each of the datasets ( S1 Fig ) . We hypothesized that this separation was due to the differential expression of core genes involved in symbiosis initiation and/or maintenance . To identify these genes , we performed four independent differential expression analyses using the exact same pipeline and parameters as described in Materials and Methods . These analyses identified between 2 , 398 to 11 , 959 differentially expressed genes ( DEGs ) , corresponding to ~10–50% of all expressed genes in the respective studies ( Table 1 ) . Since the symbiotic state was supposed to be the main factor driving expression differences between the individuals in each study , we expected to find a great overlap between these lists of DEGs . However , the overlap was poor despite the large number of DEGs identified in the individual analyses: only 393 genes were consistently differentially expressed across all four studies . Out of these 393 genes , 166 were upregulated in symbiotic anemones in all comparisons , while 134 were found to be downregulated in symbiotic animals , relative to aposymbiotic controls ( Table 1 ) . Paradoxically , the remaining 93 genes ( 23 . 7% of all overlapped DEGs ) were differentially expressed in all studies , but in different directions i . e . in some studies they were significantly upregulated while in others they were significantly downregulated . To obtain a more robust set of core genes involved in symbiosis , we performed a meta-analysis with random effects across the four independent differential gene expression analyses ( S1 Table ) . Using this approach , we identified 731 genes that exhibited a more consistent response to symbiosis . To assess the robustness of these genes , we carried out a principal variance component analysis ( PVCA ) . PVCA is an approach to partition the total variance present in the expression data by estimating the contribution of each experimental parameter ( biological or technical ) to the variance , and determine which of these sources are the most prominent [17] . By fitting the expression profiles and the different experimental parameters used in each study ( such as feeding schedule , water source , temperature , etc . , as shown in S2 Table ) into the PVCA , we were able to detect correlations between expression changes and potential effect sources ( Fig 2 ) . For the four individual studies , we found that the symbiotic state of the anemones accounted for a relatively small fraction ( 6 . 5% in raw data , 8 . 4% in normalized data ) of the observed variance . Most of the variance was introduced by study-specific variables such as feeding frequency , days between feeding and sampling , water , light intensity , and temperature . We further noticed that a large proportion of the variance across these four datasets remained unaccountable , suggesting that technical variability , e . g . RNA extraction , library preparation and sequencing , also introduces substantial unwanted heterogeneity to gene expression profiles . When the PVCA was similarly applied to the 731 genes identified through our meta-analysis , we observed that these core genes had a significantly higher association with symbiosis . Symbiosis state accounted for 46 . 6% of the expression variance observed in these genes ( Fig 2 ) . We noticed that smaller gene lists tended to have variances that were better explained by symbiosis state , exemplified by DEG_YL and DEG_EML-36 having better association with symbiosis than DEG_SB and DEG_EML ( Fig 2 ) . Thus , one could argue that the meta-analysis merely achieved better association with symbiosis as it had the fewest genes of interest . To assess this potentially confounding factor , we performed PVCAs on sets of 731 randomly picked genes from each of the DEG lists ( DEG_YL , DEG_SB , DEG_EML , and DEG_EML-36 ) . These were repeated 10 , 000 times ( i . e . , a Monte-Carlo approach ) . These simulations allowed us to estimate that the likelihood of our meta-analysis producing the observed 46 . 6% by random chance was p < 10−4 ( 0 of 40 , 000 trials had symbiosis state accounting for > 46 . 6% of the variance ) . To assess the impact of the previously identified experiment-specific biases , we conducted Gene Ontology ( GO ) and Kyoto Encyclopedia of Genes and Genomes ( KEGG ) enrichment analyses on the DEGs identified using the four independent differential gene expression analyses , respectively . Across the analyses of four independent experiments , 283–645 GO terms and 9–55 KEGG pathways were enriched . However , the functional overlap across all studies was poor: a large proportion of the putatively enriched terms were only identified in a single dataset ( ~75% in GO , and ~65% in KEGG ) ( S2 Fig ) . This finding reflected the previously observed poor overlap of differentially expressed genes across the studies and provided further evidence for the role of study-specific technical factors in driving gene expression profiles . Compared to the independent analyses , the GO and KEGG enrichment of the 731 symbiosis-associated core genes contained fewer significant GO terms ( 204 ) ( S3 Table ) , but comparatively more significantly enriched KEGG reference pathways ( 31 ) ( S4 Table ) . Many of the enriched GO terms and KEGG reference pathways , as well as their associated genes , fit well with processes previously reported to be involved in symbiosis [11 , 18] , including symbiont recognition and the establishment of symbiosis , host tolerance of symbiont , and nutrient exchange between partners and host metabolism which are discussed separately ( S1 Text ) . The enrichment of the KEGG nitrogen metabolism reference pathway ( S3 Fig ) concurs with previous studies that reported the symbiosis-induced upregulation of genes involved in glutamine synthetase / glutamine oxoglutarate aminotransferase ( GS/GOGAT ) cycle in Aiptasia [11 , 18] . The GS/GOGAT cycle has been demonstrated to be the main pathway of ammonium assimilation in plants [19] , bacteria [20] as well as in cnidarians [9 , 11] . Moreover , we found that the GS/GOGAT cycle connects to several symbiosis-related processes that were previously overlooked; of these processes , pathways associated with amino acid metabolism exhibited some of the most extensive changes in response to symbiosis . These findings therefore suggested that amino acid biosynthesis pathways might play a previously undiscovered role in the maintenance of the symbiotic relationship . Amino acid and protein metabolism represented a major symbiosis-related aspect in our meta-analysis . Nine of 31 enriched KEGG pathways ( S4 Table ) and 18 of 125 enriched biological process GO terms ( S3 Table ) were associated with amino acid and/or protein metabolism ( Fig 3 ) . A total of 97 DEGs were involved in these processes , of which 43 were upregulated in symbiotic animals . Interestingly , the DEGs involved in most of the enriched biological processes exhibited consistent expression changes ( Fig 3A ) , i . e . the genes associated with the corresponding process were either exclusively upregulated or downregulated . Further integration of these enriched biological processes and pathways revealed an amino acid metabolism hub in Aiptasia-Symbiodiniaceae symbiosis ( Fig 4 ) . We observed that genes catalyzing glycine/serine biosynthesis from food-derived choline were systematically downregulated in symbiotic anemones . In contrast , the genes involved in de novo serine biosynthesis from 3-phosphoglycerate , one of the glycolysis intermediates , and glutamine/glutamate metabolism were generally upregulated ( Fig 4A ) . The resulting change in amino acid synthesis pathways suggested that symbiotic anemones utilize glucose and waste ammonium to synthesize serine and glycine , which are both main precursors for many other amino acids ( S1 Text ) . Based on these findings , we hypothesized that the host might be using symbiont-derived glucose to assimilate waste ammonium into amino acids . To test this hypothesis , we further analyzed the metabolic profiles of anemones at different symbiotic states using 13C bicarbonate labeling , which can only be fixed by the symbiont through photosynthesis . We first investigated metabolomes of symbiotic and aposymbiotic anemones using nuclear magnetic resonance ( NMR ) spectroscopy . Three metabolites in the de novo serine biosynthesis pathway were highly abundant in symbiotic Aiptasia ( two of them significantly so , p < 0 . 05 ) , while five out of the six intermediates in the alternative glycine/serine biosynthesis pathway using food-derived choline were significantly enriched in aposymbiotic anemones as predicted ( Fig 4B , Fig 5A ) . However , as glucose produces multiple peaks in the 1H NMR spectrum , and most of these peaks overlap with many other potential metabolites in both symbiotic and aposymbiotic anemones , it was not possible to precisely determine glucose concentrations via NMR . Consequently , we performed 13C bicarbonate labeling experiments and compared metabolite profiles of symbiotic and aposymbiotic anemones using gas chromatography-mass spectrometry ( GC-MS ) , in order to test if the glucose is indeed provided by the symbiont and if the downstream usage of symbiont derived organic carbon is in the host . Our experiments confirmed that symbionts provide large amounts of 13C-labeled glucose to the host ( S4 Fig ) and that the 13C-labeling was significantly enriched in many amino acids and their precursors in symbiotic anemones compared to aposymbiotic ones ( S5 Table ) . Moreover , metabolite set enrichment analysis indicated that 13C was mainly enriched in several amino acid metabolism pathways ( Fig 5B ) , which is consistent with the enrichment analysis of the 731 differentially expressed core-symbiosis genes . For the amino acids with good abundance in both symbiotic and aposymbiotic animals , we examined the proportion of 13C in each of them , respectively . Interestingly , we observed relatively stable increases ( ~1 . 5-fold ) of 13C levels in symbiotic animals compared to aposymbiotic ones ( Fig 5C ) . This constant increase suggests a single carbon source ( photosynthesis-produced glucose ) rather than multiple sources ( glucose and symbiont-derived amino acids ) involved in host amino acid biosynthesis . In the latter case , we would expect to identify more amino acid transporter genes being differentially expressed in response to symbiosis , which is not the case according to the meta-analysis . This provides further proof for symbiont derived glucose as the carbon source used for host amino acid synthesis .
Technical variation during experimentation may introduce strong bias in high throughput sequencing studies [21] . This is especially true in the study of symbiotic systems . Since such systems usually feature highly interdependent metabolic interactions , technical variations in culturing , sampling , and/or manipulation can be expected to introduce significant noise in the metabolic processes intertwined with real symbiosis-associated signals . However , this is often overlooked in transcriptomic studies , and especially so in non-model organisms . As we have noticed , the non-experimental parameters sometimes exerted stronger effects on the expression profiles than the symbiotic state in the Aiptasia transcriptomic studies . To reduce the high signal-to-noise ratio , we suggest two potential venues for differential expression studies . Firstly , future transcriptomic efforts should take extreme care to standardize all experimental conditions save for the one under study . For example , culture conditions should be identical across treatments except for the factor under study , treatments should further be performed on multiple independent batches , RNA extractions and library preparation should be carried out on all samples simultaneously . The prepared libraries should also be sequenced on the same run to further minimize technical variation . Secondly , one should not dogmatically adhere to the convention of using p = 0 . 05 as the cutoff for statistical significance . If a study considers one in every three genes as significantly differentially expressed , to a careful reader , the proclaimed significance of those genes is diminished . As the number of DEGs increase , the rate of type I errors would also increase , which makes the discovery of meaningful biological processes more difficult . Consequently , meta-analyses have become a powerful approach for summarizing sequencing data from different trials in order to reduce the biases inherent to single experiments and to increase the statistical power for the identification of underlying common processes [15 , 16] . By applying this approach to Aiptasia RNA-seq data , we were able to deal with the specific variances present in the individual studies , improve the precision in effect size estimation for each of the genes , and eventually identify a group of high-confidence symbiosis-associated candidates . As shown in our Monte Carlo simulations of PVCA , the genes we identified using meta-analysis exhibited significantly higher association with symbiotic state than any of the single experiment analyses . Moreover , the functional enrichment analyses of our core gene set presented more symbiosis-related GO terms and KEGG pathways , rather than the very broad terms identified from individual studies , being enriched . These terms were also enriched significantly with relatively smaller p values in meta-analysis-identified genes and assisted in the understanding of metabolic interactions between host and symbiont . The metabolic interactions between host and symbionts have been of great interest in the study of the cnidarian-Symbiodiniaceae symbiosis [22 , 23] . Among these interactions , the exchange of carbonic and/or nitrogenous compounds between the two partners is arguably a central process that has been the focus of many investigations [5 , 7 , 9 , 24–26] . However , the connections between these two major compounds remains unclear and highly controversial . By combining a meta-analysis of transcriptomic data , metabolomics , and 13C profiling , we demonstrated a host-dependent negative feedback mechanism in the regulation of nitrogen availability to the symbionts , which is driven by symbiont-derived fixed carbon ( Fig 6 ) . The systematic upregulation of genes involved in choline-betaine pathway highlights the heterotrophic state of aposymbiotic Aiptasia ( Fig 6A ) . This also emphasizes the importance of regular feeding in the maintenance of aposymbiotic animals as previously stated [27] . The downregulation of choline transport in symbiotic Aiptasia indicates a decrease of the host’s demand on dietary choline ( Fig 6B and 6C ) . Correspondingly , genes involved in the downstream conversion of choline to betaine and the production of glycine from betaine are also downregulated . The decrease of glycine caused by this downregulation is likely compensated by the metabolism of serine , which can be achieved by the observed upregulation of serine hydroxymethyltransferase ( SHMT , AIPGENE4781 ) , which catalyzes the interconversion of glycine and serine . Interestingly , our results suggest that serine is one of the key components in amino acid interconversion , as the genes involved in its de novo biosynthesis from 3-phosphoglycerate ( one of the intermediates of glycolysis ) were consistently upregulated . The conversion from glutamate to 2-oxoglutarate , catalyzed by the upregulated phosphoserine aminotransferase ( PSAT , AIPGENE17104 ) , may serve as the main reaction to provide amino groups for the biosynthesis of amino acids . Since 2-oxoglutarate is also one of the intermediates in the citrate acid cycle , an increase of glucose provided by the symbionts may also increase the overall activity of the cycle , hence raising the relative abundance of 2-oxoglutarate in symbiotic animals . High levels of 2-oxoglutarate have been reported to induce ammonium assimilation through glutamine synthetase / glutamate synthase cycle in bacteria [28] . Consistent with this finding , we observed upregulation of all the genes involved in this pathway for symbiotic anemones . Metabolomic analyses of symbiotic and aposymbiotic anemones confirmed the predictions derived from our transcriptomic meta-analysis . Most of the metabolic intermediates in the de novo serine biosynthesis using symbiont-derived glucose were highly enriched in symbiotic anemones and showed increased 13C-labeling . Conversely , many of the metabolites from choline-betaine-glycine-serine conversion showed decreased abundance in symbiotic animals . Furthermore , we also identified many other amino acids with significantly increased abundance and 13C-labeling signals , suggesting that serine may serve as a metabolic intermediate for the production of other amino acids . Overall , these results highlight that symbiont-derived glucose fuels ammonium assimilation and amino acid production in the host and that serine biosynthesis acts as a main metabolic hub in symbiotic hosts ( Fig 6B and 6C ) . The strong shifts in host amino acid metabolic pathways induced by symbiont-provided glucose explains the interactions between nitrogen and carbon metabolism in the Aiptasia-Symbiodiniaceae symbiosis . The catabolism of glucose through pathways such as glycolysis , pentose phosphate pathway , and citric acid cycle , not only generates more energy ( in forms of ATP , NADH , and NADPH ) , which is critical to ammonium assimilation , but also produces more intermediate metabolites that can serve as carbon backbones in many biosynthetic pathways such as amino acid synthesis . Our findings thus highlight nitrogen conservation , i . e . the host driven assimilation of waste ammonium using symbiont-derived carbon , as a central mechanism of the cnidarian-algal endosymbiosis [7] . This metabolic interaction might serve as a self-regulating mechanism for the host to control symbiont density through the regulation of nitrogen availability [5] in a carbon dependent manner . This allows for higher nitrogen availability in the early stages of infection ( few symbionts translocating little carbon and requiring little nitrogen ) and gradual reduction of nitrogen availability with increasing symbiont densities ( many symbionts translocating more carbon and requiring more nitrogen ) . The strict dependence of this mechanism on symbiont-derived carbon highlights the sensitivity of this relationship to changes in carbon translocation from the symbiont to the host as imposed by environmental stresses ( Fig 6D ) . Heat-challenged symbionts have been shown to retain significantly more carbon for their own proliferation using the increased nitrogen availability [29] , while exhibiting a significant decline in light utilization efficiency [30] . This indicates that the balance of the negative-feedback system is tipped by climate change-induced heat stress , because such stress disrupts carbon translocation from the symbionts to the host while increasing the amount of nitrogen available to the symbionts . Overall , this sensitive metabolic equilibrium presents a potential molecular mechanism underlying the establishment , maintenance , and breakdown of symbiotic relationships between cnidarian hosts and Symbiodiniaceae .
To collect data for a meta-analysis , we screened for transcriptomic study that focused on cnidarian-Symbiodiniaceae symbiosis using the clonal Aiptasia strain CC7 . We obtained 3 previous RNA-seq studies that met our criteria and provided 4 separate datasets [11–13] . All the datasets were generated on the same platform ( Illumina HiSeq 2000 ) . Three of the datasets contained 101 bp paired-end reads , while the last one contained 36 bp single-end reads . Samples were labeled based on the initials of the first author of published papers: YL , SB , EML , EML-36 . As all raw data from Lehnert et al . [11] was provided as a monolithic FASTQ file , a custom Python script was written to split the reads into its constituent replicates , as inferred from the FASTQ annotation lines . To avoid biases stemming from the use of disparate bioinformatics tools in calling DEGs , data from the four datasets were processed with identical analytical pipelines . Gene expressions were quantified ( in TPM , transcripts per million ) based on the published Aiptasia gene models [12] using kallisto v0 . 42 . 4 [31] . DEGs were independently identified in the four datasets using sleuth v0 . 28 . 0 [32] . Genes with corrected p values < 0 . 05 were considered differentially expressed . To enable direct comparisons of gene expression values between datasets , another normalization with sleuth was carried out on all samples ( n = 17 aposymbiotic and n = 17 symbiotic ) . Principal component analysis ( PCA ) and ranked correlation analysis ( RCA ) were carried out on these normalized expression values to assess the relationship between samples and reproducibility of these studies . Principal variance components analysis ( PVCA ) , a technique that was developed to estimate the extent of batch effects in microarray experiments [17] , was used several times in our study . A PVCA was carried out on raw data to estimate the batch effects in the combined dataset and their possible source in the original experimental designs . Consistently , the normalized data was also assessed for the reduction of batch effects post-normalization . We also performed PVCA on normalized expression values of the DEGs identified in each independent analysis or the final meta-analysis to detect the robustness of DEG calling . For every gene with at least two studies with significant differential expression values , a meta-analysis was performed to determine the overall effect size and associated standard error . Effect sizes from each study i ( represented as wi ) were calculated as the natural logarithm of its expression ratio ( ln Ri ) , i . e . geometric means of all expression values in the aposymbiotic state divided by the geometric means of all expression values in the symbiotic state . Conveniently , this value is approximately equal to the βi value provided by sleuth . As sleuth also calculates the standard error of βi , the variance of ln Ri was not calculated via the typical approximation—instead , the variance vi was directly calculated as vi=SEβi2∙ni where ni represents the number of replicates in study i . To combine the studies , a random-effects model was used . While the use of this model is somewhat discouraged for meta-analyses with few studies as it is prone to produce type I errors [33] , we still opted for its use over the fixed-effects model due to the substantial inter-study variation evident in the PCAs performed previously . Also , the type I error rate could be controlled by setting a more conservative p threshold , if required . The DerSimonian and Laird [14] method was implemented as described below . Studies with individual effect sizes mi were weighted ( w* ) by a combination of the between-study variation ( τ2 ) and within-study variation ( vi ) , according to the formula wi*=1vi+τ2 The between-study variation ( τ2 ) across all k studies was calculated as τ2=max{Q−dfC , 0} where Q=∑wi ( Ti−T¯ ) 2 C=∑wi−∑wi2∑wi The weighted mean ( m* ) was calculated as m*=∑wi*Ti∑wi* while the standard error of the combined effect was SE ( m* ) =1∑wi* The two-tailed p-value was calculated using p=2[1−Φ ( |m*SE ( m* ) | ) ] and then subsequently corrected for multiple hypothesis testing with the Benjamini-Hochberg-Yekutieli procedure [34 , 35] using a Python script . Genes with corrected p < 0 . 05 were considered differentially expressed . For transparency , calculations for all equations were implemented manually in Microsoft Excel ( S1 Table ) following established guidelines [36] . Gene ontology ( GO ) and KEGG pathway enrichment analyses were both conducted on five DEG lists: one each from the four independent datasets , and one from the results of the meta-analysis . Identification of enriched GO terms were conducted using topGO [37] by a self-developed R script ( https://github . com/lyijin/topGO_pipeline ) . A GO term was considered enriched only when its p value was less than 0 . 05 . KEGG pathway enrichment analyses were performed using Fisher’s exact and subsequent multiple testing correction via false discovery rate ( FDR ) estimation . A KEGG pathway was deemed enriched ( or depleted ) only when its FDR was less than 0 . 05 . The results of enrichment analyses were visualized using GOplot [38] . Aposymbiotic Aiptasia strain CC7 and the same strain in symbiosis with Breviolum minutum strain SSB01 ( formerly Symbiodinium minutum SSB01 ) [1] were used for metabolic profiling . All the symbiotic and aposymbiotic anemones were maintained in the laboratory in autoclaved seawater ( ASW ) at 25°C in 12-hour light/12-hour dark cycle with light intensity of ~30 μmol photons m-2s-1 for over three years . Anemones were fed three times a week with freshly hatched Artemia nauplii , and water change was done on the day after feeding . Anemones were rinsed extensively to remove any external contaminations , and starved for two days in ASW and transferred into ASW with 10 mM 13C-labelled sodium bicarbonate ( Sigma-Aldrich , USA ) for another two days before sampling . The four-day starvation period ensured all Artemia had been digested and consumed , hence there was no contamination from Artemia in the samples for NMR and GC-MS . The samples were drained completely on clean tissues to remove any water on surface , then snap frozen in liquid nitrogen to avoid any further metabolite changes in downstream processing . To compare metabolomic profiles of anemones at different symbiotic states , four replicates of each state ( n = 30 individuals per replicate ) , were processed for metabolite extraction using a previously reported methanol/chloroform method [39] . The free amino acid-containing methanol phase was dried using CentriVap Complete Vacuum Concentrators ( Labconco , USA ) . For NMR metabolite profiling , samples were dissolved in 600 μl of deuterated water ( D2O ) , and vortexed vigorously for at least 30 seconds . Subsequently , 550 μL of the solution was transferred to 5 mm NMR tubes . NMR spectrum was recorded using 700 MHz AVANCE III NMR spectrometer equipped with Bruker CP TCI multinuclear CryoProbe ( BrukerBioSpin , Germany ) . To suppress any residual HDO peak , the 1H NMR spectrum were recorded using excitation sculpting pulse sequence ( zgesgp ) program from Bruker pulse library . To achieve a good signal-to-noise ratio , each spectrum was recorded by collecting 512 scans with a recycle delay time of 5 seconds digitized into 64 K complex data points over a spectral width of 16 ppm . Chemical shifts were adjusted using 3-trimethylsilylpropane-1-sulfonic acid as internal chemical shift reference . Before Fourier transformations , the FID values were multiplied by an exponential function equivalent to a 0 . 3 Hz line broadening factor . The data was collected and quantified using Bruker Topspin 3 . 0 software ( Bruker BioSpin , Germany ) , with metabolite-peak assignment using Chenomx NMR Suite v8 . 3 , with an up-to-date reference library ( Chenomx Inc . , Canada ) . For 13C-labelling investigation using GC-MS , dried samples were re-dissolved in 50 μl of Methoxamine ( MOX ) reagent ( Pierce , USA ) at room temperature and derivatized at 60°C for one hour . 100 μl of N , O-bis- ( trimethylsilyl ) trifluoroacetamide ( BSTFA ) was added and incubated at 60°C for further 30 min . 2 μl of the internal standard solution of fatty acid methyl ester ( FAME ) were then spiked in each sample and centrifuged for 5 min at 10 , 000 rpm . 1 μl of the derivatized solution was analyzed using single quadrupole GC-MS system ( Agilent 7890 GC/5975C MSD ) equipped with EI source at ionization energy of 70 eV . The temperature of the ion source and mass analyzer was set to 230°C and 150°C , respectively , and a solvent delay of 9 . 0 min . The mass analyzer was automatically tuned according to manufacturer’s instructions , and the scan was set from 35 to 700 with scan speed 2 scans/s . Chromatography separation was performed using DB-5MS fused silica capillary column ( 30m x 0 . 25 mm I . D . , 0 . 25 μm film thickness; Agilent J&W Scientific , USA ) , chemically bonded with 5% phenyl 95% methylpolysiloxane cross-linked stationary phase . Helium was used as the carrier gas with constant flow rate of 1 . 0 ml min-1 . The initial oven temperature was held at 8°C for 4 min , then ramped to 300°C at a rate of 6 . 0°C min-1 , and held at 300°C for 10 min . The temperature of the GC inlet port and the transfer line to the MS source was kept at 200°C and 320°C , respectively . 1 μl of the derivatized solution of the sample was injected into split/splitless inlet using an auto sampler equipped with 10 μl syringe . The GC inlet was operated under splitless mode . Metabolites in all samples were identified using Automated Mass Spectral Deconvolution and Identification System software ( AMDIS ) with the NIST special database 14 ( National Institute of Standards and Technology , USA ) . The mass isotopomer distributions ( MIDs ) of all compounds were detected and their 13C-labelling enrichment in symbiotic Aiptasia were investigated using MIA [40] . Pathways associated with these 13C-enriched metabolites were explored using MetaboAnalyst v3 . 0 [41] .
|
The symbiotic relationship with photosynthetic algae is key to the success of reef building corals in the nutrient poor environment of tropical waters . Extensive insight has been obtained from both physiological and “omics” level studies , yet , there are still gaps in our knowledge with respect to the metabolic interactions in this symbiotic relationship . In particular , the role of the host in nitrogen utilization and its potential link to symbiont population control still remains unclear . Using a meta-analysis approach on publicly available RNA-seq data and isotope-labeled metabolomics , we demonstrate the presence of a negative-feedback cycle in which the host uses symbiont-derived organic carbon to assimilate its own waste ammonium . This host-driven nitrogen recycling process might serve as a molecular mechanism to control symbiont densities in hospite . The dependence of this regulatory mechanism on organic carbon provided by the symbionts explains the sensitivity of this symbiotic relationship to environmental stress .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"invertebrates",
"carbohydrate",
"metabolism",
"chemical",
"compounds",
"symbiosis",
"statistics",
"metaanalysis",
"carbohydrates",
"glucose",
"metabolism",
"animals",
"organic",
"compounds",
"glucose",
"organisms",
"mathematics",
"research",
"and",
"analysis",
"methods",
"amino",
"acid",
"metabolism",
"mathematical",
"and",
"statistical",
"techniques",
"gene",
"expression",
"chemistry",
"biochemistry",
"nitrogen",
"metabolism",
"eukaryota",
"organic",
"chemistry",
"cnidaria",
"genetics",
"monosaccharides",
"biology",
"and",
"life",
"sciences",
"species",
"interactions",
"physical",
"sciences",
"metabolism",
"statistical",
"methods",
"sea",
"anemones"
] |
2019
|
Host-dependent nitrogen recycling as a mechanism of symbiont control in Aiptasia
|
Characterising the spatio-temporal dynamics of pathogens in natura is key to ensuring their efficient prevention and control . However , it is notoriously difficult to estimate dispersal parameters at scales that are relevant to real epidemics . Epidemiological surveys can provide informative data , but parameter estimation can be hampered when the timing of the epidemiological events is uncertain , and in the presence of interactions between disease spread , surveillance , and control . Further complications arise from imperfect detection of disease and from the huge number of data on individual hosts arising from landscape-level surveys . Here , we present a Bayesian framework that overcomes these barriers by integrating over associated uncertainties in a model explicitly combining the processes of disease dispersal , surveillance and control . Using a novel computationally efficient approach to account for patch geometry , we demonstrate that disease dispersal distances can be estimated accurately in a patchy ( i . e . fragmented ) landscape when disease control is ongoing . Applying this model to data for an aphid-borne virus ( Plum pox virus ) surveyed for 15 years in 605 orchards , we obtain the first estimate of the distribution of flight distances of infectious aphids at the landscape scale . About 50% of aphid flights terminate beyond 90 m , which implies that most infectious aphids leaving a tree land outside the bounds of a 1-ha orchard . Moreover , long-distance flights are not rare–10% of flights exceed 1 km . By their impact on our quantitative understanding of winged aphid dispersal , these results can inform the design of management strategies for plant viruses , which are mainly aphid-borne .
Infectious diseases of humans , animals and plants severely impact the world’s health and economy . To gain knowledge on disease dynamics , powerful mathematical models have been developed [1–3] . However , for predicting the relative efficacies of competing control strategies across realistic heterogeneous landscapes , spatially-explicit in silico simulation models provide the main avenue [2] . The dispersal parameters of such models critically affect the predicted spatio-temporal dynamics of the disease , and thus the predicted outcome of potential control strategies [4] . Obtaining reliable estimates for these parameters is therefore a fundamental issue in epidemiology [5–7] . Models frequently employ dispersal kernels to represent how the probability of dispersal events diminishes as a function of distance , and simulation studies have proven that dispersal parameters can be identified in idealised scenarios [5] . Indeed , this has been achieved for simple models or small-scale datasets [8–13] . Recent advances in Bayesian methods and computing power have enabled fitting more realistic models to larger-scale surveillance data [6 , 14–19] . However , most dispersal kernels are still unknown . Indeed , estimation gets more complex when graduating from idealised toy problems to reconstructing the spatio-temporal dynamics of real epidemics . The first issue is the mismatch between the spatio-temporal coordinates of the epidemic , sampling and model [20] . For example , the timing of key events ( e . g . when a susceptible individual becomes infected ) is often censored ( i . e . known only within certain bounds ) , and failure to account for this can bias estimates . Moreover , the challenge of inference is increased by uncertainty arising from missing observations [21 , 22] or imperfect sensitivity of disease detection [23 , 24] . Further difficulties arise when surveillance data are aggregated at the patch scale because a landscape comprising patches of various shapes or sizes often cannot be summarized by patch centroids without biasing connectivity estimates . All these issues require appropriate correction measures to avoid biased inference and prediction [25] . In the case of aerial vector- or wind-borne diseases , dispersal kernels critically depend on the flight properties of the vectors or infectious propagules [26] . When the probability of dispersal decreases more slowly than an exponential distribution , kernels are termed “long-tailed” and lead to non-negligible long-distance flights [27] . Such events are an important component of disease epidemiological–and evolutionary–dynamics and call for kernel estimation at the landscape scale [28] . However , among plant diseases , there are few available kernel estimates . The dispersal kernel of black Sigatoka ( a fungal disease of banana ) has been estimated experimentally up to 1 km from a point source , based on the direct observation of spore-induced lesions [29] . This is the only available direct estimate at this scale for the dispersal kernel of a plant disease , which reflects the extreme practical difficulties of such field studies and highlights the critical need for developing in silico solutions . A promising way forward is to infer dispersal parameters indirectly , i . e . from spatio-temporal patterns observed in epidemiological data [5] whilst accounting for the added complexity ( outlined above ) of observational studies . This approach has been used to infer the dispersal kernels of the wind-dispersed plantain fungus Podosphaera plantaginis [15] , the fungus Leptosphaeria maculans affecting oilseed rape and dispersed both by wind and wind-driven rain [30] , and two pathogens transmitted only by wind-driven rain: the oomycete Phytophthora ramorum that is responsible for sudden oak death [16] , and the bacterium Xanthomonas axonopodis that causes Citrus canker [17] . A dispersal kernel has been estimated for two other Citrus diseases: Bahia bark scaling of Citrus , a disease with an elusive etiology [13] , and Huanglongbing , which is caused by bacteria from the ‘Candidatus Liberibacter’ genus and transmitted by psyllids [18] . To date , this is the only vector-borne plant disease for which the dispersal kernel is documented . Although aphids are responsible for transmitting almost 40% of more than 700 plant viruses [31] and impose large economic burdens , their dispersal remains ill-characterized at the landscape scale [32 , 33] . For a vast number of aphid-borne diseases , this lack of basic knowledge affects science-based control strategies by undermining the reliability of quantitative risk assessment and predictive epidemiological models . Most aphid-borne viruses belong to the Potyvirus genus and are transmitted in a non-persistent manner , i . e . by winged aphids that acquire and transmit the virus immediately while probing on various plants in search of a suitable host species [31] . Potyviruses are transmitted by a wide range of aphid species , and aphid infectivity is lost after the first probes . For these reasons , estimating the natural dispersal kernel of a potyvirus provides an indirect way of estimating the dispersal kernel of infectious winged aphids . Plum pox virus ( PPV ) is a potyvirus that is listed as one of the 10 most important plant viruses [34] . This virus is the causal agent of sharka , a quarantine disease affecting trees of the Prunus genus ( i . e . mainly peach , apricot and plum ) , reducing fruit yield , quality ( modified sugar content and texture ) and visual appeal ( due to deformations and discolouration ) [33] . Sharka is a worldwide plague that has infected over 50 countries in Europe , Asia , America and Africa [33] , inflicting estimated economic losses of 10 billion Euros over 30 years [35] . The transfer of infected ( possibly symptomless ) plant material can disseminate PPV over long distances [35] , and the natural spread of the disease is ensured by more than 20 aphid species [36] . Virus-infected trees cannot be cured , and insecticides do not act fast enough to prevent the spread of the virus by non-colonising aphids [31 , 37] . In addition , resistant or tolerant peach and apricot varieties are too scarce to provide a short-term alternative to cultivated varieties . However , aphid-mediated transmission can be reduced by removing infected trees as soon as they are detected . As a result , various countries have implemented PPV eradication or control strategies based on regular surveys and removal of trees or orchards when PPV is detected [33 , 35 , 38] . Given the cost of surveillance , tree removal and compensation , these strategies should benefit from model-assisted optimisation , which requires estimating the aphid dispersal kernel . In this context , the aims of this study are: ( i ) to develop a Bayesian inference framework for estimating , from surveillance data , the parameters of a spatially-explicit epidemiological model that accounts for patch geometry and for interactions between disease spread , surveillance and control , ( ii ) to assess through simulations the accuracy and precision of the dispersal parameters estimated under various epidemic scenarios , and ( iii ) to apply our method to 15 years of geo-referenced surveillance data collected during an epidemic of Plum pox virus in order to estimate the dispersal kernel of the aphid vectors .
In the early 1990’s , an outbreak of the M strain of PPV was detected in peach/nectarine patches ( orchards ) in southern France [39] . The plant health services implemented a control strategy based on disease surveillance and removal of symptomatic trees . This process involved the routine collection of patch-level data comprising the observed number of new cases ( trees with PPV-typical discolouration symptoms on flowers and leaves ) and the corresponding inspection dates , as well as patch attributes ( location , planting and removal years , planting density , etc . ) . We aggregated the information about a 5 . 6×4 . 8 km production area over surveillance years 1992-2006 into a unique georeferenced database , with patch boundary coordinates obtained from digitised aerial photographs . With 4820 inspections over 15 years in 553 patches ( mean area: 0 . 95 ha; 52 orchards were replanted in these patches during that period ) , this database is a precious resource for inference on aphid-mediated viral dispersal in patchy ( i . e . fragmented ) landscapes . Moreover , to account for seasonal variation in the number of flying aphids , we used in our model the average ( over 17 years ) weekly number of flying aphids collected from a 12-m-high Agraphid suction tower located within the bio-geographical region of the study area . Our model has a compartmental Susceptible-Exposed-Infectious-Removed ( SEIR ) structure that aims to reduce bias in parameter estimates by accounting for irregular patch geometry , detection-dependent removal , imperfect detection sensitivity , interval censoring of between-compartment transition times , missing data and parameter uncertainty . We address these challenges by: ( i ) integrating a mixture of exponential dispersal kernels over source and receiver patches to compute between-patch connectivity; ( ii ) splitting the infectious state I into hidden ( H ) and detected ( D ) sub-states ( Fig 1 ) ; ( iii ) integrating over uncertainty in the times of transition between compartments; ( iv ) using Bayesian data augmentation and inference . Two versions of our discrete-time spatio-temporal SEHDR model–one for stochastic simulations and the other for Bayesian inference–are described below ( for further details , see Texts A and B in S1 Texts ) . Whole patches are removed and replanted at predefined dates throughout the study period . Each patch i is planted with Ni individuals . At the planting date , a proportion pi of these individuals are infectious ( in state H ) and 1-pi are susceptible ( in state S ) . If patch i is an introduction patch , pi>0; otherwise , pi = 0 . Up to four transition times ( TE , TH , TD and TR ) can be associated with any given individual ( Fig 1 ) , i . e . individuals pass sequentially from state S to E to H to D to R , and all other transitions occur with zero probability . The exposed state E accounts for the latent period , i . e . the time-lag between the infection date TE and the date at which the individual becomes infectious TH . In this discrete-time model ( whose time steps are denoted by the index r ) , the transitions ( denoted by ‘→’ ) between the five compartments are modelled as: S E → i , t r ∼ Binom ( S i , t r - 1 , 1 - e - λ i , t r ) , ( 1 ) lag ( E H → ) ∼ Gamma Tr ( θ 1 , θ 2 ) , ( 2 ) H D → i , t r ∼ Binom ( H i , t r - 1 , ρ i , t r ) , ( 3 ) lag ( D R → ) ∼ Geom Tr ( 1 / δ ) , ( 4 ) where: S i , t r - 1 ( resp . H i , t r - 1 ) is the number of individuals in patch i that are in state S ( resp . H ) at the beginning of the time interval ( tr−1 , tr] , and S E → i , t r ( resp . H D → i , t r ) represents how many of them make the transition from S to E ( resp . from H to D ) in this time interval; the corresponding transition probabilities are 1 - e - λ i , t r for a given individual in state S to incur at least one infection event ( transmission of non-persistent viruses is principally driven by independent vectors ) , and ρ i , t r for the detection of symptoms on an infectious ( H ) individual ( ρ i , t r=ρ when patch i is inspected in ( tr−1 , tr] , and ρ i , t r=0 otherwise ) ; the sojourn times in compartments E and D are determined per individual via random variables lag ( E H → ) = T H-T E and lag ( D R → ) = T R-T D , respectively; the latent period is modelled classically with the flexible gamma distribution , and here the left truncation of GammaTr represents an absolute minimal latent period for sharka [33] to account for seasonality in Prunus phenology and prevent secondary transmission prior to the first winter; the delay between detection and removal is modelled with a geometric distribution where the probability of removal is the same ( 1/δ ) at each time step , up to the right truncation of GeomTr which represents the maximal delay before removal ( detected trees must be removed before the end of the year ) . The force of infection ( i . e . the expected number of transmission events ) incurred by each individual in patch i over ( tr−1 , tr] is defined as: λ i , t r = α t r β N i - R i , t r - 1 ∑ i ′ ( m i ′ i I i ′ , t r - 1 ) , ( 5 ) where α t r is the normalized flight density , i . e . the proportion of annual flights occurring over ( tr−1 , tr]; β is the transmission coefficient , i . e . the annual number of vector flights per source ( infectious ) host that would lead to infection if the recipient host is susceptible; N i − R i , t r - 1 is the number of remaining hosts on which the incoming vectors distribute themselves in patch i , and I i ′ , t r - 1 is the number of infectious hosts in patch i′ over ( tr−1 , tr] . Note that Ni is constant ( i . e . N i = S i , t r + E i , t r + I i , t r + R i , t r ) for all tr between the planting and removal dates of patch i . Finally , the connectivity mi′i is the probability that a vector flight starting in patch i′ terminates in patch i . The connectivity between source patch i′ of area A i ′ and receiver patch i is obtained via: m i ′ i = ∫ x ∈ i ′ ∫ y ∈ i f 2 D ( ‖ x - y ‖ ) d y d x A i ′ , ( 6 ) where x and y are coordinate vectors in ℝ2 , and f2D is the 2-dimensional dispersal kernel [40] . The computation time required to calculate connectivity mi′i between several hundreds of patches prohibits the use of iterative algorithms to directly estimate the parameters of flexible ( e . g . two-parameter ) kernels . Thus , we developed an approach to approximate long-range ( e . g . exponential-power ) dispersal kernels . We defined f2D as a mixture of J components: f 2 D ( ‖ x - y ‖ ) = ∑ j = 1 J [ w j f j 2 D ( ‖ x - y ‖ ) ] , ( 7 ) where the wj are positive mixture weights summing to 1 , and 2hj is the mean dispersal distance for exponential kernel f j 2 D defined as: f j 2 D ( ‖ x - y ‖ ) = e - ‖ x - y ‖ / h j 2 π h j 2 . ( 8 ) Under this mixture formulation , the connectivity becomes: m i ′ i = ∫ x ∈ i ′ ∫ y ∈ i ∑ j = 1 J [ w j f j 2 D ( ‖ x - y ‖ ) ] d y d x A i ′ ( 9 ) = ∑ j = 1 J [ w j ∫ x ∈ i ′ ∫ y ∈ i f j 2 D ( ‖ x - y ‖ ) d y d x A i ′ ] . ( 10 ) This formulation permits the connectivity of each mixture component j to be computed just once , since only the weights wj require updating in an estimation procedure . We set h j = 3 2 × 1 . 08 j − 1 ( and J = 100 ) , to obtain kernel components with mean distances ranging from 3 to 6110 m and higher resolution at smaller distances . To simplify parametrisation , and to avoid identifiability issues with the mixture of exponentials , we restrain weights using: w j = P ( j J | s 1 , s 2 ) - P ( j - 1 J | s 1 , s 2 ) , ( 11 ) where P is the cumulative distribution function of a beta distribution with parameters s1 and s2 . We call any kernel of the form ( Eq 7 ) using exponential kernels ( Eq 8 ) weighted by ( Eq 11 ) a beta-weighted mixture of exponentials ( BWME ) kernel . In order to test whether BWME kernels provide a good approximation of other dispersal kernels , we fitted a BWME kernel to 3 standard [28] dispersal kernel types ( exponential-power , power-law , and 2Dt ) , all with the same mean distance travelled ( 100 m ) . Model fitting was performed by minimizing the total absolute difference between the marginal cumulative distribution functions at 20 , 000 points spaced evenly between 0 and 1000 m . For each type of disperal kernel , 4 values of the shape parameter were tested . Among the four transition times , only TD ( i . e . the time when an infectious individual is detected ) can be known precisely . Let ( t i , 1 , ⋯ , t i , k , ⋯ , t i , K i ) denote the set of Ki inspection dates in patch i ( which may be partly censored by omissions in surveillance records ) . Let p ( TD , i = ti , k ) denote the probability for an individual in patch i to be detected as infected at inspection date ti , k . Data provide the associated number D i , k + of newly detected individuals , and the number D i - of individuals upon which symptoms were not detected in any of the Ki inspections . These variables are modelled as: ( D i , 1 + , ⋯ , D i , K i + , D i - ) ∼ Multinomial ( N i , p ( T D , i = t i , 1 ) , ⋯ , p ( T D , i = t i , K i ) , 1 - ∑ k = 1 K i p ( T D , i = t i , k ) ) , ( 12 ) where Ni is the initial number of trees planted in patch i . A survival model [41] was used to derive p ( TD , i = ti , k ) whilst accounting for censoring , imperfect detection sensitivity , and the expected dependencies between infections ( Text A in S1 Texts ) . The probabilities p ( TD , i = ti , k ) were determined from the set of model parameters Θ , using a smoothed representation of the expected epidemic , and were not conditioned on past observations . Thus , Eq ( 12 ) provides a pseudo-likelihood for the observed data ( Text A in S1 Texts ) . Based on this pseudo-likelihood , Bayesian inference ( for parameter set Θ ) was performed via Markov chain Monte Carlo ( MCMC ) using a Gibbs sampler with embedded adaptive Metropolis-Hastings steps and data augmentation for the unknown planting and inspection dates ( Texts B and C in S1 Texts ) . By data augmentation , we mean the explicit introduction of latent variables [42–44] . To assess the accuracy ( i . e . amount of bias ) and precision ( i . e . amount of variance ) of the estimation of dispersal parameters , 10 epidemics were simulated under each combination of 7 disease introduction scenarios × 3 dispersal kernels × 4 parameter estimation scenarios . All simulations were performed under the same virtual landscape derived from the surveillance database: we retained the spatial coordinates ( and thus the geometry ) of the patch polygons , but all other potential spatio-temporal dependencies were suppressed through the random permutation of orchard-level data including planting densities and patch planting/removal/replanting dates . When density or planting date were missing in the database , their values were drawn from the corresponding empirical distribution . Simulations were performed with 1 time step per day , and 1 survey per patch per year , with inspection days drawn from the corresponding empirical distribution . The transmission coefficient β was fixed at 1 . 5 ( which leads to realistic epidemic dynamics ) and all other parameters were fixed at the expected values of their prior distributions ( Text B in S1 Texts ) . The three simulated kernels correspond to short- , medium- and long-range dispersal . They were parametrised using low-dimension mixtures of exponential kernels ( Eq 7 ) with fixed mean distances and weights ( Table 1 , mixture parameters ) . These were subsequently approximated by the BWME kernel minimizing the Kullback-Leibler ( KL ) distance [45] between the two probability density functions ( Table 1 , simulation parameters ) . The seven introduction scenarios were defined by the following number of introduction patches ( and the initial prevalence pi in these patches ) : 1 ( 25% ) , 5 ( 10% ) , 10 ( 5% ) , 15 ( 2% ) , 20 ( 1% ) , 25 ( 1% ) or 30 ( 1% ) . For a given introduction scenario , all simulations were performed with the same introduction patches , which were chosen at random with the constraint that the first introduction occurred at year 1 and all other introductions occurred before year 6 ( S1 Fig ) . In order to identify whether our MCMC estimation procedure ( Text C in S1 Texts ) encountered identifiability issues with some parameters , we tested 4 estimation scenarios targeting parameter sets of increasing size ( Table 2 ) , with all other parameters fixed at the values used for simulation . Both simulated epidemics and the smoothed epidemics of the pseudo-likelihood started at the beginning of year 1 and stopped at the end of year 22 . Because some MCMC chains became trapped in local maxima associated with negligible likelihoods , we performed 10 MCMC chains under each estimation scenario ( applied to each simulated epidemic ) , which produced 8400 MCMC chains in total . Within each combination of epidemic replicate × kernel × introduction × estimation scenario , we retained the MCMC chain with the highest mean posterior log-likelihood . Then , for each of these 840 chains , indices of accuracy ( resp . precision ) were defined as the mean ( resp . span of the 95% credibility interval ) of the posterior KL distances between the probability density functions f2D ( Eq 7 ) of simulated and estimated kernels . For ease of interpretation , simulated and estimated kernels were plotted using the distribution function of the distance travelled: F 1 D ( ‖ x - y ‖ ) = ∑ j = 1 J ( w j [ 1 - ( 1 + ‖ x - y ‖ h j ) e - ‖ x - y ‖ h j ] ) . ( 13 ) This function is the cumulative version of the 1-dimensional f1D ( i . e . the probability density function of the distance travelled ) , which is obtained by integrating ( marginalising ) f2D ( Eq 7 ) over all directions . Finally , to assess the impact of detection sensitivity ( ρ ) on the accuracy and precision of the estimation of the dispersal kernel , we performed an additional simulation-estimation study . For 99 equally spaced values of ρ between 0 . 01 and 0 . 99 , a unique epidemic was simulated . Each epidemic started at year 1 from a single introduction patch with 25% prevalence , and spread under the long-range kernel scenario ( Table 1 ) . Default values were used for all other parameters . For each of the 99 simulated epidemics , independent estimations were carried out under the most exhaustive scheme ( Θ4 ) with 3 prior distributions for detection sensitivity ρ corresponding to different levels of available prior information ( Text B in S1 Texts ) . For each combination of prior × detection sensitivity , 10 MCMC chains were run , leading to 2970 MCMC chains . For each value of ρ , posterior distributions were inferred using all chains with non-negligible mean posterior likelihood . Using PPV surveillance data , estimation was carried out under the most exhaustive scheme ( Θ4 ) to infer parameters of the spatial SEHDR model . As above , and for the same reasons , we ran multiple MCMC chains and retained the chain with the highest mean posterior log-likelihood ( Text C in S1 Texts ) . The number of introduction patches κ was fixed at integer values in the range 1-24 , and 30 chains were run per fixed κ . This approach was taken because each unit increase in κ adds two parameters ( additional introduction patch identity and initial prevalence ) to Θ , which always increases the posterior log-likelihood ( various uninformative and weakly informative priors were tested ) . Thus , to avoid over-fitting , identification of κ was treated as a model selection problem for which we maximised the Fisher information criterion I ( κ ) ( Text D in S1 Texts ) .
The parameter combinations chosen to test the inference procedure cover a wide range of epidemic behaviour , from local to widespread epidemics and from low to high incidence ( Fig 2 ) . The general trends are that the stochastic variability has less effect than the introduction scenario or kernel type , that more introduction patches generally lead to more widespread epidemics , and that higher disease prevalence in the introduction patches does not necessarily increase the final local cumulative incidence ( S2 and S3 Figs ) . Increasing kernel range generally decreases the cumulative incidence ( S2 and S3 Figs ) , especially near the introduction patches , although these epidemics are more widespread ( Fig 2 ) . A key inovation in our estimation procedure is the BWME dispersal kernel . This kernel provides close approximations to exponential-power and power-law kernels for all tested values of the shape parameter ( S4 and S5 Figs ) . Such flexibility is an interesting property when one does not know which kernel type to assume , which is a common issue . However , the fit to the 2Dt kernels was more approximate ( S6 Fig ) . This is not surprising since the 2Dt kernel is essentially a continuous mixture of Gaussians . Thus , switching the basis functions from exponential to Gaussian ( giving a BWMG kernel ) may greatly improve the fit . The distribution of Kullback-Leibler ( KL ) distances between simulated and estimated kernels demonstrates that estimation accuracy is not affected by the inclusion of sensitivity and latent period parameters in the estimation scheme ( Fig 3A ) . Neither is the median accuracy of the estimated kernels affected much by the range of the dispersal kernel ( Fig 3B ) . However , for longer-range dispersal kernels , KL distances can become more extreme ( Fig 3B ) , and the span and variance of their 95% credibility intervals increase ( S7B Fig ) . This shows that the precision of the estimated kernel decreases with increasing dispersal range . The most influential factor on the accuracy and precision of estimated dispersal kernels is the introduction scenario ( Fig 3C and S7C Fig ) . However , the effect of the introduction scenario is neither strongly related to the number of introduction patches nor to the associated initial prevalence , but rather to the presence of an introduction patch in the dense central cluster of patches ( Fig 3 and S1 Fig ) . The impact of kernel range and introduction scenario on kernel estimation can also be seen by the visual comparison between simulated and estimated kernels ( S8 , S9 and S10 Figs ) . For each of the 3 simulated kernels , the distribution of KL distances was summarised by its minimum , quartile and maximum values across all 7 introduction scenarios × 10 epidemics per scenario . The comparison between simulated kernels and their estimates within the most exhaustive scheme ( Θ4 ) shows that the 3 kernels are very accurately estimated for some simulated epidemics ( left column in Fig 4 and S11 Fig ) . However , dispersal distances are often overestimated , with the median KL distance increasing from 5 . 2×10−2 to 6 . 1×10−2 with increasing kernel range . A closer look at the estimation curves corresponding to the median KL distance reveals that estimated distances do not exceed the simulated distances by more than 0 . 25 on the log10 scale . Dispersal distances are thus overestimated by a factor below 1 . 8 ( 1 . 2 for the mode; see central column in S11 Fig ) . Even for the most challenging of the 70 epidemics simulated with the long-range dispersal kernel ( bottom-right panel in Fig 4 and S11 Fig ) , the difference between the two curves remains below 0 . 6 on the log10 scale . This value translates into less than 4-fold estimation errors ( less than 4 . 3 for the mode; see right column in S11 Fig ) , which is high but still within one order of magnitude . By contrast , precision is very high for all kernel ranges , as indicated by a median span below 0 . 04 for the 95% posterior credibility interval of KL distances ( S7 Fig ) and the corresponding overlapping red lines in each plot of Fig 4 and S11 Fig . The estimated values of the other parameters are generally close to the values used for simulation , but the relative bias varies among parameters , kernel ranges , and introduction scenarios ( S12 Fig ) . Detection sensitivity ( ρ ) is the most precisely estimated parameter , followed by the shape of the latent period ( θ1 ) for which the estimates are also almost unbiased . Bias can be more severe for the scale of the latent period ( θ2 ) and the transmission coefficient ( β ) , with up to 45% under- and over-estimation ( respectively ) in the worst-case combinations of kernel and introduction scenarios ( S12 Fig , top row for θ2 and bottom row for β ) . For these two parameters , the impact of the introduction scenario on parameter estimation increases with kernel range . The simulation-estimation study on ρ shows that the estimation procedure is robust to detection sensitivities below the default value ( 0 . 8 ) used in the rest of this work . Indeed , although reducing ρ reduces ( by definition ) the proportion of detected cases , the link between detection and epidemic control results in a disproportionate increase in the total number of infected hosts as ρ decreases , providing more data for statistical inference–except when ρ reaches extremely small values ( S13 Fig ) . As a result ( see S14 and S15 Figs ) : ( i ) accuracy of kernel estimation is not reduced as detection sensitivity decreases; ( ii ) precision of kernel estimation is only affected when ρ is very close to 0 or 1; ( iii ) increasing the precision of the prior on ρ only affects the accuracy of kernel estimation for ρ>0 . 8 ( i . e . when epidemic size–and thus data available for inference–is strongly reduced by effective control ) . Finally , we note that stochastic variations among replicated epidemics have more influence than ρ on the KL distance between simulated and estimated kernels ( S15 Fig ) . Once validated on simulated epidemics , we used the developed inference framework to estimate the dispersal kernel of Plum pox virus ( and thus of the flight distances of the infectious aphid vectors ) based on survey data . As a first step , we inferred the number of introduction patches . For κ<10 , no combination of introduction patches returned a finite posterior log-likelihood . The Fisher information criterion was maximised at κ = 11 ( Fig 5 ) , indicating that improvement in model fit saturates beyond this point . This suggests that the most robust inference is obtained with κ = 11 . These 11 introduction events among 547-579 orchards planted over 22 years ( planting date is unknown for 32 orchards ) correspond to disease introduction probabilities of 0 . 5 per year and 1 . 90-2 . 01×10−2 per orchard planted . Summary statistics of the posterior distributions of key parameters and percentiles of the dispersal kernel were tabulated for κ = 11 ( Table 3 ) . From the estimated values of s1 and s2 , we derived the weights of the kernel components ( S16 Fig ) , the dispersal kernel , the cumulative distribution function ( Fig 6 ) and the probability density function ( S17 Fig ) of aphid flight distances . These figures , and the estimated quantiles shown in the second part of Table 3 , demonstrate the substantial contribution of long-range dispersal to aphid-borne virus epidemics . Indeed , almost 50% of the infectious aphids leaving a tree land beyond 100 m ( median distance = 92 . 8 m; CI95% = [82 . 6-104 m] ) , and nearly 10% land beyond 1 km ( last decile = 998 m; CI95% = [913-1084 m] ) .
Since the dispersal kernel is the key component of spatial epidemiological models , we focused attention on its estimation and treated the other parameters as nuisance parameters ( i . e . parameters than are inferred to limit bias in the estimation of the distribution of interest ) . S12 Fig shows how simulated and estimated values compare for all nuisance parameters . Recent methodological advances have permitted the extraction of crucial information on the dispersal kernel of four plant diseases from surveillance data [13 , 16–18] and observational studies [15 , 30] . These estimation procedures all account for unobserved infection times , with additional methodological challenges related to large heterogeneous landscapes [16] , introduction from external sources [17 , 18] , or active disease control [18] . The present work handles these various processes and , contrary to the abovementioned studies which all assume a known detection sensitivity , also accounts for this poorly known variable which adds a layer of uncertainty into the surveillance process . Inclusion of parameters for detection sensitivity and the latent period in the estimation procedure ( Table 2 ) barely affects the KL distance between simulated and estimated kernels ( Fig 3 and S7 Fig ) ; hence the inclusion of these extra parameters during inference based on PPV surveillance data . The resulting estimate of detection sensitivity is ρ = 0 . 66 ( Table 3 ) . Although a previous analysis showed that the presence of undetected infectious individuals resulted in slightly overestimated dispersal distances [19] , here we show that our estimation procedure is robust to detection sensitivities far below one , even with weak prior information on ρ ( S14 and S15 Figs ) . The dataset used for inference contains information on the disease status of more than 401 , 000 trees over 15 years , and is associated with a substantial level of censoring ( on the dates of planting , inspection , infection , end of the latent period , and removal ) . For these reasons , using data augmentation to infer the transition times was unlikely to scale successfully to our analysis . Instead , we used a pseudo-likelihood where the unknown numbers of infectious and removed trees were replaced by their expected values . Intuitively , this approach can be expected to work best in highly connected landscapes , where epidemics are more likely to follow their expected course , and to become more erroneous in patchy landscapes where stochastic events can deflect epidemics away from their expected course . This might explain in part why the smaller KL distances in Fig 3C correspond to those introduction scenarios where a source patch was located in the most highly connected region of the study area . A unique feature of the present work is the validation of the estimation of the dispersal kernel through comparing known functions used in simulations and the corresponding functions estimated from these simulated epidemiological data sets . Although this is an intuitive and standard practice [46–48] , previous estimations of plant disease dispersal parameters instead used goodness-of-fit statistics between actual and simulated spatiotemporal patterns as a way to validate their inference models [16–18] . This general trend to rely on goodness-of-fit statistics , without performing simulation-based validation tests , may be due to the high computational burden associated with such validation procedures which require several simulation scenarios and several independent estimations per scenario to assess the accuracy and precision of the estimation algorithms . Since we focus on dispersal kernel estimation , rather than on model predictions as in [16 , 17] , simulation-based validation was useful to demonstrate that , despite the approximations of the pseudo-likelihood , dispersal kernel estimation was generally very precise . Accuracy was often high for short-range kernels , and dispersal distance estimates ranged from very accurate to overestimated for longer-range kernels ( Figs 3 and 4 ) . The same approach also showed that both the precision and the accuracy of dispersal kernel estimation is unaltered when the probability ρ to detect a symptomatic/infectious tree is in the range 0 . 05–0 . 8 ( S15 Fig ) . The observed overestimation is not likely to be caused by insufficient flexibility in the BWME kernel because , even for the 2Dt dispersal kernel ( which the BWME kernel does not fit perfectly ) , the magnitude of the difference between the two kernels is negligible in comparison with the difference between simulated and estimated kernels . It is not likely either to be caused by choosing the MCMC chain with the highest mean posterior likelihood ( among 10 chains ) since this procedure was just used to remove degenerate chains ( and coherence between all other chains was high ) . Although this procedure is rather wasteful of problem-free chains , and provides lower precision than alternative approaches to multi-chain analysis , there is no reason to expect any bias concerning the mean ( or other statistics ) of the posterior distribution . It is most likely that the estimation bias reported here arose from approximations made ( for practical reasons ) within the pseudo-likelihood . Our inference procedure explicitly accounts for patch geometry and patch-level aggregation of surveillance data . Although this choice was data-driven ( infected tree numbers–not individual locations–were included in the database ) , for landscape-scale studies this approach appears to strike an interesting compromise between computational feasibility and spatial realism . Indeed , considering the disease status of over 401 , 000 individuals simultaneously would cause major computational issues given the size of the resulting connectivity matrix . Conversely , spatial models commonly use the coordinates of patch centroids in connectivity calculations ( e . g . [14 , 19] ) . However , this neglects patch geometry and can be expected to bias connectivity estimates ( i ) when patch shapes and sizes are disparate , or ( ii ) when patch dimensions are of the same order of magnitude as the distances between patches . To exemplify ( i ) , consider a small patch located next to a large patch , where many of the propagules leaving the small patch can be expected to land , but a much lower proportion of the propagules leaving the large patch are expected to fall in the small patch . To exemplify the importance of ( ii ) , consider that many more propagules can be exchanged between two large adjacent orchards than would be calculated using the distance between their distant centroids . Although our approach neglects the effects of disease aggregation within patches , it does account for patch size and geometry that both impact disease spread [49] . The use of Eq ( 6 ) to integrate patch geometry , combined with the BWME kernel , can thus be useful for the inference of the landscape-scale dispersal kernels of many wind- and vector-borne diseases . A rigourous assessment of connectivity between patches is also necessary because of its influence on parameter estimation . Our study shows that kernel range affects both the KL distance between simulated and estimated dispersal kernels ( Fig 3 and S7 Fig ) and the cumulative incidence ( S2 and S3 Figs ) . This pattern reflects how parameter identifiability depends on statistical power , which depends on cumulative disease incidence , which in turn depends on landscape connectivity . Short-range kernels imply greater local connectivity than long-range kernels , leading to relatively intense local transmission but reduced transmission at greater distances . Whether or not shorter-range kernels generate larger epidemics depends on the proportion of potential transmission events falling outside host patches , and thus on landscape configuration . Here , larger cumulative incidences were obtained using smaller kernels because , in our patchy agricultural landscape , many dispersal events generated by long-tailed kernels do not end within host patches . Disease introduction scenarios had a substantial effect on the accuracy and precision of the inferred dispersal kernel ( Fig 3 and S7 Fig ) . Surprisingly , this effect does not seem related to either the number of introduction patches or the associated initial prevalence . However , we note that lower KL distances between simulated and estimated dispersal kernels ( in introduction scenarios 1 , 6 and 7 ) are associated with introductions occurring in the highly connected central patches ( S1 Fig ) . The resulting higher cumulative incidence probably improves estimation for the reasons given above . During parameter estimation , we did encounter multi-modality in the posterior likelihood surface , which may arise when fitting ecological dynamic models to data , even without observation error and model mis-specification [50] . For epidemic scenarios with both a short-range kernel and a high number of introduction events , misidentifying some of the introduction patches had a large negative effect on the likelihood , and some MCMC chains were trapped in degenerate solutions . For this reason , we ran the MCMC algorithms many times and carefully compared the posterior likelihoods and parameter estimates of all chains before making inference . We also considered alternative algorithms such as parallel tempering [51] or equi-energy sampling [52] , which increase the likelihood of between-mode transitions . However , the extra computational burden of these approaches was considered superfluous given that the observed differences in the posterior likelihoods of various modes were typically relatively large . Thus , launching a large number of chains to increase the likelihood of identifying the global mode was a reasonable compromise . We have extensively tested this approach , reporting here the results of several thousand MCMC chains , and have found that in practice results are consistent . Overall , inference of epidemiological parameters is easier for epidemics where disease introductions are well characterized , or at least infrequent . Unfortunately , this was not the case with the PPV-M dataset , and estimating the number of introduction patches κ was challenging . Such difficulty is by no means unique to the current study ( see e . g . [17] ) . Reversible-jump MCMC ( RJMCMC ) [53] is a method for performing MCMC when the dimension of the parameter space is unknown and inferred from data . We initially attempted various implementations of RJMCMC , but found it impossible to construct priors that could both prevent over-fitting and provide robust posterior probabilities for κ under a wide variety of epidemiological scenarios . To circumvent this issue we inferred κ based on the Fisher information . This gives a minimum-variance estimator that provides robust inference with a good balance between under- and over-fitting–although it does not permit the estimation of posterior probabilities associated with the various κ . This approach has been used successfully in similar situations [54] . Like most plant viruses , PPV is transmitted by winged non-colonising aphids in a non-persistent manner [33] . To match the characteristics of this widespread transmission process , in our model transmission events are independent ( conditional on infection sources ) and transmission distances directly depend on host locations and on the distance travelled by an aphid within a single infectious flight . Although estimating this aphid dispersal kernel is crucial to plant virus epidemiology , it has long remained elusive . Traditional ecological methods such as capture-mark-recapture provide little information regarding aphid dispersal at the landscape scale [32] . This has been a major obstacle to the parametrisation of models simulating the dispersal of these vectors and the pathogens they spread , as exemplified by the scarcity of landscape-scale models on cereal aphids [55] and by the informed guesses of flight-distance parameters in such models [56] . Here we estimated , for the first time , the dispersal of aphid vectors at the landscape scale . This estimation indicates that 50% of the infectious aphids leaving a tree land within about 90 meters , while about 10% of flights terminate beyond 1 km . Although dispersal estimation from simulated epidemics suggests that these distances may be overestimated , the large number of flights estimated to terminate within some tens of meters of the source tree is consistent with previous studies of within-patch clustering of trees infected by PPV-M [33 , 57 , 58] or PPV-D [38 , 59] . Indeed , one of these studies [38] shows that 50% of the new PPV cases occur within 35-70 m of the nearest previous case; in addition , 10% of the new PPV cases were found beyond 200-460 m from the nearest previous case . Although the proportion of new PPV cases captured within a given radius is not equivalent to a dispersal kernel ( e . g . because the trees are not always infected by the nearest previously detected neighbour ) , the figures are of the same order of magnitude . In particular , both studies highlight the long range of the dispersal kernel . Our estimation of the dispersal kernel at the landscape scale has important consequences . For example , current French regulations enforce at least one visual inspection per year within 2 . 5 km of a detected sharka case ( followed by the removal of all trees with sharka symptoms ) . Our results suggest that less than 3% of flights should thus go beyond this radius ( Fig 6 ) . In a patchy French landscape , most of these aphids would land outside a peach orchard and thus lead to no infection . Such procedures are thus likely to efficiently detect most of the aphid-mediated secondary infections; actually , given the cost of surveillance and the speed of disease spread , this radius may even be oversized . Future work based on this study could aim at the definition of new management strategies against PPV . More generally , our results provide a unique reference point on the epidemiology , simulation and control of the principal group of plant viruses ( i . e . those caused by non-persistant aphid-borne viruses ) , which have a major epidemiological and economic impact . Finally , by focusing on incidence data the presented estimation approach is adaptable to many epidemiological situations , including other vector-borne and airborne fungal diseases .
|
In spatial epidemiology , dispersal kernels quantify how the probability of pathogen dissemination varies with distance from an infection source . Spatial models of pathogen spread are sensitive to kernel parameters; yet these parameters have rarely been estimated using field data gathered at relevant scales . Robust estimation is rendered difficult by practical constraints limiting the number of surveyed individuals , and uncertainties concerning their disease status . Here , we present a framework that overcomes these barriers to permit inference for a between-patch transmission model . Extensive simulations show that dispersal kernels can be estimated from epidemiological surveillance data . When applied to such data collected from more than 600 orchards during 15 years of a plant virus epidemic our approach enables the estimation of the dispersal kernel of infectious winged aphids . This kernel is long-tailed , as 50% of infectious aphids leaving a tree terminate their infectious flight beyond 90 m whilst 10% fly beyond 1 km . This first estimate of flight distances at the landscape scale for aphids–a group of vectors transmitting numerous viruses–is crucial for the science-based design of control strategies targeting plant virus epidemics .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"orchards",
"medicine",
"and",
"health",
"sciences",
"infectious",
"disease",
"epidemiology",
"kernel",
"functions",
"plant",
"science",
"farms",
"probability",
"distribution",
"mathematics",
"plant",
"pathology",
"infectious",
"disease",
"control",
"infectious",
"diseases",
"operator",
"theory",
"probability",
"density",
"epidemiology",
"probability",
"theory",
"agriculture",
"infectious",
"disease",
"surveillance",
"disease",
"surveillance",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences"
] |
2018
|
Estimation of the dispersal distances of an aphid-borne virus in a patchy landscape
|
Oxidative stress is a common etiological feature of neurological disorders , although the pathways that govern defence against reactive oxygen species ( ROS ) in neurodegeneration remain unclear . We have identified the role of oxidation resistance 1 ( Oxr1 ) as a vital protein that controls the sensitivity of neuronal cells to oxidative stress; mice lacking Oxr1 display cerebellar neurodegeneration , and neurons are less susceptible to exogenous stress when the gene is over-expressed . A conserved short isoform of Oxr1 is also sufficient to confer this neuroprotective property both in vitro and in vivo . In addition , biochemical assays indicate that Oxr1 itself is susceptible to cysteine-mediated oxidation . Finally we show up-regulation of Oxr1 in both human and pre-symptomatic mouse models of amyotrophic lateral sclerosis , indicating that Oxr1 is potentially a novel neuroprotective factor in neurodegenerative disease .
Reactive oxygen species ( ROS ) are the natural by-products of many essential biological processes such as mitochondrial respiration , although they are also potentially damaging to cells . Consequently , eukaryotic organisms have evolved a comprehensive range of proteins to detoxify ROS and repair against any unwanted oxidative damage to DNA , lipids or proteins . These antioxidants include enzymatic scavengers such as superoxide dismutase ( SOD ) and catalase , glutathione peroxidase ( Gpx ) and peroxyredoxins , as well as non-enzymatic factors including glutathione , flavonoids and vitamins [1]–[2] . Oxidative stress occurs when the antioxidant response is insufficient to balance the production of ROS; this state can ultimately cause cell death by apoptosis or necrosis via an array of signalling pathways , and many studies both in vitro and in vivo have demonstrated that the normal function of antioxidant defence systems is vital for cell survival [3] . For example , mouse knockouts representing the most critical mitochondrial antioxidant genes are often lethal at the pre- or early post-natal stage , including glutathione peroxidase 4 ( Gpx4 ) , thioredoxin 2 ( Thx2 ) and SOD2 [4]–[6] . In recent years there has been a particular focus on the role of ROS in neurons , driven by the consistent presence of various oxidative stress markers in neurodegenerative disease , as well as several pathogenic mutations in proteins that feature prominently in antioxidant pathways [3] . Furthermore , it appears that the brain is more vulnerable to ROS damage compared to other organs due to its high metabolic rate combined with a relatively low concentration of antioxidant proteins [7] . Indeed , oxidative stress and mitochondrial dysfunction have been implicated in all major neurodegenerative disorders , including amyotrophic lateral sclerosis ( ALS ) , Parkinson's and Alzheimer's disease ( PD and AD ) [3] , [8]–[9]; yet , despite numerous attempts to recapitulate human disease pathology in mouse models , it is unclear how the timing and disruption of endogenous ROS defence pathways might lead to such heterogeneous neuropathological features [10] . Consequently , with speculation that the up-regulation of antioxidants may be a practical therapeutic target for neurological disease [11] , the hunt continues for new proteins that are key players in the oxidative stress response . In one such search for human factors induced under oxidative stress , Volkert et al . identified oxidation resistance 1 ( OXR1 ) as a novel gene that was able to suppress DNA damage in Escherichia coli oxidative repair-deficient mutants [12] . They went on to report that the human protein , when localised to the mitochondria , was sufficient to prevent oxidative damage in Saccharomyces cerevisiae mutants lacking Oxr1 [13] . Indeed , the gene is found in all eukaryote genomes , although in lower organisms its sequence is restricted predominantly to the highly conserved C-terminal ( TLDc ) domain [14] . In humans , the TLDc domain-containing gene family is composed of four proteins in addition to OXR1 , including nuclear receptor coactivator 7 ( NCOA7 ) and TBC1D24 [14]–[15] . Significantly , a mutation in the TLDc domain of TBC1D24 recently has been found in Familial Infantile Myoclonic Epilepsy ( FIME ) [16] . The function of this domain has not been established , yet it was originally identified as a catalytic motif [17] . Studies have demonstrated that Oxr1 is induced under oxidative stress [13] , [18]; however , virtually nothing is known about this obviously evolutionary significant gene or the TLDc domain itself in mammalian systems . Here we have used a combination of in vivo and in vitro approaches to show that the levels of Oxr1 are critical for neuronal survival and that up-regulation occurs in both human disease and mouse models of neurodegeneration . In addition , we demonstrate that the conserved TLDc domain alone is sufficient to confer functionality in the mouse . This study therefore reveals the vital role of Oxr1 in oxidative stress-related neurodegeneration .
We identified the recessive Bella ( bel ) mouse as part of our screen for mouse models of human movement disorders and ataxia from a large-scale mutagenesis programme . Bel mice are indistinguishable from their control littermates at 2 weeks of age ( P14 ) ; however they rapidly develop a severe ataxic gait ( see Video S1 ) , fail to gain weight as quickly as controls , and do not survive beyond P26 . Pathological analysis of the bel CNS revealed significant and increasing number of apoptotic cells in the granule cell ( GC ) layer ( GCL ) of the cerebellum ( Figure 1A and 1B ) . The onset of cell death occurs from P18-19 , after which there is a highly significant increase in apoptotic cells in the following days ( Figure 1C ) . No cell death was observed in any other region of the brain or spinal cord in end-stage mutants , however ( data not shown ) . The relative size , structure and foliation pattern of the cerebellum was not affected in late-stage bel mutants as determined by quantitative histological methods ( Figure S1A , S1D , and S1E ) and no significant difference in the GCL width was observed , reflecting the relatively small proportion of apoptotic cells in mutant mice ( Figure S1B and S1F ) . Purkinje cell ( PC ) death is frequently associated with GC loss [19]–[22] therefore the relative density of PCs was calculated from bel mice , although no reduction was observed compared to controls ( Figure S1C and S1G ) . Quantitative histopathology of skeletal muscle was also carried out on end-stage bel mice . A significant increase in centrally nucleated fibres was observed in the diaphragm of mutants compared to controls , indicative of muscle degeneration , but not in the tibialis anterior ( TA ) or soleus muscles of the hindlimb ( Figure S1H and S1I ) . Heterozygous ( bel/+ ) mice aged up to 18 months of age display no neuropathological or gait abnormalities ( data not shown ) . An initial genome scan followed by further genetic mapping using polymorphic microsatellite and SNP markers reduced the critical region containing the bel mutation to 5 . 5 Mb on chromosome 15 . Unexpectedly , during candidate gene sequencing , exons representing the genes Oxr1 and Muscle Activator of Rho Signalling ( STARS or Abra [23] ) could not be amplified from bel DNA . Therefore , genomic walking using chromosomome 15-specific PCR primers followed by inverse PCR was used to identify the boundaries of the apparent spontaneous deletion; the missing region was confirmed as 193 . 5 kb , ablating the expression of both Oxr1 and Abra ( Figure S2A and S2B ) . To confirm no additional ENU-generated mutation was segregating with the bel phenotype , all annotated coding and non-coding transcripts in the critical region were sequenced and no mutations were identified . In addition , qRT-PCR confirmed that the loss of potential regulatory sequences did not influence the expression of all adjacent transcripts within the bel critical region ( data not shown ) . Expression studies were then carried out to determine the distribution of both deleted genes in the central nervous system ( CNS ) . In situ hybridisation and RT-PCR showed that while Oxr1 was expressed in the cerebellar GCL , Abra could not be detected in the cerebellum or the rest of the brain ( Figure 2A and Figure S3D ) . Further analysis of the developmental expression patterns showed that Oxr1 is highly expressed in all major regions of the postnatal brain and spinal cord at the RNA level ( Figure 2B and Figure S3A ) , although Abra could only be detected in skeletal muscle tissue by in situ hybridisation and RT-PCR ( Figure S3B and S3D ) ; these data are consistent with previously published expression data on both genes [13] , [23] . In the mouse , several isoforms of Oxr1 have been described , including the shortest isoform that includes only the TLDc domain-containing exons 10 to 16 ( or 11 to 16 ) with a unique first exon ( exon 9 ) ( Oxr1-C , also known as C7C [24]; for detail see Figure S7 and Figure S3C ) . In situ hybridisation using isoform-specific probes demonstrated that both the Oxr1-C and full-length ( Oxr1-FL ) transcript variants showed an essentially identical expression pattern ( Figure S3E ) , in agreement with the riboprobe common to both isoforms used above ( Figure 2B ) . An antibody raised against the same common C-terminal end of Oxr1 ( Figure S7 ) confirmed high levels of Oxr1 protein in the brain , with no signal in bel tissue as expected ( Figure 2C ) ; these data also serve to demonstrate the specificity of the antibody . Taken together , these data suggest that loss of Oxr1 and not Abra is responsible for the neuropathology observed in bel mutants . As conclusive proof that deletion of Oxr1 causes the bel phenotype , we performed a genetic rescue experiment with two independent Oxr1 transgenic lines . Ubiquitous expression of the full-length Oxr1 cDNA ( Oxr1-FL ) in the brain was confirmed by in situ hybridisation in bel mutants carrying the transgene ( bel/bel; Tg ( CAG-Oxr1 ) +/− ) ( Figure 3A ) . Animals of this genotype displayed no ataxia or growth defects , and no cell death was detected in any region of the brain , including the cerebellar GCL , compared to littermates that did not contain the Oxr1 transgene ( bel/bel; Tg ( CAG-Oxr1 ) −/− ) ( Figure 3B ) . This rescue of the bel phenotype is maintained in bel/bel; Tg ( CAG-Oxr1 ) +/− mice aged to 8 months of age ( Figure 3B ) . These data confirm that neurodegeneration in bel mice is caused by Oxr1 . To demonstrate that loss of Oxr1 rendered neurons from bel mice more vulnerable to ROS , primary GCs were assayed for hydrogen peroxide ( H2O2 ) sensitivity . The assay conditions were first optimised to facilitate measurements of cell death in the presence or absence of Oxr1 ( Figure S4A ) . These data confirmed that bel mutant GCs are significantly more susceptible to exogenous peroxide-induced apoptosis than controls ( Figure 4A ) . To further investigate the specificity of this effect , all Oxr1 isoforms were then knocked-down by an shRNA to <10% of endogenous levels in wild-type GCs , resulting in almost twice the level of cell death compared to neurons transfected with control constructs ( Figure 4B and Figure S4B ) . Conversely , replacement of Oxr1 in bel GCs by lentiviral expression rescued the level of apoptotic cell death in H2O2-treated cells down to wild-type levels ( Figure 4C and Figure S4B ) ; thus once again strongly suggesting that loss of Oxr1 alone is responsible for the bel phenotype . Significantly , lentiviral over-expression of Oxr1 in wild-type GCs lead to a significant reduction in apoptosis compared to cells expressing endogenous levels of the gene ( Figure 4C and Figure S4B ) , demonstrating that Oxr1 can also be protective to neurons exposed to stress . In bel end-stage mice , apoptotic cell death is specific to the cerebellar GCL , despite high levels of expression in other brain regions including the wild-type cortex ( Figure 2B ) . To therefore investigate whether loss of Oxr1 would also render non-cerebellar neurons susceptible to oxidative damage , we assayed cell death in primary cortical cells ( CCs ) from bel mice in parallel with cerebellar GCs; both cell populations were cultured to correspond to P14 and P21 in vivo , respectively ( Figure 4D ) . These data show that no increase in cell death occurs in response to peroxide treatment in bel GCs after 7 days in culture; however , after 14 days of culturing a significant increase in apoptosis ( approximately 80% ) is observed in mutant cells versus wild-type . Although similar results were obtained from CCs , interestingly a much smaller increase in cell death is seen ( approximately 20% ) in mutants at the second timepoint ( Figure 4D ) . This suggests that in the cerebellum Oxr1 levels play a more defining role in neuronal survival , consistent with the specificity of neurodegeneration in bel mice . Previous studies have described the presence of OXR1 in the mitochondria of HeLa cells [13] , but also in the nucleus and nucleolus in other mammalian cell lines using a different antibody [24] . Thus to clarify the localisation of Oxr1 in neuronal cells , immunofluorescence was carried out in wild-type GCs . Using our antibody , Oxr1 was not detectable in GCs unless the cells were treated with H2O2 , which clearly induced protein expression ( Figure S4C ) . In these treated cells , Oxr1 also co-localised with the mitochondrial marker Cox4 , consistent with published studies [13] . To determine whether similar induction and localisation was common to other neuronal cell lines , the localisation studies were repeated in N2A cells , generating essentially identical results ( Figure S4C ) . These data demonstrate stress-induction and predominantly mitochondrial localisation of Oxr1 in neurons . In view of the link between Oxr1 and oxidative stress , we then screened for markers of oxidative stress in bel mice . 8-OHdG staining was detected exclusively in the mutant GCL at P24 , indicative of oxidative DNA damage ( Figure 5A ) . In agreement with the apoptotic markers , virtually no DNA damage was detectable prior to P24 ( Figure 5B ) . To further quantify DNA fragmentation due to loss of Oxr1 , the DNA strand scission factor from GCs was calculated using a picogreen assay , showing a significant increase in DNA breaks in bel GCs subjected to H2O2 treatment ( Figure 5C ) . We then analysed a large range of both direct and indirect markers of oxidative stress in addition to antioxidant enzymes from the cerebellum of end-stage ( P24 ) bel mice by qRT-PCR ( Figure S5A ) . These data identified an approximate 70% reduction in expression of glutathione peroxidase 1 ( Gpx1 ) in mutants , although no other genes showed significant differences between the genotypes ( Figure S5A and S5B ) . We went on to test key antioxidants at the protein level , but found no deregulation of the protein expression or enzyme activities of Gpx or catalase in the bel cerebellum ( Figure S5C , S5D , S5E ) . Using the same assays , there was no evidence for oxidative stress in brain regions outside of the cerebellum ( data not shown ) . These data combined with the 8-OHdG results suggest that the bel cerebellum does show some signs of oxidative stress response due to the loss of Oxr1; although these are clearly limited in vivo by the relatively small proportion of neurons affected in end-stage mutant animals . The cellular and tissue data combined suggest that the effect of Oxr1 deletion is highly specific to GCs in bel mice in vivo . We therefore investigated whether Oxr1 may also influence sensitivity to other cellular stress factors using serum starvation in cultured GCs . These data show that there was a significant ( approximately 7-fold ) increase in apoptosis in GC neurons cultured without serum , although no difference in the levels of cell death was observed between bel and control GCs ( Figure S6 ) . These data suggest that loss of Oxr1 does not influence sensitivity to all cellular stress conditions . As discussed above , the C-terminal TLDc domain is highly conserved in all Oxr1 orthologues as well as being highly expressed in the brain ( Figure S3D ) . To therefore investigate whether the short Oxr1-C isoform was functional in neurons , we repeated the peroxide sensitivity assays in bel GCs using constructs coding for this isoform as well as Oxr1-FL . In these experiments , a bicistronic vector containing GFP was used to assay the proportion of transfected cells that were apoptotic ( Figure 6A ) . These data show that , despite the removal of over 500 amino-acids from the N-terminus , Oxr1-C is able to confer protection against oxidative stress as efficiently as the full-length protein in both wild-type and bel GC culture . To determine whether the presence of only short Oxr1 isoforms would be detrimental to neuronal survival in vivo , a gene-trap mouse ( Oxr1Gt ( RRR195 ) Byg ) was rederived containing the vector insertion between exons 3 and 4 of Oxr1 ( Figure 6B and Figure S7 ) ; mice carrying two copies of this insertion are therefore expected to only express shorter isoforms of the gene . Mice homozygous for the insertion were successfully generated and the exact position of the trap vector confirmed ( Figure 6B ) . These mice displayed no gait or pathological abnormalities in the CNS up to 12 months of age ( data not shown ) and western blotting confirmed that the gene-trap insertion had almost completely ablated the expression of Oxr1-FL as expected ( Figure 6C ) . Interestingly , the proportion of the smallest isoform ( Oxr1-C at approximately 25 kDa ) was much higher in these cerebellar extracts than in whole brain ( Figure 6C and Figure 2C ) ; this is consistent with the isoform-specific in situ hybridisation data ( Figure S3E ) and suggests that Oxr1-C may play a more significant functional role in the cerebellum than other regions of the CNS . These data demonstrate that TLDc domain-containing Oxr1 isoforms other than the full-length protein are functional . To gain some insight for the first time into the mechanism of Oxr1 function , taking into account the results from peroxide stress experiments in GCs , we investigated whether Oxr1 could react directly with H2O2 . Recombinant Oxr1-C protein was purified ( Figure S8A ) and an Amplex Red assay was used to quantify decreasing H2O2 concentration in the presence of increasing concentrations of Oxr1-C . These data demonstrate that Oxr1-C is able to significantly decrease the H2O2 levels in a dose-dependent manner ( Figure 7A ) . As a negative control , the same assay was carried out in the absence of horseradish peroxidase ( HRP ) that is an essential part of the Amplex Red reaction . These data show no change in Amplex Red signal in the presence of Oxr1-C , suggesting that Oxr1 is not able to compensate for the HRP activity in this assay and is therefore unlikely to possess peroxidase activity . We next examined whether these data could be due to direct oxidation of the Oxr1 protein . Several amino-acids have the potential to undergo oxidative modification [25] , but we began by analysing the oxidation state of cysteine residues considering that a C-terminal cysteine in the TLDc domain is conserved in Oxr1- and Ncoa7-related sequences found in human , mouse , fly and yeast ( Cys753 in mouse Oxr1; Figure S8B ) . To quantify the oxidation of sulfhydryl ( SH ) side chains , recombinant wild-type Oxr1-C protein was incubated with H2O2 and then reacted with ThioGlo-1 , a thiol-active fluorophore [26] . Samples were then separated by gel electrophoresis followed by densitometric analysis ( Figure 7B and 7C ) . These data show that a significant loss of thiol labelling in Oxr1-C of approximately 2-fold occurs in the presence of H2O2 , indicating that oxidation of free SH groups is taking place . To further examine the significance of cysteine residues in the TLDc domain of Oxr1 , the conserved cysteine was mutated ( C753A ) and the recombinant protein assayed as above ( Figure 7B and 7C and Figure S8A ) . Independently , a second cysteine found in Oxr1 proteins in vertebrates but not the related Ncoa7 was also mutated and analysed ( C704A; Figure 7B and 7C , Figure S8A ) . Peroxide-treated C704A Oxr1-C showed a similar 2-fold reduction in ThioGlo-1 labelling as wild-type Oxr1-C; however , the C753A mutant protein showed a non-significant loss of fluorescence , suggesting that this particular cysteine is more important for the oxidation state of Oxr1 than C704 ( Figure 7B and 7C ) . As a positive control for these studies , DJ-1 ( PARK7 ) , a protein that has been well-studied with respect to cysteine oxidation [27]–[28] , was analysed in parallel ( Figure S8A ) . Recombinant wild-type mouse DJ-1 showed a similar reduction in ThioGlo-1 labelling upon peroxide treatment to wild-type Oxr1-C ( Figure 7B and 7C ) . To ascertain the rate of the reaction between wild-type Oxr1 and H2O2 , direct kinetic measurements using HRP competition assays were attempted [29] , but the apparent low levels of reactivity between Oxr1-C and peroxide rendered this approach impractical ( data not shown ) . Therefore , ThioGlo-1 labelling experiments were repeated over a time-course , generating a rate constant of Oxr1-C oxidation by H2O2 of 0 . 82 M−1⋅s−1 based on second-order kinetics ( Figure 7D ) . To then relate the significance of the cysteine mutants to consumption of H2O2 in the Amplex Red assay , both were assayed as above in parallel with wild-type Oxr1-C recombinant protein . In agreement with the thiol labelling assay , the C753A mutant showed a non-significant level of peroxide consumption , whereas a significant drop in fluorescence was observed using the C704A Oxr1-C mutant ( Figure 7E ) . In summary , these data suggest that Oxr1 can react directly with H2O2 , although primarily through the oxidation reactive cysteine residues . Considering the numerous links between oxidative stress and neurodegenerative disorders , and the high levels of Oxr1 in the spinal cord ( Figure S3A ) , we then analysed OXR1 expression in ALS human biopsy samples . Western blots of ALS spinal cord tissue show an obvious up-regulation of the intermediate TLDc-domain-containing OXR1 isoforms compared to age-matched controls ( Figure 8A ) . As these data were obtained from patients at the end-stage of disease , it was also important to ascertain whether up-regulation of Oxr1 occurs before any major neuropathological changes . Therefore , we analysed protein levels in spinal cord tissue from pre-symptomatic low-copy G93A mutant superoxide dismutase 1 ( SOD1 ) expressing transgenic mice , a model of ALS . These data show a significant up-regulation of Oxr1 in SOD1 mutants at 5 months of age compared to littermate controls ( Figure 8B and 8C ) . Importantly , this represents a timepoint prior to the first reported signs of neuropathology or oxidative stress in this particular line [30]–[31] , suggesting that Oxr1 may be a novel early marker of specific neurodegenerative pathways .
Combining results from three mutant mouse lines , cellular assays and biopsy samples , we have demonstrated for the first time the importance of Oxr1 in neuronal survival; indeed , our data show that the sensitivity of neurons to exogenous stress can be exquisitely controlled by the level of Oxr1 expression . Oxr1 therefore has much in common with some of the most important antioxidants [3]; proteins such as SOD2 can be lethal when disrupted but neuroprotective when over-expressed in vivo and have therefore been nominated as potential therapeutic targets in neurodegenerative disease [6] , [32] . Other key mitochondrial proteins have also been linked to ataxia in mouse models . Apoptosis-inducing factor ( Aif ) is vital to oxidative phosphorylation , and an 80% reduction in expression of the gene in the Harlequin ( Hq ) mutant causes ataxia and oxidative stress-related GC loss [33] . The Hq phenotype is far less severe than bel , however , with the first signs of apoptosis in the cerebellum not appearing until 4 months of age , followed by necrotic Purkinje cell death and degeneration of other brain regions [19] , [34] . These models , along with bel mutants , emphasise that the brain is clearly vulnerable to oxidative stress . Taken further , the fact that GCs are more susceptible to ROS insults than other neuronal populations , as observed in bel and Hq mice [33] , has been considered recently in detail . Using a combination of expression and biochemical data , Wang et al . discovered that cerebellar GCs were more susceptible to exogenous oxidative stress than neurons from the cerebral cortex or hippocampal CA3 region [35] . Lower expression of energy generating genes , combined with a greater depletion of stored ATP , was also observed in GCs versus stress-resistant neurons; this suggested that a shortfall in the energy required to carry out cellular repair might render GCs particularly sensitive to ROS . These data may also explain why we observed that bel cortical cells were less sensitive to peroxide treatment than GCs from bel mice . Neurodegenerative disease often presents with highly specific pathological lesions despite widespread or ubiquitous expression of the mutated gene ( s ) involved . Therefore , examples of selective neuronal vulnerability to oxidative stress , such as bel , are vital to understand why certain neurons are targeted while others are spared , particularly in the early stages of disease [36] . Several splice variants of Oxr1 have been described previously , although we are the first to show that the shortest of these , Oxr1-C , is able to protect neurons from oxidative stress as efficiently as the full-length ( Oxr1-FL ) protein . Our western blot data is also the first to demonstrate what appears to be a complex differential distribution of Oxr1 isoforms at the protein level . For example , Oxr1-C ( 25 kDa ) is present in the whole brain at very low levels , although in the cerebellum this splice variant is as highly expressed as Oxr1-FL ( 85 kDa ) whereas other intermediate isoforms are absent . As Oxr1 in the apparently normal gene-trap mouse is almost exclusively represented by the short Oxr1-C protein , we postulate that Oxr1-C plays a more important role in the response of GCs to stress than elsewhere in the CNS . Therefore , loss of this particular isoform as well as the full-length protein , combined with the apparent vulnerability of GCs , leads to cerebellar-specific apoptotic cell death in bel mice . Original descriptions of human OXR1 induction by oxidative stress were restricted to intermediate isoforms ( at approximately 40 and 58 kDa ) , based on known splice variants starting upstream of , but including , the TLDc domain [12]–[13] . Although the full-length protein was therefore not addressed in these early experiments , a construct representing the 40 kDa isoform was still able to confer protection against ROS in yeast [13] . Importantly , these original studies are in agreement with our investigation that unequivocally demonstrates that these shorter splice variants , all containing the TLDc domain , are indeed functional . We chose ALS to model the in vivo induction of Oxr1 as oxidative stress has been consistently implicated in the human disease and in ALS mouse models , with multiple markers of oxidative damage observed in ALS post-mortem tissue recapitulated in SOD1 transgenic lines [37] . The striking up-regulation of intermediate OXR1 isoforms we observed in ALS may be a consequence of significant neurodegeneration in the spinal cord; but crucially , full-length Oxr1 protein levels were significantly increased in SOD1 G93A mice before any overt phenotypic abnormalities , suggesting Oxr1 may be an early marker of neurodegeneration [30]–[31] , [38]–[39] . The fact that alternate isoforms were differentially regulated between human and mouse may reflect species-specific post-transcriptional regulation of Oxr1 or the dissimilar stages of disease examined . In summary , it is likely that alternate Oxr1 isoforms have specific purposes; however it is clear from our work that deciphering the function of the highly conserved TLDc domain will be key to understanding the role of Oxr1 and Oxr1-related proteins [14] . In view of this , we assayed recombinant Oxr1-C in a peroxide scavenging assay and discovered that this region of the protein can react directly with H2O2 in vitro . This is the first direct evidence that the TLDc domain , originally predicted to be catalytic [17] , may act as an antioxidant protein . Consequently , neurodegeneration in bel mice may be due to an increase in ROS that would normally be detoxified by Oxr1 . Importantly , however , the calculated rate constant for Oxr1 oxidation by H2O2 argues against a vital role for the protein as an antioxidant enzyme . Although the value of 0 . 82 M−1⋅s−1 for Oxr1 is of the same order of magnitude as reactive cysteine-mediated oxidation of BSA and DJ-1 [28] , [40] , it is thousands of times lower than key antioxidants such as catalase and peroxiredoxins that have reported oxidative rate constants of over 107 M−1⋅s−1 [29] , [41]–[42] . Therefore , although one attribute of Oxr1 may be to reduce ROS directly , it appears more likely that the oxidation of Oxr1 itself as a consequence of oxidative stress in the cell has a more important functional and/or regulatory role . Such redox-controlled modifications are important for a variety of proteins , including the regulation of conformational changes , often related to the formation or alteration of disulphide bonds [43] , [44] . For example , detailed studies of oxidised forms of DJ-1 have focussed on Cys106 as a key residue that mediates the function of the protein , although the mechanistic link between oxidation of this particular amino-acid and the multiple proposed roles for DJ-1 in vivo is still unclear [27] . It will therefore be important in the future to ascertain the relationship between our data regarding cysteine oxidation of Oxr1 with three-dimensional structural information of the TLDc domain; for instance , differences in the accessibility of Cys753 and Cys704 to peroxide may go some way to explain the results described here . A recent study , utilising Oxr1 knockdown in the mosquito A . gambiae , proposed that Oxr1 down-regulates the transcription of the antioxidants catalase and Gpx downstream of the stress-related Jun N-terminal kinase ( JNK ) [45] . Our transcriptional analysis of oxidative stress-related genes in the bel cerebellum also identified a significant reduction in Gpx1 expression , although this difference was not recapitulated at the protein level or using quantitative enzyme assays . This is most likely due to the small number of GCs affected in end-stage bel mice . As the study in A . gambiae was limited to transcriptional data , it would be interesting to examine whether the expression changes observed in mosquitoes equate to detectable alterations at the protein level . Although the interaction between Oxr1 and antioxidant enzymes is a plausible functional hypothesis , no mechanism for this particular pathway was investigated . Indeed , our biochemical data suggest that Oxr1 can react directly with ROS , although an indirect influence on other antioxidants , such as Gpx , cannot be ruled out . In the bel cerebellum , further evidence for oxidative stress was shown by the large increase in oxidative DNA damage as quantified by 8-OHdG immunostaining . The fact that few markers for the oxidative stress response were altered overall may simply reflect the limited lifespan of the bel mutant; only a small proportion of neurons undergo apoptosis before death ( approximately 1–2% of all GCs ) indicating the relative subtlety of the neuropathology . We can only speculate that if bel mice survived for longer whether additional regions of the CNS would be similarly affected . Future work using conditional or inducible disruption of Oxr1 will shed further light on such region-specific mechanisms . The bel mutant is an important new model of oxidative stress-related neurodegeneration , although the short lifespan of mutants limits the study of non-cerebellar neurons in vivo . Importantly , however , the fact that Oxr1 is expressed in all major regions of the brain and spinal cord , combined with our data from ALS and SOD1 mutant tissue , suggests that it plays a widespread and vital neuroprotective role . Indeed , it is intriguing that down-regulation of OXR1 has been recently reported as one of the major differences in a microarray study of the cortex in PD [46] . It is frequently postulated that stimulating endogenous defence pathways would be an effective strategy in combating cell death in disease [9]; our findings therefore provide the first indication that the enhancement of Oxr1 activity in vivo may counteract or even prevent the damage carried out by ROS in the progression of neurodegenerative disorders . The apparent functional compensation of OXR1 between yeast and human [13] and the high degree of sequence conservation at particular amino acid residues in the TLDc domain can now be investigated further to help decipher the molecular mechanisms involved . Indeed , it is noteworthy that the alanine mutated in the TLDc domain of the TBC1D24 protein in human FIME is not only conserved in OXR1 ( Figure S8B ) , but has also been shown to inhibit neurite outgrowth in vitro [16] , suggesting that further study into this family of proteins will also be important for neurological disorders outside of those directly linked with oxidative stress [47] .
All experiments were performed in accordance with the UK Home Office regulations and approved by the University of Oxford Ethical Review Panel . The bel phenotype was first identified from a screen for recessive ENU mutants at MRC Harwell , UK . To genetically map the trait , 13 bel mutants were initially screened for polymorphic SNP markers between the parental C3H/HeH and BALB/c ( ENU treated ) strains followed by fine mapping using additional microsatellite and SNP markers . Inverse PCR was carried out by digesting bel genomic DNA with a range of restriction enzymes , ligating the products , amplifying around the circular DNAs using nested primers and sequencing . A BglII restriction fragment spanning the deletion was consistently identified , which was confirmed using bel-specific PCR primers ( 5′ CGACTAGGCCATCTTCTATTAC and 5′ GCTAATGGCTGCCGAGTTTG ) . Mice were genotyped using these deletion primers in combination with wild-type control primer ( 5′ GTGACTGGAGGTGAGCTTTG ) or using D15Mit229 , a polymorphic microsatellite marker in very close proximity to the bel deletion . In situ hybridisation was carried out as previously described on 12 µM frozen tissue sections [48] . Regions of Oxr1 and Abra mouse cDNA sequences ( see Figure S7 ) were subcloned into pCR4-TOPO ( Invitrogen ) prior to DIG-labelled riboprobe synthesis and hybridisation . Slides were exposed for 16 hours in all cases . TUNEL staining for apoptotic cells was carried out on frozen sections using the in situ cell death kit ( Roche ) . Antibodies for cleaved caspase-3 ( Cell Signalling , 1∶500 dilution , 24 hours at 4°C ) and 8-OHdG ( QED Biosciences , 1∶250 ) immunostaining were used on 4% paraformaldehyde perfused , paraffin wax embedded sections; 8-OHdG staining was carried out as previously described [49] . Primary antibody staining was visualised using Vectastain Elite ABC kit ( Vectorlabs ) or Alexa Fluor 488 or 594 secondary antibodies ( Invitrogen ) for immunofluorescence . Five 10 µm sections taken at 40 µm intervals from the midline of 3 bel/+ and 3 bel/bel mice were stained with cresyl violet . The total area of each section corresponding to the cerebellum and the remainder of the brain was calculated using Axiovision 4 . 6 software ( Zeiss ) and averaged over each genotype . For GCL analysis , the midpoint of lobes III , IV/V and IX was determined as the distance between the apex to the abyss of the fissure . A region representing 0 . 4 mm , 0 . 2 mm either side of this midpoint , was used to determine the GCL width by dividing the area of the GCL in this region by 0 . 4 mm to obtain an average value for each lobe . To examine Purkinje cell numbers , adjacent sections , 5 from each animal , were immunostained using anti-calbindin 28 K ( Swant , 1∶15000 dilution , 48 hours at 4°C ) as previously described [48] . The total number of Purkinje cells on each section was counted and divided by the total length of the Purkinje cell layer . Adjacent sections to those above were used to count caspase-3 and 8-OHdG immunopositive cells in all cerebellar lobes . Quantification of apoptosis by TUNEL staining was carried out on five 10 µm midline sections at 40 µM intervals from 3 mice of each genotype . For muscle histopathology , tissue samples were dissected and snap frozen in OCT ( VWR ) on isopentane in liquid nitrogen . Frozen transverse sections were cut at 10 µM for haematoxylin and eosin ( H&E ) staining using standard methods . Counts of centrally nucleated fibres were averaged from H&E stained sections from 4 mice of each genotype . Culturing of granule and cortical cell cultures was carried out as previously described ( Amaxa Nucleofector protocol ( Lonza ) and Bilimoria et al . [50] . Bel mutant and control granule or cortical neurons were obtained from postnatal day 7 ( P7 ) or P2 animals , respectively , and cultured for 7 to 19 days prior to treatment . For cell death experiments , cells were treated with 150 µM H2O2 for 4 hours before being fixed in 4% paraformaldehyde for 15 minutes prior to using the TUNEL assay as above . Cells were assayed for survival by counting 1 , 500 cells for granule cells or 500 cells for cortical cells . For immunofluorescence , GCs and N2As were treated with 1 mM H2O2 for 30 minutes prior to recovery in fresh media for 1 hour . Cell counts were analyzed using Prism software; the difference between wild-type and mutant or between the various treatments was compared using ANOVA . P values<0 . 05 were considered significant . All experiments were carried out on 3 or more occasions with cultures obtained from independent mouse litters . For knockdown of Oxr1 expression , a Mission shRNA construct ( Sigma ) specific to all TLDc-containing isoforms of the gene ( see Figure S7 ) was used . Primary cells were electroporated with constructs using the Amaxa Nucleofector method ( Lonza ) . The relative level of knockdown was consistently over 90% as shown by qRT-PCR using Oxr1 exon-spanning primers ( Figure S4B ) . For over-expression of Oxr1 , the full-length mouse coding sequence ( NM_130885 ) with a C-terminal HA-tag was cloned into pLenti6/V5-D-TOPO vector ( Invitrogen ) with a stop codon introduced before the V5 sequence . The constructs were transfected into HEK293T cells with packaging vectors and virus-containing supernatants were collected 3 days later . GCs were infected after 11 days in culture by adding lentivirus-containing medium ( 1∶50 dilution ) and H2O2 treatment was carried out after 3 days of infection for 4 hours prior to cell survival estimation as above . Lentiviral Oxr1 expression equivalent to endogenous levels were consistently obtained as shown by qRT-PCR using Oxr1 primers as above ( Figure S4B ) . For over-expression studies comparing Oxr1-FL and Oxr1-C sequences , the coding regions ( NM_130885 and NM_001130164 , respectively ) were cloned into a bicistronic pCAGGS-based vector with additional internal ribosomal entry site ( IRES ) upstream of GFP . Primary cells were electroporated as above . Tissue or cell extracts were prepared using standard RIPA buffer and protein levels were quantified using BSA assays ( Pierce Thermo Scientific ) . After primary antibody ( Oxr1 1∶100 ( see above ) ; catalase ( Abcam ) ; Gpx1 ( Epitomics ) ; SOD1 ( Abcam ) ) and peroxidase-conjugated secondary antibody incubation , blots were developed with the ECL kit ( Amersham ) . Frozen thoracic spinal cord samples from non-SOD-related sporadic ALS patients and age-matched controls were obtained from the Thomas Willis Oxford Brain Collection . The lumbar enlargement of the spinal cord from 5-month old male SOD1 G93A mutants ( TgN[SOD1*G93A]Gur1 ) and littermate wild-type controls were dissected and protein extracts prepared immediately as above . Band intensity relative to internal controls was carried out using ImageJ software . Expression studies were carried out from total RNA purified using the RNeasy kit ( Qiagen ) . cDNA was generated using Expand Reverse Transcriptase ( Roche ) and triplicate qRT-PCR reactions carried out using SYBR green ( Applied Biosystems ) . Data were analysed using StepOne software ( Applied Biosystems ) and normalised to the control β-actin gene in all cases . All data shown are generated from at least 3 independent samples . Primer sequences are shown in Dataset S1 . The full-length mouse Oxr1 coding sequence was cloned into a pCAGGS-derived vector ( containing the chicken β-actin promoter with a CMV enhancer and a rabbit β-globin intron ) , freed of the plasmid backbone by restriction digest and injected into the pronuclei of superovulated CBAB6F1 mice . Founder mice were initially identified using Oxr1 exon-spanning primers for subsequent breeding . Two independent founder females ( Tg ( CAG-Oxr1 ) +/− ) were bred to heterozygous bel/+ males over two generations to generate mice homozygous for the bel deletion but also expressing the Oxr1 transgene ( bel/bel , Tg ( CAG-Oxr1 ) +/− ) to determine genetic rescue . Genetic background effects were controlled by assessing the onset of ataxia and neuropathology of non-transgenic bel/bel mutants which proved to be identical to the original bel line . Gene-trap ES cell line RRR195 ( Oxr1Gt ( RRR195 ) Byg ) was obtained from Bay Genomics and the correct identity of the insertion was confirmed by RT-PCR from cultured cells prior to rederivation . RRR195 ES cells were injected into preimplantation mouse embryos and chimeras were generated and bred with C57BL/6J mice . Chimera ES cell contribution and germline transmission were assessed by coat colour and confirmed by genotyping . The exact position of the insertion was determined by PCR to generate primers for genotyping; 5′ GTGTTGAGTTCCCCATC and 5′ CCGCAAACTCCTATTTCTGAG for the gene-trap vector or 5′ CAATCTAAATCCACTGCTGAC for the wild-type intron 3/4 control . Mice heterozygous for the insertion were bred together to generate homozygous animals . The full-length coding sequence of Oxr1-C ( NM_001130164 ) and a region representing the TLDc domain of mouse Oxr1 ( C7C , see Figure S7 ) were subcloned into the pET-22b ( + ) expression vector ( Novagen ) in-frame with a polyhistidine tag ( 6× His ) at the C-terminus . The coding sequence of mouse DJ-1 ( NM_020569 ) was cloned in the same manner . Oxr1 cysteine mutants were generated by QuikChange site-directed mutagenesis ( Stratagene ) and sequenced prior to use . Constructs were transformed into BL21 ( DE3 ) E . Coli cells ( Invitrogen ) and protein expression was induced overnight at 18°C at O . D600∼0 . 8 by addition of isopropyl-β-D-thiogalactopyranoside ( IPTG ) to a final concentration of 0 . 1 mM ( Oxr1 ) or 0 . 25 mM ( DJ-1 ) . Bacterial cultures were sonicated and recombinant His-tagged proteins were purified from the soluble fraction using BD Talon metal affinity resin ( BD Biosciences Clontech ) according to the manufacturer's recommendations . Antiserum was raised in rabbits against the C7C TLDc domain fusion protein ( Eurogentec ) and affinity purified . For protein oxidation studies , proteins were reduced in 2 mM DTT and subjected to buffer exchange into 50 mM phosphate buffer , pH 7 . 4 on PD-10 filtration columns ( GE Healthcare ) prior to use . The Amplex Red assay was used to determine the presence and/or depletion of H2O2 essentially as described by the manufacturer ( Molecular Probes ) . Working solutions of the dye and horseradish peroxidase ( HRP ) were made fresh for each assay and added to varying amounts of purified recombinant protein . Based on predicted molecular weight of Oxr1-C ( 28 . 55 kDa ) , protein concentrations ranged from 0 . 175 to 1 . 4 µM . The reaction was initiated with the addition of H2O2 at a final concentration of 1 . 4 µM ( e . g . ratio of H2O2 to Oxr1-C of up to 1∶1 ) . Samples were incubated for 30 minutes at 37°C in the dark and fluorescence readings were obtained at 580 nm . All wells were counted in triplicate correcting for background fluorescence from a blank sample and all experiments were repeated on 3 separate occasions . Pre-reduced recombinant Oxr1-C protein was incubated with 100 µM H2O2 ( final concentration ) at 37°C for up to 30 minutes . Individual aliquots ( corresponding to 5 µg of protein ) were taken at a range of timepoints , excess H2O2 was removed with purified catalase ( Sigma ) , and the amount of reactive thiols determined by incubation with 30 µM ThioGlo-1 for 90 minutes at 60°C . Samples were mixed with Laemmli loading dye ( without bromophenol blue to prevent background fluorescence ) and subjected to SDS-PAGE . Gels were visualised with an ultraviolet light source and subjected to densitometry using a Fluor-S MultiImager ( BioRad ) . To confirm equal loading of protein , gels were post-stained using SimplyBlue ( Invitrogen ) . The rate constant for this reaction was estimated by plotting ( 1/nHo−So ) ln[So ( Ho−S ) /Ho ( So−nS ) ] versus time as previously described [40] , where Ho represents the initial concentration of H2O2 , So is the initial concentration of free SH groups , S is the SH content reacted and n is the moles of free SH oxidised per mole of H2O2; n was taken to be a value of 2 given two potential reactive cysteines in the Oxr1-C recombinant protein fragment used . Catalase activity from cerebellar tissue samples was carried out using Amplex Red Catalase Assay Kit ( Molecular Probes ) and Gpx activity was measured using the Glutathione Peroxidase Assay Kit ( Calbiochem ) , both according to the manufacturer's instructions based on standard curves of enzyme activity . Genomic DNA from cells was extracted using the Genomic DNA Extraction Kit for tissues ( Qiagen ) . Double-stranded DNA was quantified using PicoGreen ( Thermo Scientific ) as per manufacturer's instructions and compared to a standard curve generated from λ/HindIII DNA to determine a ratio of dsDNA to ssDNA ( strand scission factor ) in the sample .
|
Oxygen is vital for life , but it can also cause damage to cells . Consequently , protective proteins ( antioxidants ) are utilised to maintain the fine balance between oxygen metabolism and the production of potentially toxic reactive oxygen species ( ROS ) . If this balance is not maintained , oxidative stress occurs and excess ROS are generated , causing damage to DNA , proteins , and lipids . The brain is particularly susceptible to oxidative stress , and ROS–induced damage is a common feature of all major neurodegenerative disorders , including amyotrophic lateral sclerosis ( ALS ) and Parkinson's disease ( PD ) . However , the molecular mechanisms of ROS defence in neurons are still under investigation . Here we describe the characterisation of oxidation resistance 1 ( Oxr1 ) , a gene previously shown to be induced under oxidative stress . We show both in mice and in cells that loss of Oxr1 causes cell death and that increasing protein levels can protect against ROS . In addition , Oxr1 is over-expressed in the spinal cord in ALS patients , as well as in a pre-symptomatic ALS mouse model . These data demonstrate for the first time that Oxr1 is vital for the protection of neuronal cells against oxidative stress and that induction of Oxr1 may be relevant to neurodegenerative pathways in disease .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"animal",
"models",
"medicine",
"neurobiology",
"of",
"disease",
"and",
"regeneration",
"model",
"organisms",
"neurological",
"disorders",
"neurology",
"neurodegenerative",
"diseases",
"biology",
"neuroscience",
"cerebellar",
"disorders"
] |
2011
|
Oxr1 Is Essential for Protection against Oxidative Stress-Induced Neurodegeneration
|
The dynamic nature of technological developments invites us to rethink the learning spaces . In this context , science education can be enriched by the contribution of new computational resources , making the educational process more up-to-date , challenging , and attractive . Bioinformatics is a key interdisciplinary field , contributing to the understanding of biological processes that is often underrated in secondary schools . As a useful resource in learning activities , bioinformatics could help in engaging students to integrate multiple fields of knowledge ( logical-mathematical , biological , computational , etc . ) and generate an enriched and long-lasting learning environment . Here , we report our recent project in which high school students learned basic concepts of programming applied to solving biological problems . The students were taught the Python syntax , and they coded simple tools to answer biological questions using resources at hand . Notably , these were built mostly on the students’ own smartphones , which proved to be capable , readily available , and relevant complementary tools for teaching . This project resulted in an empowering and inclusive experience that challenged differences in social background and technological accessibility .
Biology-oriented programming workshops using Python for students and teachers were performed in one private and two public schools from La Plata , the capital city of the province of Buenos Aires , Argentina . These schools were selected on the basis of their willingness to take part in the project and the possibility to accommodate the instructors’ schedule in their weekly activities . Participation of the students in the workshops was voluntary in all cases , although framed in the context of a particular science course like biology or genetics . Although representing schools with different educational goals and infrastructures , the three groups of students that took part comprised similar demographic profiles , with mixed socioeconomic backgrounds and a balanced gender ratio . The workshops were delivered as three weekly 90-minute-long face-to-face classes offered during consecutive weeks , with one teacher or teaching assistant per ten students . They took place in the schools with the technological resources available at each of them , using installed versions of Python 2 . 7 and/or 3 . 6 on PCs and students’ smartphones . Online Python terminals ( http://repl . it/languages/python3 , http://www . tutorialspoint . com/execute_python_online . php ) were also presented in order to show additional ways to use the language . Internet was only requested for the first meeting for Python installation . Overall , more than 90% of the students completed the practical exercises on their smartphones whereas the rest used netbooks or notebooks . The workshops were aimed at students of the last years of their secondary school ( a five/six-year-long stage equivalent to high school in the United States ) due to the science background needed to face the biological problems presented during the course . In spite of the public or private nature of the school , the curriculum design in the province of Buenos Aires establishes a common core in natural sciences and mathematics . Apart from this shared nucleus , the students can follow different orientations with additional workload in distinctive subjects ( a detailed description of the curriculum , in Spanish , is included in http://servicios . abc . gov . ar/lainstitucion/organismos/consejogeneral/disenioscurriculares/ ) . In particular , the different activities proposed during the workshops require basic operations in mathematics , understanding of logical operators , and knowledge of the classical perspective on the molecular basis of information flow from DNA to proteins . Because these contents are included in the shared nucleus of all orientations , students from both years in each school joined a unique , integrated class for the workshop . The contents covered in each class are shown in Table 1 . An interactive guide was given to students ( see S1 File ) in which many exercises were proposed . Some of them were taken as examples to solve during the workshop by the students with teachers’ assistance . Possible solutions were shared and evaluated collectively in order to take the maximum advantage of every different proposal . After finishing the workshops , the participants were offered to take part in a bioinformatics challenge that was set up as a contest . Each school could present multiple groups of up to five students accompanied by a teacher . Three problems ( see S2 File ) were given to the students to be solved in a three-week period , during which the groups were monitored by teachers and workshop trainers . The exercises were written with increasing complexity , and each had extra goals to tackle in order to encourage a deeper analysis for working solutions . For example , the first question asked the participants to construct an algorithm for translating a hidden message between nucleotide and amino acid alphabets using the standard genetic code , with additional points awarded for showing the number and identity of codon sequences that could encode the message . Each group delivered their scripts and a written report detailing the general approach they applied , the difficulties they faced , and the major decisions they took toward their goal . Submissions were evaluated by an ad hoc committee . They provided feedback on early versions of the work and ranked the final submissions by testing that the programs worked as intended and evaluating the extra effort put into solving the optional exercises and the attention to documentation , presentation , and general style of code . All the examples of the final scripts built by the students answering the required questions are shown in S2 File . There are many smartphone applications ( apps ) available for the different operating systems ( OSs ) , which may be more or less useful for working in the classroom , depending on the type of tools to create . When developing simple tools that could be run from the interactive interpreter or by loading single scripts , and that only require standard libraries , distributions that offer a Python terminal are sufficient and recommended . There are many Python apps available for Android and iOS , both free and paid , but fewer options can be used in Windows smartphones and are generally not optimal for running external scripts . In Table 2 , we summarize some useful free apps for the classroom , among which QPython and QPython 3 for Android , Python 3 for Windows , and Python 2 . 5 for iOS were recommended to the students because these proved to be stable and responsive based in our preliminary evaluation in several smartphone platforms . These Python applications allowed students to test the code proposed in class quickly and easily , making the overall experience less passive and noticeably more engaging . Other more comprehensive apps may be needed in complex scenarios , especially if there is the need to load big external data files or use third-party libraries with multiple dependencies . For a preliminary evaluation of different standard smartphone platforms , we tested several Python apps using a simple script and recording its calculation times . The script we implemented ( see S3 File ) is a “Translator” that receives a phrase in “human language” and translate it into “cells language . ” Using the universal genetic code and the standard one-letter amino acid representation , most letters from the English alphabet could be written as one or more codons . Any word using these letters can therefore be translated to a large number of codon combinations . This idea was later explained and proposed to the students as an exercise too . The main purpose of this evaluation was to compare calculation times between PCs , online resources , and mobile smartphones that are available for students to perform bioinformatics calculations in a classroom . Our intention is not to benchmark smartphones performances , which depends on too many variables that could not be addressed here , but to assess whether those smartphones that are commonly accessible to students in our local communities would be able to complete the proposed tasks efficiently . From Fig 1 , it is possible to infer that smartphones are on average slower than PCs for calculation times . Online resources such as Repl . it ( https://repl . it/ ) can perform somewhere between PC and mobile phones , although permanent access to the internet should be provided . It is also possible to see in Fig 1 that some smartphones could perform even better than current personal computers . This does not seem to be highly dependent on the OS of the phone but on the available processor speed ( Fig 2 ) and RAM memory ( Fig 3 ) , with other factors such as the Python interface used and the system load possibly affecting the running times . According to our results , smartphones were at most an order slower than a typical desktop PC hardware setup ( Intel i5-6400 2 . 7GHz quad-core processor with 8Gb RAM ) , proving very capable of serving as programming tools for this kind of course .
A total of 100 students aged between 16 and 19 years old were part of the project , all of them owning a smartphone; 92 . 9% of the students didn’t have previous programming knowledge , and most of them ( 87 . 5% ) did not know about bioinformatics . More than half ( 53 . 57% ) of the participant students were from the natural sciences orientation . Students enrolled in the social sciences ( 35 . 71% ) and economy ( 10 . 72% ) orientations also took part in the workshop , showing that the proposal of learning a programming language was transversal and attractive for the students in general . The tools chosen for the workshops ( smartphones apps and online terminals of Python ) made the teaching–learning process , as well as the exchange of knowledge among students , engaging and effective . As derived from student’s feedback , the exchange of ideas was fluid and the immediacy provided by these technological devices allowed students and teachers to explore different variants for the proposed exercises , generally derived from questions raised in the classroom , making it possible to evaluate several possible paths toward a solution . The biological questions proposed and solved in the workshops ( see Table 3 ) triggered challenges in programming and enriched the overall learning–teaching process . Our results show that the use of smartphones could help to surmount the limitations related with the availability of computers in high schools . The easy setup of this kind of workshop , based almost entirely in smartphones and thus independent of the available equipment in schools , triggered a great interest of the educational community and generated enthusiastic responses in students . Although it is yet not possible to collect enough evidence to address the impact of our workshops , this novel approach should let students deepen their knowledge and interest in the field by revisiting biological concepts under a new light . The workshop should have also helped students to realize the potential of acquiring programming skills , giving them a tool not only for understanding and experiencing science , but also for developing strategies to help solve different challenges of their future professional life . Altogether , we think that these practices reinforce the notion that bioinformatics provides a suitable framework to improve the learning-teaching experience of biology and programming .
|
Bioinformatics is an interdisciplinary activity that complements and connects several fields with biology and can also be used as an educational tool for science . During 2017 , the Structural Bioinformatics Group at National University of Quilmes in Buenos Aires , Argentina , worked together with public and private schools to promote the usage of bioinformatics towards a better understanding of biology . We performed short biology-oriented programming workshops using Python , aimed at students in different schools , who were later invited to participate in a specially organized and challenging bioinformatics contest ( http://ufq . unq . edu . ar/sbg/education/index . html ) . The choice of computational tools , with a major role of smartphone applications , made the teaching–learning process easier , dynamic , and accessible . This experience allowed us to build bridges with the participating schools and develop a great commitment toward expanding the project in the near future . The great interest shown by educational communities and the positive responses of students reinforce the idea of bioinformatics as a plausible tool for the learning–teaching of biology .
|
[
"Abstract",
"Background",
"and",
"motivation",
"Discussion"
] |
[
"learning",
"education",
"engineering",
"and",
"technology",
"sociology",
"social",
"sciences",
"teachers",
"neuroscience",
"learning",
"and",
"memory",
"cell",
"phones",
"computers",
"cognitive",
"psychology",
"research",
"and",
"analysis",
"methods",
"computer",
"and",
"information",
"sciences",
"bioinformatics",
"human",
"learning",
"schools",
"communication",
"equipment",
"people",
"and",
"places",
"professions",
"psychology",
"programming",
"languages",
"equipment",
"database",
"and",
"informatics",
"methods",
"biology",
"and",
"life",
"sciences",
"population",
"groupings",
"cognitive",
"science",
"workshops"
] |
2019
|
Bioinformatics calls the school: Use of smartphones to introduce Python for bioinformatics in high schools
|
The Ebola virus disease epidemic between 2013 and 2016 in West Africa was unprecedented . It resulted in approximately 28 . 000 cases and 10 . 000 Ebola survivors . Many survivors face social , economic and health-related predicaments and media reporting is crucially important in infectious disease outbreaks . However , there is little research on reporting of the social situation of Ebola survivors in Liberia . The study used a mixed methods approach and analysed media reports from the Liberian Daily Observer ( DOL ) , a daily newspaper available online in English . We were interested to know how the situation of Ebola survivors was portrayed; in what way issues such as stigma and discrimination were addressed; and which stigma reduction interventions were covered and how . We included all articles on the situation of Ebola survivors in the quantitative and in-depth qualitative analysis published between April 2014 and March 2016 . The DOL published 148 articles that portrayed the social situation of Ebola survivors between the 24 months observation period . In these articles , Ebola survivors were often defined beyond biological terms , reflecting on a broader social definition of survivorship . Survivorship was associated with challenges such as suffering from after-effects , social and economic consequences and psychological distress . Almost 50% of the articles explicitly mentioned stigmatisation in their reporting on Ebola survivors . This was contextualised in untrustworthiness towards international responses and the local health care system and inconclusive knowledge on cures and transmission routes . In the majority of DOL articles stigma reduction and engaging survivors in the response was reported as crucially important . Reporting in the DOL was educational-didactical and well-balanced in terms of disseminating available medical knowledge and reflecting the social situation of Ebola survivors . While the articles contextualised factors contributing to stigmatisation throughout the reporting , journalistic scrutiny regarding effectiveness of interventions by government and NGOs was missing .
The Ebola virus disease epidemic in West Africa 2013–2016 was unprecedented and resulted in over 28 . 000 cases . Up to 10 . 000 persons are estimated to have survived the disease [1] . In Liberia , there are over 1500 survivors registered in the database of the Liberian Ministry of Health [2] but considering the high overall WHO estimates it is assumed that the actual number of survivors is significantly higher than the ones registered . During the height of the epidemic an important component of a comprehensive disease response was to include survivors in the treatment and care of EVD patients , in training of health workers , as well as in health sensitizing and social mobilization efforts [3–5] . However , the reintegration of survivors into their communities proved to be challenging . While at the beginning of the epidemic survivors were often welcomed by their communities [6] , EVD-related stigma that led to social isolation of survivors in families and communities , physical violence and loss of jobs has been reported from the affected countries [3 , 7–9] . Stigma affects the social lives of individuals and communities , leading to suffering and loss of important networks . Stigma and social distancing is also associated with avoidance of seeking health care [10 , 11] which in the case of Ebola virus disease is important regarding isolation and treatment of sick persons , and quarantine of contacts . Stigmatization of survivors in the health system might also lead to neglect or survivors not disclosing their status . Survivors have lost their possessions to prevent disease transmission and they often suffered income loss . Moreover , many are suffering from physical after-effects of the disease [12] , as well as mental and psychological consequences [8 , 13–16] . Considering the high numbers of survivors in this latest epidemic there is little research on the social situation of survivors in the most affected countries [17] . In this study we therefore wanted to analyse how one of Liberia’s largest newspapers , the Liberian Observer , portrayed and informed the public on the situation of Ebola survivors 2014–2016 . We were interested in who is considered a survivor , the challenges survivors face in day-to-day life , and how survival was explained and disease transmission linked to survivors . Furthermore , we were interested how the discourse on stigma and discrimination was conveyed to the readership . Stigmatization of survivors has recently received heightened attention through the tragic and unnecessary death of a female survivor who was not treated after complications in childbirth [18] . Moreover , we wanted to understand how these themes changed over time–from the beginning of the epidemic in March 2014 , through the peak in December 2014 , to its consolidation in 2015 and up to March 2016 when the epidemic waned . There are ongoing discussions over who is to be considered a survivor of EVD . This question is important because 1 ) survivors are eligible for social and medical support , 2 ) survivors are often part of clinical studies , and 3 ) the identity of survivorship is important when discussing the social implications of the epidemic . One of the first agreed upon biological definitions of survivorship in Liberia was discussed at a meeting of the Liberian Ministry of Health and Social Welfare . The group came to the consensus that a survivor should be defined as “A person who tested positive for the Ebola virus disease and after receiving care and treatment , recovered and tested negative”[19] . However , one of the problems is the often poor record keeping in facilities that makes certification of survivors difficult [19 , 20] . Another question was what should be done with persons who were sick and survived but were not treated in a facility and thus not tested ? Furthermore , towards the end of the epidemic it became clear , that minimally symptomatic Ebola virus infections were likely to have taken place [21] and the question arose “Should the notion of survivorship be extended to all those who are IgG positive , including those who had minimally symptomatic infection or who were sick but were never tested at the time of illness ? ” [21] . Thus WHO [1] defines a survivor as a person “With a confirmed positive result on RT-PCR testing for Ebola virus on any body fluid who subsequently recovered; and/or who is IgM and/or IgG positive on serological testing for EVD and has not been vaccinated against Ebola virus . ” This definition , however , does not answer the social and economic implications of survivorship . This study contributes to the research field of media reporting in public health emergencies . Media reporting is crucially important in infectious disease outbreaks . One study , for instance , found an association between media reports of interventions that provided education and reductions in transmissions of Ebola in Sierra Leone and Liberia [22] . Real-time information from trustworthy news reports have been used to characterize epidemiological patterns [23] . Media can be equally important allies for raising awareness , inducing positive healthy behaviour practice [24] and disseminating vital information in emergencies . Media reporting gives visibility to what is seen as important , it frames risk perceptions and allows reflection of public health policy such as promoting necessary intervention strategies . It can also be useful in promoting behaviour change and mitigating stigmatization [25–27] . However , any form of communication in public health emergencies has to consider people’s priorities as well as social , political , economic and historic contexts [28 , 29] . Several studies have dealt with the role of media in the Ebola epidemic in West Africa , most of them focusing on the situation in the global north , especially the U . S . [30–35] . Only few studies investigate media reporting on Ebola in the affected countries themselves . Two studies looked at how the Ebola crisis was framed by newspapers from the US and Canada , compared to an “insider” perspective of an Ebola affected country ( Sierra Leone and Nigeria ) [36 , 37]; one assessed the role of four Nigerian newspapers in creating awareness to stop the spread of Ebola [38] . To our knowledge , however , there is no study on media reporting in the Ebola epidemic in Liberia and no research so far has been published on the portrayal of Ebola survivors in the media during the epidemic in West Africa . The main research questions are therefore: How is the situation of Ebola survivors portrayed in one of Liberia’s largest newspapers ? In what way are issues such as stigma and discrimination against Ebola survivors addressed and what factors are identified to contribute to it ? Which stigma reduction interventions are covered and how ? According to Peacebuilding-Data newspapers are consumed at least occasionally by one third of the population in Liberia ( 29% ) . The most commonly read newspaper is the Daily Observer ( 21% ) , followed by the Inquirer ( 11% ) and the New Democrat ( 10% ) [39] . The Daily Observer is available in English which is the country’s only official language . Besides varieties of Liberian English , there are over 30 other languages spoken in the country . A total of 47 , 6% of the population is literate [40] and it can be assumed that the Daily Observer addresses a distinct , educated readership rather than the general population .
This study used a mixed method approach , whereby the Liberian Daily Observer ( DOL ) a daily newspaper released in English was analysed . It is available online ( www . liberianobserver . com ) and is reportedly the largest Liberian newspaper [41] . All articles that were published in the DOL on the situation of Ebola survivors between April 2014 and March 2016 were collected through the archive and search function . We used the following simple search terms: “Ebola survivors” , “Ebola” , “survivors” and “survive” , screened the articles applying inclusion and exclusion criteria . We performed a basic quantitative content analysis [42] and an in-depth qualitative analysis [43 , 44] , guided by the research question . Using the above mentioned search terms EM initially collected 630 articles and screened the articles for abstract and titles . Articles from all rubrics were principally included . The article had to report on any aspect of Ebola survivorship–challenges faced , after effects , and survivors as factors of EVD transmission . Articles were excluded if they only reported about survivors ( of other disease ) but not in regard to Ebola , if they only mentioned Ebola survivors as receiver group of donations but lacked any further or contextual information , or if they had fictional character such as short stories or poetry . The inclusion and exclusion criteria were firstly discussed in the research team ( EM , RK , TN ) and consecutively applied by EM resulting in 148 final articles for analysis . The selected articles were imported into the Qualitative Data Analysis Software atlas . ti in word or pdf format and were analysed [45] . Two of the authors ( EM , RK ) developed the coding framework . First , all collected articles were reviewed by EM and main themes were identified following an open inductive coding approach [45] . In discussion with RK who has worked with Ebola survivors in Liberia , EM compared these main themes , further categorized and supplemented them with sub-themes . The data was then quantitatively analysed; in a first step the frequency of publication of articles per month was recorded , and descriptive information such as date of publication and type of article was registered ( EM ) [42] . This was followed by an in-depth qualitative analysis to investigate which themes were addressed and how the situation of Ebola survivors was portrayed in the newspaper [43] . In accordance to the main research questions , we looked at how survivors were defined , what and how challenges of living as survivor were described , and how the themes stigma and discrimination were presented . Related themes to codes that emerged in open coding through re-reviewing the articles were noted and adapted with deductive categories that were discussed among all three authors after one third of the articles were coded in atlas . ti . The following codes were used: personal story/situation and release , health care workers , after-effects , how/why survived , transmission through survivors , stigma explicit , stigma implicit , valuable involvement of survivors , support received , identity , what is lost ( for an overview of codes see Table 1 ) . We assessed inter-coder reliability , or more specifically inter-coder agreement , using Cohen’s kappa technique [46 , 47] . RK coded twenty articles independently ( 14% of articles chosen randomly from the sample ) and agreement between 0 . 67 and 1 . 00 was reached between the two coders EM ( coder 1 ) and RK ( coder 2 ) , see Table 1 . Codes not applied to the 20 articles are not appearing in the table . In accordance to Landis and Koch ( 1977 ) strength of agreement division for Kappa Statistics ( <0 . 00 Poor; 0 . 00–0 . 20 Slight; 0 . 21–0 . 40 Fair; 0 . 41–0 . 60 Moderate; 0 . 61–0 . 80 Substantial; 0 . 81-1-00 Almost Perfect ) it lies between substantial and almost perfect [48] . This media analysis was conducted with already published newspaper articles; therefore , no ethical approval was necessary .
The DOL referred to two groups as survivors: persons who were confirmed Ebola positive cases who later tested negative; and persons who were not infected themselves but were closely associated to the disease , e . g . who lost partners and family members , orphans , persons discharged from ETUs without being infected or persons who “survived” quarantine ( A1-A64 , B1-B77 , C1-C7 ) . The representation of survivors used by the DOL is assumed to have originated from how people in Liberia were using the term Ebola survivor . Some articles also referred to survivors as “Ebola victims” ( A6 , A12 , A19 , A50 , A52 , B8 , B22 , B31 , B51 , B63 , B64 , C3 ) or “viral victims” ( B3 ) . Campaigns against stigmatization often targeted “Ebola victims & survivors” or “Ebola affected children/persons and survivors” , highlighting that psychosocial support was pivotal for survivors themselves but also for their families , friends and communities ( e . g . A19 , A27 , A45 , A46 , A47 , A52 , A55 , B1 , B3 , B8 , B9 , B11 , B39 , B41 , B63 , B65 , B66 ) . The complex representation of who is a survivor is linked to the fact that both survivors and victims are rather classified by their marginalization and affectedness by the disease than by biological survival alone . Survivorship per se was associated with support by various programs . However , this identity might change in the course of time , as one business woman who offered training to survivors is cited: While being called an “Ebola survivor” by others is often felt stigmatizing , the term “survivor” might turn into a positive identity if used by the ones affected . The denomination of survivor becomes also important when highlighting the fact that survivors are immune to the disease . This was used in several awareness campaigns: Over the course of the epidemic the DOL increasingly cited Ebola survivors as heroes and heroines . Already in November 2014 two discharged health care workers were referred to as “healthcare heroes” ( A49 ) and also later discharged health care workers were described as the “true heroes and heroines” ( B17 ) , or heroic icons who cared for the infected ( B6 , C1 ) . General views that survivors should be viewed as heroes are also found in some of the articles ( B3 , B24 , B46 ) . The country director of an NGO for instance explained: “[…] that survivors , mainly orphans , should be taken care of and embraced in the society because they are heroes of the Ebola crisis” ( A55 ) . The DOL reports that Ebola survivors face a multitude of health related , economic and social difficulties and were furthermore subject to discrimination and stigma during and in the aftermath of the epidemic . Eighteen articles ( 12% ) address the issue of health related after-effects of Ebola survivors which included visual problems or blindness , headaches , hearing impairment , joint- , muscle- and back pain , weakness , breathlessness , diarrhoea , abdominal pain , changes in the immune system , fever , swollen legs , loss of appetite , impotency or uncontrollable erection for men and failure to conceive or get pregnant by women ( A14 , A24 , A59 , B3 , B34 , B49 , B60 , B65 , C2 , C4 ) . The articles also report on neurological symptoms , psychological distress and mental health issues ( A64 , B9 , B34 , B60 , B61 , B76 ) . The reporting of sequelae increased in frequency when the PREVAIL Study lll , which is a Liberia-U . S . clinical research partnership on Ebola survivors to study long-term health effects of EVD , was launched and recruitment of study participants apparently started . One survivor explained her feeling of overall powerlessness: Another study participant describes his psychological problems: Social and economic consequences of the disease were portrayed in the DOL especially with regards to survivors’ experiences of discrimination and stigmatization . In this context the problematic situation of orphans who survived was also highlighted in several articles ( A27 , A52 , A53 , A55 , A56 , A61 , A64 , B8 , B11 , B15 , B31 , B34 , B37 , B39 , B41 , B50 , B51 , C5 ) . The overall situation of survivors is portrayed as severely distressing; several articles refer to the loss of children and family members due to the disease ( A7 , A26 , A27 , A30 , A31 , A43 , A44 , A46 , A47 , A48 , A50 , A51 , A53 , A56 , A64 , B7 , B13 , B15 , B22 , B34 , B37 , B46 , B60 , B68 , B72 ) , the loss of hope ( A3 , A28 , B68 ) and a situation of despair . One survivor explains: Loss also included “capacity to sustain herself […] and thinking about nothing except death” ( A38 ) and lost sources of income ( A44 , A53 ) “[…] and every other hope of livelihood has been lost” ( A27 ) . A female survivor describes the difficulty to sustain normal life: Also in terms of material losses DOL articles reported that household items , beddings and other belongings were burned: Almost fifty percent of the articles ( 74 articles ) explicitly mention “stigmatization” or “stigma” in their reporting on Ebola survivors . While in 2014 only 34% articles mention stigma or stigmatization the number increases in 2015 to 64% and in 2016 to 43% . Of those 74 articles which explicitly mention stigma or stigmatization in their reporting on Ebola survivors , 50% ( 37 articles ) were published in the news section , 30% ( 22 articles ) in the column section ( including education , health , women and youth ) , 11% ( 8 articles ) in the sports section , three articles mentioning stigma or stigmatization occurred in the business section and two articles respectively as opinion commentary or could not be assigned . The articles often describe that stigmatization resulting in discrimination was experienced as a community attitude and resentment and ostracism was often expressed by shunning and avoiding survivors or other persons associated with the disease ( A25 , B32 , B77 ) . Articles also describe survivors in a situation where they experience isolation ( A49 ) and where they have nowhere to go ( A64 ) and desperation on their situation was linked to mental health issues ( see above ) . Other acts of discrimination referred to in the DOL spanned from evictions from home , joblessness , separation from community and family members to acts of violence , such as forced prohibition to use community wells ( A51 ) . Some articles reported on stigma leading to discrimination at the work place , which had devastating economic effects and often result in aid dependency ( A51 , B34 , B44 , C5 ) . Another aspect reported on was that many Liberians refuse Ebola survivors as tenants ( A47 , A49 , A59 , C5 ) , sometimes leading to eviction and homelessness ( A59 , B12 ) . It was reported that survivors also had problems buying food because marketers refused to accept money from them ( A31 , B12 , B41 ) . Furthermore , DOL articles point to stigma and fear of patients who recovered from Ebola when seeking health care in non-Ebola healthcare structures ( B40 , B44 ) , which is specifically problematic in the face of the after-effects survivors often experience . One article reports that female survivors suffer the burden of social stigmatization more than man due to their traditional role as care givers ( B50 ) . Several articles stressed the problem of failing communication between sick relatives that were treated in Ebola Treatment Units ( ETUs ) and their families and communities , as factors reinforcing fear . It was reported that for a number of days , there was no communication with patients; from a patient’s wife perspective an article narrates: “no one on the outside was telling her anything about his wellbeing even if she asked” ( A5 ) . Explanations for stigmatizing behaviour were also viewed in the context of a general untrustworthiness towards the international health response and their local establishments , the health care system as well as estimated casualty numbers , noted in some articles ( A8 , A29 , A32 , A44 , A50 , A62 ) . Many articles identified the fear of contracting the disease as the main underlying reason for any form of stigmatization and/or discrimination . The DOL cited a WHO employee , explaining: Other articles also reported fear of contracting the disease as a contributing factor for rejection , stigma and discrimination of survivors and their families by their relatives or communities ( A51 , A63 ) . In this context , several articles cite explanations for survival , e . g . early treatment and special care ( A6 , A9 , A11 , A12 , A15 , A23 , A44 , A54 ) , other articles attribute survival to God and miracles ( A2 , A3 , A5 , A11 , A17 , A19 , A28 , A31 , A38 , A55 ) . From the beginning of 2015 onwards no direct explanations for survival were portrayed in the newspaper articles any longer ( B1-B77 , C1-C7 ) . Many of the DOL articles point to fear of transmission in the view of inconclusive knowledge or rather absence of cures . After May 2015 the DOL material did not again point to the fact that there is “no cure” or secure healing method for Ebola . Yet , another major theme that emerged from the articles was the uncertainty about possible transmission routes . Already in one of the first selected article from April 2014 the Ministry of Health was cited to encourage the public to take the following precautionary measures: Until the end of the observation period the second and fourth point became central in the debates and reports around transmission routes . Transmission became increasingly linked to biological survivors . While in 2014 only six articles reported a possible transmission through survivors ( A11 , A35 , A36 , A42 , A45 , A58 ) , the frequency increased in 2015 where 18 articles linked transmission to survivors ( B26 , B28 , B29 , B30 , B33 , B35 , B36 , B38 , B43 , B45 , B54 , B57 , B62 , B69 , B70 , B73 , B74 , B75 ) and in 2016 with two articles ( C1 , C7 ) . In February 2015 an article still stated that: In March 25th 2015 the first EVD confirmed case was reported where transmission routes initially remained undetermined but in which a survivor was mentioned as a possible source of infection: The patient passed away on March 27th and three days afterwards on March 30th 2015 the headline read: “‘Sexual Transmission’ May Have Infected Ebola Patient” and the article reports: As visualised in Fig 2 , on May 9th 2015 the original Ebola outbreak was declared over and Liberia was first declared free of Ebola transmission . On June 30th 2015 the disease re-emerged , one case was tested positive and succumbed to the disease . On September 3rd 2015 the WHO declared Liberia Ebola free for the second time . Eleven weeks later on November 19th 2015 once again re-infection was reported , with a number of two cases . The 3rd declaration of the end of Ebola transmission was announced on January 14th 2016 . New flare-ups were reported starting March 16th 2016 , where 8 people died and two survived . Until on June 9th 2016 the end of the EVD outbreak in Liberia was finally declared over [49] . The uncertainty of sexually transmitted flare ups through survivors is reflected in several articles . In November 2015 DOL relates that the transmission routes were obscure: The DOL material shows the uncertainty of sexual transmission of Ebola regarding the perceived time span a survivor is still infectious , which is described as between “two to three months” ( P70 ) , “the virus is still active in male survivors for 82 days” ( A42 ) , to reporting a CDC statement: “contact with semen from male survivors should be avoided ! Until research finds out how long the Ebola virus remains in the semen” ( B43 ) . Similarly to the latter precautious measures the head of the Incident Management System stressed: Up to the end of the observation period it remains unclear how large the infection potential in male survivors is and it is reported that: The DOL also reported on stigma reduction measures and general support initiatives for survivors . It was emphasized that in order to minimize stigmatization a reintegration process was particularly important for individuals following isolation or treatment for Ebola ( A9 ) . Especially in articles from the year 2014 reintegration , reunion and reunification ceremonies and efforts were highlighted ( A1 , A2 , A2 , A9 , A18 , A12 , A20 , A30 , A45 , A51 , A59 ) . Later in the epidemic we found fewer reports on reintegration , or the importance of the process , in 2015 there were four articles reporting on that issue ( B4 , B18 , B68 , B76 ) . In an article from 15th January 2016 titled “Liberia , Guinea , S/Leone All Ebola-Free” a WHO spokesmen is quoted outlining a comprehensive approach on survivors required: Specific slogans targeting survivors were promoted in the Ebola response and reports about campaigns , workshops and billboard ads reoccurred in the DOL articles . These slogans often picked up the issue of stigmatization and linguistic demarcations: Furthermore , also different representatives and authorities were cited: “to ensure that […] survivors of the deadly virus are not stigmatized […] cautioned the public against stigmatizing survivors” ( A37 ) and calling upon communities to show tolerance and dignity to Ebola survivors as well as health care workers treating former Ebola patients and members of burial teams ( A39 ) : Identification with Ebola survivors through media was also noted in the DOL , e . g . churches ( A19 ) , organizations , football associations and individuals announced in the articles that they identified with Ebola survivors ( A21 , A46 , A53 , A60 , A61 , B3 ) . Others also identified with the overall struggle to defeat Ebola and its stigma on survivors ( B2 ) and articles demanded to: The role of news coverage in stigma reduction is already brought up in August 2014 , in an article where the head of the Liberian Business Association was cited , who: Another stigma reduction measure specified in the DOL was the adoption of the “Health Workers , Infants and Epidemic Survivors Protection Bill” adopted on November 4th 2014 , which provides for compensation schemes to epidemic survivors and insurance incentives for health workers ( A22 ) . In how far the bill was enforced does not become evident through the articles . The DOL reported on the one hand , on general support measures; eighty articles ( 54% ) reported on support ranging from a variety of material donations ( including food , beddings , etc . ) and solidarity packages to radios and solar lanterns ( B11 ) , dignity kits ( B13 ) to direct cash dispensing ( B5 , B70 ) . The latter was not presented without reservations: On the other hand , several of the articles pointed to supporting networks and associations which were established and which organized workshops , counselling , psychosocial support ( B8 , B9 , B11 , B28 , B42 , B75 , C3 ) , capacity building management seminars ( C3 ) and e . g . counselling manuals “to help people live , love and laugh again after the virus” ( B75 ) . Approaches to counteract factors contributing to stigma were also portrayed as reducing stigma; thereby we found that awareness-raising about transmission , protection and chances of recovery was most prominently put forward ( A9 , A10 , A15 , A36 , A48 , A51 , A52 , A63 ) in the battle against stigmatization ( A63 ) . Some articles cited the importance of Ebola survivor involvement in the fight against the epidemic . Involvement as in donating blood ( A25 , A28 ) , inclusion in the National Ebola Taskforce , in the recovery plan or as psycho-social counsellors ( A10 , B84 ) or through telling their stories , sharing their experience ( A37 , A63 , B26 , B35 ) and helping to fight the epidemic ( A2 , A6 , A46 ) . It is argued that involving survivors may or will lead to de-stigmatization also in terms of countering factors which contribute to stigma: The necessity of involvement was voiced both , from the DOL article author and/or from survivors who expressed the wish to be involved in the fight of the epidemic . The total number of articles lacked a critical appraisal on success of the support and stigma reduction measures , furthermore we could not find reflection on evidence based measures or evaluation results of measures in the articles ( A1-A46 , B1-B76 , C1-C7 ) .
We observed that the Liberian Observer defined survivors beyond biological terms–something that might reflect the actual perception in Liberian society which is expressed by the often cited statement “we are all survivors” . Broadly speaking , we identified two ways of defining a survivor in the media portrayal analyzed–in social and biological terms . The DOL reflects a broader social definition of survivorship and argues that persons who were not tested positive but „survived”the ETU or quarantine measures , or persons who lost a loved one are to be considered survivors . This reflects day-to-day experiences of affected people , families and communities . Being a survivor as portrayed in the DOL is often associated with being a victim . The identity as victims and the broad concept of survivorship , however , do not go uncontested . Some articles call for an empowerment of survivors . Others debate its broad definition because being a biological survivor with a certificate from an ETU entitles a person to support through NGOs and free access to health care . The narrative that survivors should be considered heroes or at least the “social acknowledgment of their status” [19] was supported by the government and several NGOs as an effort to de-stigmatization and was reflected in several campaigns . The heroic narrative is also explicitly used in the DOL reporting , especially to refer to local health care workers , who were particularly affected during and after the epidemic . It can be assumed that an appreciative and empowering definition and attitude towards survivors in the reporting aimed at reaching this group and advocating for them . The situation of survivors in the DOL was portrayed comprehensively and included physical and mental problems survivors faced , as well as social and economic challenges rarely studied systematically [17 , 51–53] . The DOL reports on various different dimensions of survivorship and consequences of having lived through the disease or being highly affected by the epidemic . Almost in an educational-didactical approach the newspaper explains how realities of survivors are experienced by quoting their stories and providing context to their views and needs . Health related after-effects of Ebola virus disease are well documented in scientific literature [12 , 54] . However , the social and economic effects of survivorship have rarely been studied and survivors advocate for the improvement of their situation . In the reporting of the DOL the topics stigma and discrimination emerge as very important topics , they increase in frequency over the time period studied . We found that the DOL takes a clearly supportive stance towards Ebola survivors and takes their sides . Survivors are given a strong voice and representation in reporting on stigma and how discrimination is experienced and the DOL takes a very clear educational role towards the general public and its readership . This finding was actually surprising , as this does not seem to reflect what ordinary people believe . The reality on the ground suggests that stigma and discrimination against survivors is taking place widely [19] and that the DOL seems to act as a corrective to what the larger public believes . As the Liberian Observer online newspaper is only available in English it addresses a distinct readership . Based on the latest published data the illiteracy rates in Liberia are at 52 , 4% [40] which indicates that the readership might not being representative of the total Liberian population . The uniformity of survivors’ support is unusual and clearly different to media reporting on other victims , i . e . suicidal individuals who are often actively stigmatized in media and in the general public [55] . The DOL gave a broad overview what contributed to stigma and discrimination of survivors , e . g . failing communication , fear of contracting the disease in light of uncertain knowledge on transmission routes , absence of a cure and high case fatality rates . During the reporting period transmission became increasingly linked to survivors but the uncertainty of medical knowledge was actually admitted in the reports , yet of course it remained a major challenge in the hazardous situation . To reduce stigma , support and reintegration of survivors was often propagated and reported in the DOL . It could accordingly be shown in the research literature that the involvement of communities and families helped the reintegration process of survivors which again reduced stigmatization [3] . The very fact of talking about stigmatization and de-stigmatization campaigns and increasing the visibility of survivors is important and considered to potentially reduce stigma . The DOL broadly reported how individuals and organization of public interest took the sides of survivors and how survivors were helped with goods and psychosocial support . Awareness-raising on the modes of transmission and on how to protect oneself was a prominent issue; engaging survivors in the response was also reported about and portrayed as crucially important to improve the perception as well as the situation of survivors . However , the effectiveness of these interventions was never critically questioned , specifically whether the stigma reduction measures , support and reintegration schemes were appropriate to reduce stigmatization in communities and at the workplace . The portrayal of stigmatization of survivors resembles what has been reported in scientific studies from earlier EVD outbreaks [53 , 56–59] . Approaches that compared stigmatizing attitudes towards people living with HIV and persons who suffered from EVD found striking similarities between the two groups and conclusions are quite similar to what the DOL was actually promoting: i . e . , to empower survivors , mobilize opinion leaders in the communities , and disseminate accurate information [60] . However , specifically in the case of infectious diseases it is also important to acknowledge that there is a fine line of what constitutes justified risk behavior for fear of contagion and irrational blaming and shunning of victims [61] . The persistence of Ebola RNA in semen of survivors [62] and minimally symptomatic individuals [21] might add to these biomedical uncertainties . To the best knowledge of the authors this is the first analysis of media reporting on the situation of Ebola virus disease survivors . Even though there is acknowledgment that stigma and discrimination regarding EVD survivors has to be addressed [20] , in fact very little research has been conducted so far . The greatest strength of this study is that it scrutinizes questions of portrayal and acknowledgment in one of the largest online newspapers in one of the most affected countries of the recent Ebola outbreak . This study points to the portrayal of social aspects of survivorship and its implications on stigma , transmission and the potential for improved public health ( emergency ) response [32] . One limitation of this study is that it is restricted to the Liberian Observer online newspaper available in English and therefore identifies a specific portrayal in this Ebola epidemic . Based on the latest published data the illiteracy rates in Liberia are at 52 , 4% [40] . This means that limitations are also evident due to the readership not being representative of the total population . In a survey on mass media access and consumption in Liberia 63% identified radio as the main source of information , and 29% mentioned friends and family [39] . In the face of social media and internet-based data potential for real-time reporting and surveillance and controlling of infectious disease [63] future studies on media reporting on Ebola survivors are recommended to include multiple media sources and types .
|
The largest Ebola epidemic occurred in West Africa between 2013 and 2016 . Liberia was one of the most affected countries with more than 1500 survivors registered . In the height of the Ebola outbreak survivors were increasingly included in the treatment and care of patients and in health sensitizing and social mobilisation efforts . However , the reintegration of survivors back into their communities proved to be challenging across West Africa . Media reporting plays a crucial role in health emergency situations . It gives visibility to what is considered as relevant , frames risk perception and can induce positive health behaviour practices and attitudes . In this study we analysed how one of Liberia’s largest newspapers portrayed and informed the public on the social situation of survivors , in what way it addressed the issues of stigma and discrimination and which stigma reduction interventions were covered and how . We found that reporting was overall comprehensive and well-balanced in terms of disseminating available medical knowledge , scrutinizing stigma . Reports also reflected on contributing factors such as untrustworthiness towards response as well as inconclusive understanding of cures and transmission routes . In a larger context this specific reporting was acting as corrective to what the larger public believed .
|
[
"Abstract",
"Introduction",
"Material",
"and",
"methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"sociology",
"tropical",
"diseases",
"microbiology",
"geographical",
"locations",
"ebola",
"hemorrhagic",
"fever",
"social",
"sciences",
"health",
"care",
"viruses",
"filoviruses",
"rna",
"viruses",
"neglected",
"tropical",
"diseases",
"africa",
"viral",
"hemorrhagic",
"fevers",
"infectious",
"diseases",
"social",
"discrimination",
"semen",
"health",
"economics",
"medical",
"microbiology",
"epidemiology",
"microbial",
"pathogens",
"economics",
"people",
"and",
"places",
"ebola",
"virus",
"anatomy",
"viral",
"pathogens",
"physiology",
"liberia",
"biology",
"and",
"life",
"sciences",
"viral",
"diseases",
"health",
"education",
"and",
"awareness",
"hemorrhagic",
"fever",
"viruses",
"organisms"
] |
2017
|
"We are survivors and not a virus:" Content analysis of media reporting on Ebola survivors in Liberia
|
Self-renewing organs often experience a decline in function in the course of aging . It is unclear whether chronological age or external factors control this decline , or whether it is driven by stem cell self-renewal—for example , because cycling cells exhaust their replicative capacity and become senescent . Here we assay the relationship between stem cell cycling and senescence in the Caenorhabditis elegans reproductive system , defining this senescence as the progressive decline in “reproductive capacity , ” i . e . in the number of progeny that can be produced until cessation of reproduction . We show that stem cell cycling diminishes remaining reproductive capacity , at least in part through the DNA damage response . Paradoxically , gonads kept under conditions that preclude reproduction keep cycling and producing cells that undergo apoptosis or are laid as unfertilized gametes , thus squandering reproductive capacity . We show that continued activity is in fact beneficial inasmuch as gonads that are active when reproduction is initiated have more sustained early progeny production . Intriguingly , continued cycling is intermittent—gonads switch between active and dormant states—and in all likelihood stochastic . Other organs face tradeoffs whereby stem cell cycling has the beneficial effect of providing freshly-differentiated cells and the detrimental effect of increasing the likelihood of cancer or senescence; stochastic stem cell cycling may allow for a subset of cells to preserve proliferative potential in old age , which may implement a strategy to deal with uncertainty as to the total amount of proliferation to be undergone over an organism’s lifespan .
An important goal of aging research is not just to extend lifespan—which in C . elegans can be simply achieved by a pause in developmental and reproductive activities in the “dauer” state [1]—but to do so in a way that increases “healthspan” without diminishing organ activity . To this end , it is critical to understand whether aging is driven by organ activity or whether it is a simple function of chronological age [2] . The C . elegans gonad provides a powerful model system to address this question . Previous studies have identified mechanisms by which the “reproductive lifespan”—the period of adulthood over which C . elegans hermaphrodites can bear progeny—can be extended ( e . g . [3] ) . But this extension does not increase the brood size , which is in fact substantially reduced . This suggests a tradeoff between reproductive lifespan and brood size , compatible with reproductive senescence being driven by reproductive activity ( see also [4] ) . That reproductive senescence is driven by reproductive activity is however contradicted by a report that aging individuals lose “reproductive capacity”—the maximum brood size an individual is capable of producing from a given point in time until cessation of reproduction—as a function of chronological age rather than reproductive activity [2] . Here we resolve this apparent contradiction by showing that the loss in reproductive capacity—a phenomenon we refer to as “reproductive senescence” because it mimics the loss of function in other self-renewing organs—is driven chiefly not by increasing chronological age , but by activity of the gonad , and in particular by germline stem cell cycling . To ask whether reproductive senescence is a simple function of chronological age , or whether it is driven by reproductive activity itself , it is useful to manipulate that reproductive activity ( e . g . [2] ) . There are two naturally-occurring C . elegans sexes: males and hermaphrodites . Hermaphrodites can either self-fertilize ( abbreviated as “self” below ) with the ~300 stored self-sperm they produce during development , or be cross-fertilized with male sperm transferred during mating , which allows brood sizes of up to 1 , 200 [5] . Brood size of mated hermaphrodites is limited by senescence of the reproductive system , which ultimately stops producing fertilizable oocytes [3] . Reproductive activity can be modulated in a physiological way by controlled mating of hermaphrodites that are feminized—i . e . turned into “females”—by mutation of genes such as fog-1 or fog-2 [6–8] . These females do not produce self-sperm , but form an otherwise fully-developed reproductive system in which oocyte maturation and growth by cytoplasmic streaming are substantially reduced [9 , 10] . Females can bear progeny only after mating with males , whose sperm trigger oocyte maturation and fertilization . Virgin fog-2 females were shown to undergo reproductive senescence at roughly the same rate as reproductively-active hermaphrodites [2] despite the fact that the ovulation rate of fog-2 females is much reduced [9] , suggesting a time-intrinsic senescence mechanism . But the mitotic zone of the germ line in feminized worms was subsequently shown to possess M-phase cells [11] , suggesting that stem cells keep actively cycling even in virgin females . Reproductive senescence could therefore be driven by activity rather than being a function of chronological age . Here , we use genetic , environmental , and pharmacological manipulations to assay the relationship between gonad activity and senescence . We establish a causal relationship between the two . We further characterize germ cell cycle behavior on a gonad by gonad basis , using a new technique we developed , and find intermittent activity . Our results strongly suggest that switching between active and inactive states is stochastic .
To begin identifying causes of reproductive senescence , we followed a two-fold approach . First , we characterized loss of reproductive capacity over time in various genetic backgrounds known to differ in proximal gonad activity . Second , we asked whether high gonad activity early in life diminishes remaining reproductive capacity . To modulate proximal gonad activity , we selected fog-1 and fog-2 females described above , which have low ovulation rates [9 , 10] . We used for comparison inx-22; fog-2 females and spe-8 hermaphrodites—both of which are also sterile ( unless mated ) ; loss of inx-22 results in precocious oocyte maturation in feminized gonads even in the absence of the sperm signal [9 , 10 , 12] , while loss of spe-8 preserves stimulation of oocyte maturation by self-sperm that are incapable of fertilizing the oocytes [13 , 14] . Virgin spe-8 and inx-22; fog-2 females ovulate at a rate close to wild-type ( 1 . 7/h , 0 . 9/h , and 2 . 2/h , respectively [9 , 14] ) , substantially higher than that for fog-2 ( 0 . 2/h; n = 31 ) , which is in turn significantly higher than that for fog-1 ( 0 . 1/h; n = 35; p < 0 . 04 ) . We mated virgins at either day 0 , 1 , 2 , 3 , 7 or 10 of adulthood and assayed total brood size ( Fig 1A ) . We found that spe-8 and inx-22; fog-2 undergo faster reproductive senescence than either fog-1 or fog-2 . A two-way analysis of variance considering age of mating and genotype identified a significant main effect of genotype on reproductive capacity , as well as a significant interaction effect ( S1A Table ) . For example , at day 2 of adulthood ( S1 Fig ) post-hoc analysis showed all pairwise differences to be significant except for the inx-22; fog-2 / spe-8 pair ( S1B Table ) . Thus , across the genotypes that we studied , the ranking of reproductive senescence rates from fastest to slowest is: inx-22; fog-2 = spe-8 > fog-2 > fog-1 . Therefore , gonad activity as measured by oocyte production correlates positively with the rate of reproductive senescence . To ascertain whether early gonad activity diminishes remaining reproductive capacity , we compared the reproductive capacities of three groups of fog-2 females at day 7 of adulthood ( Fig 1B ) . The first group was mated at the onset of adulthood , which causes more active germ cell cycling ( see below for detailed cell cycle analysis ) . To verify that females in the first group did not run out of sperm late in life , a second group was mated both at the onset of adulthood and again at day 5 , which did not lead to a significant increase in brood size ( S1C Table; see also [2] ) . The third group was only mated at day 5 and had more than 3 times as many progeny after mating as the first or second group did over the same period ( S1C Table ) . Therefore , consistent with [4] , increased germline activity caused by mating is associated with hastened reproductive senescence . We next asked whether strong inhibition of germline activity using conditions likely to be encountered in the wild also led to a delay in reproductive senescence . We starved fog-2 females from the last larval stage ( L4 ) for two days ( Fig 2A ) . Despite being starved and experiencing a ~30-fold drop in germline mitotic index ( S2A Table ) , females progressed from L4 to the adult stage and produced fully-formed oocytes . We did not observe shrinking of the germ line or the reproductive diapause identified in wild-type by [15] ( Fig 2B ) , likely because the absence of sperm in females prevents the redirection of resources to slowly-growing embryos proposed by [16]; consistent with this , starvation for 2 days markedly reduced the number of apoptotic cells identified in females using a feminized ced-1::gfp reporter strain , from 18 . 3 per gonadal arm to just 1 . 9 ( S2B Table and S2 Fig; note that this is in contrast to starvation of hermaphrodites over a 6-hour period [17] ) . We then returned the females to food for 24 h , mated them with males , and assayed their brood sizes . Compared to continuously-fed controls of the same age , brood size was increased by almost 2-fold ( S2C Table and Fig 2C ) . This increase occurred even as apoptosis is restored to normal levels when females are returned to food ( S2D Table and S2 Fig ) . Starved worms thus retain a higher percentage of their reproductive capacity during the starvation period compared to controls ( Fig 2D ) . While starvation has pleiotropic effects such as increased stress resistance [18] , these results are also compatible with the idea that germline activity—measured either by germ cell cycling or oocyte production—drives reproductive senescence . To ask directly if reproductive senescence is driven by germline activity , and to distinguish between the influence of cell cycling distally and oocyte maturation proximally , we next performed cell cycle inhibition experiments ( Fig 3A ) . We fed the small molecule hydroxyurea ( HU ) , a specific inhibitor of DNA synthesis ( germ cells are the only mitotically active cells in adults ) . We found that HU treatment substantially reduces the incidence of M-phase cells within 12 h ( from 3 . 2 to 0 . 35; n = 18 and 17 respectively ) and eliminates it after 1 day ( n = 20 ) . We tested whether germ cell cycle arrest was accompanied by a reduction in ovulation rate by exposing virgin females to HU for 24 h at day 1 of adulthood and counting the number of oocytes laid during this period . We found that for both fog-2 , which has a low ovulation rate , and inx-22; fog-2 , which has a much higher ovulation rate , there was no effect of HU treatment on ovulation ( S3A Table and Fig 3B ) . We compared reproductive capacities of females that prior to mating received a 24 h treatment with HU or a control treatment without HU . HU treatment increased fog-2 reproductive capacity by 27% ( S3B Table and Fig 3C ) and increased inx-22; fog-2 reproductive capacity by 50% ( S3B Table and Fig 3C ) . Cell-cycle arrested females thus retain a higher percentage of their reproductive capacity compared to controls ( Fig 3D ) . HU treatment has a number of effects other than cell cycle arrest . While these effects are detrimental [19 , 20] and are expected if anything to hasten senescence , the possibility remained that HU had a hormetic effect . To test this idea we asked whether HU treatment increased lifespan or thermotolerance , but found that it in fact decreased them ( S3C Table and S3A and S3B Fig ) . While this does not formally exclude the possibility of a hormetic effect on reproductive lifespan , it shows that HU does not have a global beneficial effect on worm health . Overall , the increase in brood size that results from the HU treatment is thus remarkable . To confirm that delayed reproductive senescence induced by HU is not an off-target effect , we also performed a cell cycle inhibition experiment using the selective cyclin-dependent kinase inhibitor ( CDKI ) Roscovitine . Roscovitine has previously been used to reversibly inhibit mitosis in starfish and sea urchin embryos [21]; we found that it effectively reduces the mitotic index in worm germ lines ( S3D Table and Fig 3E ) . We found that the reproductive capacity of females treated with Roscovitine for 24 h prior to mating was over 2-fold larger than that of females that received a DMSO control treatment ( S3E Table and Fig 3F ) . While HU- and Roscovitine-treatment brood sizes cannot be compared directly , because treatment with DMSO—used as a Roscovitine solvent—decreases brood size [22] , the qualitative effects of HU and Roscovitine treatments are the same . The role that we uncovered for cell cycling in driving reproductive senescence , combined with the role of DNA damage in driving senescence in other systems [23] , led us to wonder if cell cycling could curtail reproductive output at least in part through the DNA damage response ( DDR ) . To test whether cell cycling leads to increased DDR , we focused on the single stranded DNA binding Replication Protein A ( RPA-1 in worms ) . While RPA-1 is activated in response to multiple forms of DNA damage [24–26] , it plays in particular a role in the repair of induced double strand breaks ( DSBs ) associated with meiotic crossovers . This role leads to presence of RPA-1 foci in the pachytene stage of meiosis [27 , 28] . Defects in chromosome segregation in the mitotic zone do not lead to increased prevalence of foci in the mitotic zone itself , but instead lead to accumulation in late pachytene of meiosis-induced DSBs that are not resolved [29] . We quantified RPA-1::YFP foci in pachytene ( zones 5 and 6 defined by [28] ) , in young rpa-1::yfp hermaphrodites taken at day 1 of adulthood , and in rpa-1::yfp hermaphrodites taken at day 4 of adulthood that had been selfed or that had been mated at day 0 of adulthood . We found over ~5-fold more foci per nucleus in the mated group than in either of the selfed groups ( S4A Table and Fig 4A ) . Consistent with the pattern of increased cycling being associated with increased RPA-1 foci , selfed hermaphrodites had more foci at day 4 than at day 1 ( S4A Table ) . Overall , although it remains to be established how increased cycling may lead to chromosome segregation defects or other defects in the mitotic zone , these results strongly suggest that this increased germ cell cycling leads to increased proximal germline DDR . Our results show that gonad activity strongly contributes to reproductive senescence , mostly as a result of germ cell cycling . We thus hypothesized that , at any given point in the reproductive life of a worm , remaining reproductive capacity is inversely related to the total amount of germ cell cycling that has occurred up to that point . To test this hypothesis , we decided to ask whether a dose-dependent relationship exists between average cell cycle rate and the rate of reproductive senescence by using genetic or environmental mutations that modulate cell cycling . To identify suitable genetic manipulations , we compared fog-1 and fog-2 females with inx-22; fog-2 and spe-8 . Since the former two strains undergo slower reproductive senescence than the latter ( see above; Fig 1A ) and have lower proximal gonad activity , we surmised that they might also have slower distal cycling . We also assayed for changes in cell cycle when oogenesis rates are high due to the presence of self or male sperm , or low due to depletion of self sperm . We first detail our cell cycle analysis , and subsequently compare results to reproductive senescence data . We first attempted to determine average cell cycle speeds . To this end we carried out pulse-chase experiments using bacterial food labeled with 5-Ethynyl-2’-deoxyuridine ( EdU ) , which is incorporated by cells in S phase . Although virtually all young , selfed wild-type mitotic zones contained EdU-positive cells following a 30-minute pulse ( consistent with previous reports; [32 , 33] ) , many mitotic zones from virgin females contained no labeled cells . If this lack of S-phase labeling was due to all cells in a given gonad being found by chance in G1- , G2- , or M-phase , given that M-phase length is about 10% that of G1 , G2 , and M combined ( e . g . [33 , 34] ) , and mitotic zones contain ~260 cells ( S5E Table ) , one would expect to find on average in such mitotic zones 10% of 260 cells , i . e . 26 cells , in M-phase . Since neither we nor others [35] have observed mitotic zones with such a large number of M-phase cells , lack of any S-phase cell in a mitotic zone indicates that the mitotic zone as a whole is in a dormant state . The two most likely explanations for the presence of unlabeled mitotic zones were thus either that the unlabeled worms had not ingested the EdU-labeled bacteria , or that the mitotic zones were dormant at the time of the pulse . To explore these possibilities , we switched to a continuous EdU labeling assay . In wild-type gonads at day 1 of adulthood , virtually all mitotic zones had at least one labeled cell within the first time point we assayed , with only 5% remaining unlabeled ( 1 h labeling; see example images in Fig 5A and detailed graphs in Fig 5B ) . The proportion of unlabeled mitotic zones at 1 h was substantially higher for fog-1 ( 55% ) and for fog-2 ( 17% ) than for wild-type; it took in excess of 6 h for labeled cells to have appeared in all feminized mitotic zones ( Fig 5A and 5B ) . To test whether labeling delays could be due to pauses in feeding , we fed virgin worms bacterial food mixed with fluorescent beads . At the very first time point we assayed ( 1 h ) , 100% of wild-type , fog-1 , and fog-2 worms had ingested the fluorescent beads ( see example in S4A Fig ) . This strongly suggests that different labeling rates are due to bona fide differences in mitotic zone cycling , with a substantial proportion of fog-1 and fog-2 gonads being in a dormant state in which germ cells are not progressing through the cell cycle . Continuous EdU labeling results show the existence of at least two states in which reproductively-inactive gonads can reside—an actively-cycling state , and a dormant state—and suggest that stochastic switching occurs between these states . Specifically , that different cycling states exist within the population of feminized , reproductively-inactive gonads can be inferred from the large differences in labeling times for e . g . fog-2 gonads ( the same reasoning applies to fog-1 gonads ) : some gonads label almost immediately , showing that they are actively progressing through S-phase at the time of label application , while others take up to 8 h and thus do not have cells actively progressing through S-phase during these 8 h . If fog-1 and fog-2 simply cycled continuously , in the same way as wild-type but with uniform slowing down of all cell cycle phases , the distribution of cell cycle phase indices would be unchanged and the number of gonads with at least one cell progressing through S-phase would be the same as in wild-type at any given time—and the continuous labeling time courses would therefore appear identical for all genotypes . Stochasticity can be inferred from the fact that EdU labeling of reproductively-inactive gonads happens immediately for some gonads but takes up to 8 h for others: this shows independent behavior of gonads with identical genotypes , for which stochastic switching between active and dormant states is the most parsimonious explanation . To begin confirming the existence of such states by independent means , we quantified the coefficient of variation ( CV ) in mitotic index . The CV is defined as the ratio of standard deviation to mean and thus provides a unit-free measure of noise . The fog-1 and fog-2 CVs were 2 . 3- and 1 . 7-fold higher , respectively , than the wild-type CV , differences that are significant at a 95% confidence level ( S5A and S5B Table ) . These differences show a larger spread among the population of the mitotic index measured at a given point in time , as expected if there is a mixture of gonads with little or no mitosis ( the dormant subpopulation ) and gonads cycling as normal ( the active subpopulation ) . This supports the idea that feminized gonads switch back and forth between active and dormant states . This idea is further tested by independent means below . We asked whether the dormant state we identified was an artificial byproduct of mutations in fog-1 or fog-2 , or whether it was a state naturally occupied by sperm-deprived gonads . We first performed continuous EdU labeling of selfed wild-type hermaphrodites taken at day 3 of adulthood , by which time reproductive activity has started declining; full labeling took 4 h ( S4B Fig ) , twice as long as for hermaphrodites at day 1 ( which are fully active ) . Mitotic zones from mated hermaphrodites assayed at day 3 of adulthood , however , labeled at the same speed as selfed worms at day 1 ( S4B Fig ) . We next performed continuous EdU labeling using inx-22; fog-2 or spe-8 gonads , which do not produce embryos but retain sustained oocyte maturation . The fractions of dormant inx-22; fog-2 and spe-8 mitotic zones were closer to that of wild-type ( Fig 5B ) than to that of fog-1 or fog-2 . Similarly , inx-22; fog-2 and spe-8 times to labeling of all mitotic zones were closer to that of wild-type than to that of fog-1 or fog-2 ( Fig 5B ) . The differences in median labeling times were significant for all genotype pairs ( p < 7 . 8E-3 , Wilcoxon test with Bonferroni correction for 15 tests; see Methods ) , except for pairs taken within the group formed by wild-type , spe-8 , and inx-22; fog-2 , which all behave similarly ( p > 0 . 05 ) . Differences in fractions of dormant gonads matched differences in mitotic index CV: the CV for fog-1 was significantly higher than that for fog-2 , which in turn was significantly higher than the CV for wild-type , inx-22; fog-2 or spe-8 ( S5B Table ) . Finally , we asked whether the behavior we observed was particular to C . elegans , or whether it was shared with other nematode species . We chose C . remanei [36] , for which genetic manipulation is unnecessary for feminization because it is a male/female species , but which undergoes reproductive senescence much like C . elegans females ( S4C Fig ) . We found that the fraction of dormant mitotic zones was 27% for virgin C . remanei females , but that mating decreased that fraction to 5% ( S4D Fig ) . Therefore , the mitotic zones are dormant at a substantially higher frequency in reproductively-inactive gonads than in reproductively-active gonads , both in C . elegans and in the related nematode species C . remanei . Having established that less active gonads experience periods of cell cycle dormancy , we returned to our pulse-chase dataset . To quantify cell cycle rates of active mitotic zones , and to verify that mitotic zones switch back to the dormant state from the active state , we developed a method that , instead of relying on population averages , computes cell cycle progression of individual mitotic zones . We segmented individual cells in confocal stacks of mitotic zones that were fixed and stained for DNA and EdU , and fit the data to computational simulations of germ cell cycling ( see S1 Text ) . Overall progression of mitotic zones through the cell cycle is computed as a phase that defines a position on a circle , with one full revolution corresponding to a complete cell cycle ( Fig 6A ) . A mitotic zone that labels during the EdU pulse but becomes dormant during the chase would stop progressing around the circle ( Fig 6B ) . We validated our technique using wild-type young adult mitotic zones , for which a full revolution took ~5 . 5 h to complete ( Fig 6C ) ; this estimate of cell cycle length is consistent with previous findings [33 , 37] . We then compared initial cell cycle progression during the EdU chase between hermaphrodites and virgin females . Initial progression rates ( i . e . rates of non-dormant mitotic zones ) of wild-type , fog-1 , fog-2 , inx-22; fog-2 or spe-8 were largely similar ( S5C Table ) . We estimated average cycling rates as the product of the fraction of active mitotic zones and their initial progression rates . Rankings are: wild type ( 0 . 17 cycles / h ) ≃ spe-8 ( 0 . 20 cycles / h ) ≃ inx-22; fog-2 ( 0 . 18 cycles / h ) > fog-2 ( 0 . 13 cycles / h ) > fog-1 ( 0 . 08 cycles / h ) , which are equivalent to the rankings for reproductive senescence rates . We do not know the molecular basis for the difference in average fog-1 and fog-2 cell cycle rates—slower oocyte maturation in fog-1 ( S4E Fig ) could increase mitotic zone dormancy—but in any case this difference provides a useful tool to investigate the effects of cell cycling activity . To ask if mitotic zones switch from the active to the dormant state , similar to switching from the dormant to the active state shown by continuous EdU labeling experiments , we further assayed the cell cycle progression of mitotic zones pulsed with EdU . Individual mitotic zones of young adult hermaphrodites pulsed with EdU showed little dispersion in their overall cell cycle progression , even after 6 h of chase—covering over a full cycle length ( as shown by visual inspection of phase plots and by quantification of inter-gonad synchrony depicted as wedge width in Fig 6C ) . This shows that reproductively-active gonads have mitotic zones that progress through the cell cycle in a highly-similar fashion . By contrast , individual mitotic zones from virgin , feminized gonads showed substantially larger dispersion along the set of possible phases as shown by increased wedge width in Fig 6C , 6 h time point , for virgin fog-1 and virgin fog-2 ( compare to wild-type day 1 or to mated wild-type day 3 ) . Dispersion was even stronger in wild-type hermaphrodites at day 3—to the point that cell cycle progression appeared largely randomized . Given that initial progression rates are highly similar across the genotypes and conditions that we tested , the simplest interpretation of these results is that the time at which labeled mitotic zones return to the dormant state is stochastically distributed . To test whether dispersion in the amounts of cell cycle progression after labeling was due to reduced reproductive activity , we measured cell cycle activity in mated young females , and in hermaphrodites mated at day 1 of adulthood and assayed at day 3 . In both cases , the increase in reproductive activity caused by mating was accompanied by a switch of cell cycle behavior to that of young adult hermaphrodites ( Fig 6C ) and by a reduction in the mitotic index CV ( S5B Table ) , showing that occupancy of the dormant state is regulated by reproductive activity . Therefore , gonads with reduced reproductive activity switch back from the actively-cycling state to the dormant state in a way that is in all likelihood stochastic . A prediction from a model whereby feminized gonads switch back and forth stochastically between active and dormant states , and whereby gonad activity leads to loss of reproductive capacity , is that there may be an increase in variability of remaining reproductive capacity as the population ages ( Fig 7A ) . This is because stochasticity of switching may lead to mitotic zones spending different total amounts of time in the active state . We thus tested whether inter-individual variability in remaining reproductive capacity does increase with age . To this end , we computed the coefficients of variation ( CVs ) of brood size in fog-1 populations mated at different times ( Fig 7B ) . The CV is defined as the ratio of standard deviation to mean and thus provides a unit-free measure of noise . When mated at the onset of adulthood ( day 0 ) , the brood size CV was 0 . 12 ( Fig 7C and S6 Table ) . By contrast , the brood size CV for mating on day 2 of adulthood was 0 . 33 , i . e . ~3 fold higher . To distinguish between a simple increase in variability associated with aging ( as is observed for many phenotypes; [38] ) and an increase in variability driven by stochastic cell cycling , we considered brood sizes computed from day 2 of adulthood onwards , for worms mated at the onset of adulthood . The CV was 0 . 13 ( Fig 7C and S6 Table ) , virtually identical to the CV of brood sizes scored from day 0 and significantly lower than the CV for worms mated on day 2 of adulthood . Therefore , the age-dependent increase in variability of remaining reproductive capacity for reproductively-inactive females is stronger than the increase that occurs for reproductively-active worms . We verified using a simple simulation that , given a suitable probability distribution of times spent in the dormant and in the active state , a population of females losing the same reproductive capacity as fog-1 over a period of two days can indeed experience a ~3-fold increase in the CV in remaining reproductive capacity ( see S1 Text ) . Overall , these results support the idea that stochastic bursts of cycling drive reproductive senescence . To start querying organismal regulation of the dormant state , we first asked whether the state of a mitotic zone in one gonadal arm relates to that of the mitotic zone in the sister arm from the same worm . At 1 h of continuous EdU labeling , we found 92% agreement in EdU status ( defined by presence or absence of at least one labeled cell; n = 36 pairs ) between pairs of mitotic zones from the same worm; this substantial synchrony in gonadal arm states suggests that dormancy may be regulated at the organismal level , which could perhaps occur through neuronal control of TGF-beta signaling [39] or insulin signaling [40 , 41] . Since population density affects a number of aspects of worm physiology [42–44] , we next asked whether that density also affects mitotic zone dormancy . Starting from the L2 stage , we kept populations of virgin fog-1 females either singled ( low density ) or as group of 70 individuals ( high density ) on 35 mm plates , and mated them at day 4 of adulthood to assay reproductive capacity . We found that females that had been kept at low density had a significantly reduced brood size compared to those that had been kept at high density ( Fig 8A and S7A Table ) . We then assayed mitotic zone dormancy at day 1 of adulthood , and found a significant effect of population density on that dormancy ( Fig 8B and S7B Table ) : 23% of singled female mitotic zones were dormant vs 46% for the higher-density population . Females thus undergo faster reproductive senescence and cycle more actively when they are singled . To ask whether hermaphrodite dormancy is also density-dependent , we performed the same experiment on selfed hermaphrodites assayed at day 3 of adulthood . We observed a similar effect as for females , with 9% dormancy for singled hermaphrodites vs 27% for the higher-density population ( S7C Table ) . As a first step in elucidating the mechanism underlying this population-density dependence , we repeated the experiment using the daf-22 ( m130 ) mutation , which abrogates dauer pheromone synthesis [45] , and found that density dependence of mitotic zone dormancy was lost ( 42% vs 47% for high and low densities , respectively; Fig 8B and S7D Table ) . Mitotic zone dormancy is therefore regulated by population density , most likely through a mechanism that involves dauer pheromone . Finally , we asked whether intermittent cycling is a behavior that is always associated with reduced overall reproductive activity , or whether it is specific to a state in which worms are well fed but deprived of sperm . Caloric restriction strongly reduces the rate of reproduction as well as total brood size [2 , 46] . In our hands , hermaphrodites kept in liquid culture with 1x1010 bacteria/mL ( “high” concentration ) had a similar brood size to hermaphrodites kept on solid medium under standard conditions ( 294 vs . 306; n = 11 and 12 , respectively; p > 0 . 4 ) , but brood size dropped more than two-fold ( 138; n = 11; p < 1 . 5E-6 ) at 1x109 bacteria/mL ( “medium” concentration ) . Further , [46] showed a ~4-fold drop between 1x109 bacteria/mL and 1x108 bacteria/mL ( “low” concentration ) . At day 1 of adulthood , we transferred wild-type adult hermaphrodites to low , medium or high E . coli concentrations for 24 h , and subsequently performed 1 h continuous EdU labeling . In contrast to mitotic zones from females or older hermaphrodites , those from food-restricted young hermaphrodites were all active ( n = 20 each ) , even at the lowest food concentration . This suggests that intermittent cycling is a specific response to sperm deprivation . If stem cell cycling drives reproductive senescence , why would gonads that are not reproductively active maintain cycling and thus hasten their demise ? We hypothesized that gonads that are in the active state are poised for reproduction , and thus respond quickly to a favorable change in environmental conditions . To test this hypothesis we characterized the dynamics of reproduction initiation after mating aged females . After mating of inx-22; fog-2 females at day 3 of adulthood , the rate of viable progeny production increased sharply and remained sustained over the first ~18 h ( Fig 9 ) . By contrast , mated fog-2 females of the same age experienced a transient increase followed by a trough in progeny production between 9 h and 13 h after mating; this trough was more marked and longer-lasting for fog-1 females ( Fig 9; note that the maximal rates were lower for inx-22; fog-2 than for fog-1 or fog-2 , which is expected given the faster reproductive senescence of inx-22; fog-2 ) . The trough likely resulted from a drop in the numbers of developing oocytes in diplotene or diakinesis , which was not experienced by inx-22; fog-2 ( S8A Table and S6A Fig ) , and appeared to be independent of apoptosis: initial progeny production occurred at the same rate in apoptosis-deficient strains carrying a ced-3 mutation ( S8B Table and S6B Fig ) , and apoptosis levels decreased within the first 8 h after mating ( S8C Table and S6C Fig ) and thus likely did not account for a diminished supply of oocytes . In summary , mutant populations that maintain a higher proportion of dormant gonads are less capable of sustaining a high rate of progeny production shortly after mating . Overall , our results are consistent with dormant gonads needing time to make the switch to a state with maximal reproduction rate . Diminished initial reproductive activity of gonads that have been in a dormant state might stem from a detrimental impact of prolonged meiotic arrest on gamete quality or on the gonad region that houses them [47] , or from a delay in returning to the active state . In any case , there is strong selection against reproduction delays [5 , 48] . Avoiding delays provides a powerful rationale against gonads staying fully dormant until reproduction becomes possible .
Our findings significantly extend understanding of C . elegans reproductive senescence . We showed , using direct experimental manipulation of the cell cycle as well as using comparisons of senescence rates in strains that differ in gonad activity , that activity of the gonad is largely responsible for the progression of reproductive senescence—in particular as a result of germ cell cycling . In the course of this study we discovered that germ cell cycling can be intermittent , which as discussed further below suggests that worm populations may employ an active strategy to manage the progression of reproductive senescence . These results are in contrast to the idea that this progression is solely the result of passive damage accumulation as a function of elapsed time . What mechanisms tie reproductive senescence to germ cell cycling ? Mutations that delay senescence may provide an entry point to start addressing this question . But although a number of mutations have been uncovered that increase reproductive lifespan , most do not result in an increase in total brood size and many in fact result in a decrease in brood size—i . e . the reproductive system is less active for a longer period of time , with reduced total output ( e . g . [3] ) . Given our observations that gonad activity drives reproductive senescence , the primary effect of many of these mutations may thus simply be to alter the rate of reproduction , rather than to alter the relationship between reproductive senescence and gonad activity . By contrast , the DDR reduction of function mutant hus-1 ( op241 ) is to the best of our knowledge the first worm mutant reported to have a total reproductive capacity greater than wild-type . What link between cell cycling and reproductive senescence does this mutant suggest ? It is possible that an intact DDR is required for full-speed germ cell cycling , and that hus-1 ( op241 ) germ cells thus cycle more slowly , leading to delayed reproductive senescence; given that hus-1 ( op241 ) does not display slower reproduction as might be expected from slower germ cell cycling , we do not favor this possibility . Another possibility is instead that the DDR , whose intensity according to RPA-1 foci correlates positively with the amount of past cell cycling , could mediate the effects of germ cell cycling on reproductive senescence—through molecular pathways that will require further study . An interesting additional problem will be identifying the downsides of increased reproductive output in hus-1 ( op241 ) mutants . We speculate that late-born progeny are more prone to mutation accrual or genomic rearrangements; further study will be required to characterize the fascinating tradeoff between increased brood size and decreased genome quality in progeny . We also note that germ cell cycling has been shown by others to shorten lifespan , in addition to reducing reproductive capacity as we report here: mating acts via DAF-12 –a key mediator of lifespan control by the reproductive system [49]–to cause somatic “shrinking” and death , while cell cycle inhibition protects from this effect [50] . It will prove interesting to ask whether the reproductive senescence and somatic lifespan effects of germ cell cycling are enacted by overlapping mechanisms . The role of germ cell cycling in driving reproductive senescence makes intermittent cycling particularly intriguing . Much of this intermittent cycling phenomenon remains to be characterized . The lack of suitable live imaging techniques makes it necessary to infer the existence of dormant and active states and the rules for transition between these states , and certainly leaves open the possibility that , in the future , more fine-grained studies will identify a broader array of sub-states and more complex transitions between them . The existence of at least two states , an active and a dormant one , is strongly supported by EdU continuous labeling studies that show that some virgin females mitotic zones label immediately when placed on EdU while some do not . This is compatible with germ cells being capable of stopping replication within S-phase or significantly lengthening S-phase [51 , 52] . Further , analysis of mitotic index CV shows greater variability between mitotic zones of virgin female populations than between those of mated female or young hermaphroditic populations . This is compatible with the existence of at least two cell cycle states—dormant and active—but there could be a continuum of intermediary states , whose practical relevance may depend on the amount of time they perdure . Until cell cycle states are more finely defined at the molecular level , control of transitions between states can only be addressed in qualitative terms . The spread in virgin female mitotic zone cell cycle progression during EdU pulse-chase experiments ( Fig 6C ) suggests that the transition between dormant and active states has a stochastic aspect ( although we cannot formally exclude that individuals possess strongly intrinsic differences in switching rates ) ; this is also suggested by the larger reproductive capacity CV in virgin females mated at day 2 than in those mated at day 0 , which likely stems from a random distribution of time between days 0 and 2 spent in the active state that reduces reproductive capacity . We note however that biological phenomena are often described as “stochastic” until a deterministic underlying is identified; for example , control of the lysis-lysogeny decision of the lambda phage was described as stochastic but cell volume turned out to be a strong predictor of cell fate [53] . The switch between actively-cycling and dormant states could thus be downstream of an unknown highly-deterministic mechanism rather than being controlled by molecular noise . But in any case stochastic switching between active and dormant states provides a fitting and parsimonious model for our current data . The molecular controls of intermittent cycling remain to be fully established . It may seem surprising that fog-1 and fog-2 , both known for their role in germline sex determination , have different cell cycle dynamics: fog-1 undergoes slower average cell cycling that does fog-2 . This could be consistent with the previously-established , dose-dependent role of FOG-1 in promoting proliferation [54] . But a complication in comparing fog-1 and fog-2 exists in that virgin fog-1 ovulates at a slower rate than virgin fog-2; it is possible that a slower loss of cells to oogenesis leads to slower homeostatic mitotic zone cycling , rather than there being a direct role for fog-1 in controlling cell cycling . In line with this idea , it has most interestingly been shown recently that oocyte accumulation inhibits germ cell proliferation [41] . Lastly , fog-1 and fog-2 differ in the foraging behaviors [55] , which could impact germ cell cycling . Overall , further experiments will be required to address the different behavior of fog-1 and fog-2 . In addition to the fog-1 and fog-2 genes better known for their role in germline sex determination , our results and those from another recent study [41] implicate a number of genes and pathways in control of proliferation in virgin female mitotic zones . We have identified a role played by daf-22 , likely through the dauer pheromone pathway , while the daf-18 and insulin pathways were implicated by [41] ( in a way that did not address dormant and active states but that is compatible with our results when considering average cell cycle speed ) . Further experiments will be required to derive a more comprehensive understanding of the integration of these molecular controls . What is the relevance of intermittent cycling ? First , we address whether it is a general response to a reduced need for germ cell production . A germline proliferation stop occurs quickly in females after deprivation of food [51 , 52] . In cases where a reduced average proliferation rate is warranted , different possibilities would be for mitotic zones to be reduced in size , for them to cycle at a slower but steady rate , or for them to go through periods of dormancy . We have not observed substantial changes in mitotic zone size . We did observe changes in cell cycle length across developmental stages: we recently reported that germ cells cycle ~60% more slowly in young adult hermaphrodites than in L4 hermaphrodites , in a way that does not involve intermittent cycling but rather S-phase lengthening [34]; we have also shown in the present study that caloric restriction that reduces the rate of reproduction over 2-fold also does not result in intermittent cycling . Therefore , intermittent cycling is a response that so far appears specific to well-fed females and hermaphrodites with a diminished sperm supply . Next , we address whether intermittent cycling may occur in the wild in a fashion that is relevant to “fitness . ” We focused largely on genetically-feminized C . elegans hermaphrodites in the present study , because C . elegans has been more thoroughly studied than other species and can be investigated with better established genetic tools . Although C . elegans females do not occur naturally , we showed that the same intermittent cycling behavior occurs in C . remanei , which is a gonochoristic ( i . e . male/female ) species , and in older hermaphrodites . What is the relevance of germ cell behavior in older , sperm-depleted hermaphrodites ? Two important questions to address this relevance are 1 ) how strong a selective pressure applies to hermaphrodites around the end of reproduction and 2 ) whether hermaphrodite mating is of much relevance in a mostly selfing species . We discuss these two questions in turn . Regarding whether reproduction of older hermaphrodites is under active selection , there is a general expectation that late-life reproduction might only be under weak selection ( [56]; although see e . g . [57] for recent developments ) . Interestingly however , hermaphrodites show mating behavior that is specific to old age . Specifically , older hermaphrodites release a volatile pheromone that attracts males [58] , which is regulated by the CEH-18 sperm sensing pathway [59] . In addition , older hermaphrodites mate more rapidly with males and eject sperm with reduced frequency than young hermaphrodites [60] . If evolutionary pressures were such that behavior of older hermaphrodites was of little relevance to fitness , one would expect that mutation accumulation would have caused these old-age specific behaviors to be lost [see e . g . 61] . It may of course be that these behaviors are just a remnant of a not-too-distant gonochoristic past . But in that case , the intermittent cycling behavior we observe would likely also be preserved from the previous gonochoristic state and thus amenable to study in C . elegans—even if it has lost its direct purpose . Overall , hermaphrodite mating that occurs under conditions of intermittent cycling is likely of relevance to fitness and is thus under active selection , or was in a suitably-recent past . Regarding the general in-the-wild relevance of outcrossing ( irrespective of hermaphrodite age at the time of mating ) , although C . elegans males have only been isolated at a very low frequency in the wild [62 , 63] heterozygosity data show that outcrossing does occur in the wild—crucially , at a higher frequency than expected from mere spontaneous generation of males through meiotic non-disjunction of the X chromosome [reviewed by 61] . Males thus appear to be actively maintained in the wild . Consistent with this , experiments in the laboratory have identified conditions likely to be encountered in the wild , such as increased mutational pressure or exposure to changing environments , that lead to maintenance of males at a high frequency [61 , 64] . Lower than expected observed frequencies in the wild may be due to outbreeding depression [61] , and male frequency may thus be actively maintained at an equilibrium frequency by counteracting evolutionary forces . Taken in combination with the fact that older hermaphrodites are more likely than young hermaphrodites to have cross progeny with males , this strengthens the idea that intermittent cycling in older hermaphrodites is a behavior that is of relevance in the wild and actively selected for , even if that selection is weaker than the selection for early hermaphroditic reproduction . If future reproduction of intermittently-cycling hermaphrodites is relevant to in-the-wild conditions and is specifically selected for , how would intermittency be beneficial ? We speculate that uncertainty in the time at which reproduction will become possible could perhaps underlie a bet-hedging strategy . Bet-hedging is a well-established strategy followed by unicellular and multicellular organisms to avoid or spread risks in the face of uncertain environmental conditions [65 , 66] , and phenotypic heterogeneity is a mechanism by which bet hedging can be implemented [67] . Specifically , a model is conceivable whereby in unfavorable conditions C . elegans populations hedge their bets by maintaining individuals that are primed for reproduction at the cost of faster senescence , and individuals whose gonads are dormant and that thus senesce more slowly , helping preserve reproductive capacity of the population over time . Stochasticity in the behavior of individual gonads would then have two roles . First , it would allow individuals to modulate the average rate of germ cell cycling . Second , and more interestingly , stochasticity is a parsimonious mechanism to develop a broad distribution of effective reproductive senescence rates in the population , without those rates being pre-assigned to each individual . Individuals that “take a chance” by cycling allow the population to quickly initiate reproduction if conditions become favorable shortly after they cycle , while individuals that stay in a dormant state avoid the population going extinct if conditions only become favorable after an extended period of time ( during which individuals that frequently cycled exhausted their reproductive capacity ) . To our knowledge , such bet hedging would be the first to directly control senescence of a self-renewing organ , and the first to be implemented by integration of stochastic switching between different states . An interesting problem for future studies would be to ask whether bet-hedging strategies are used by other self-renewing organs in which stem cell cycling contributes to maintenance of tissue function , but increases the probability of a cell becoming senescent to avoid cancer [68] . Finally , we note that there may be a number of alternative reasons for intermittent cycling . Intermittent cycling may be an unselected-for , side-effect behavior that derives from the particular structure of the yet-to-be-characterized gene network that regulates it , while the increase in phenotype variability in old age could be explained by phenotypic drift [69] . Rather than implementing a bet hedging strategy , intermittent cycling could provide a convenient means for cells to alternate between different metabolic states , or could be a way of maintaining the same distribution of cell cycle phases in the mitotic zone irrespective of the average cycling rates—so that mitotic zones can assume an optimal distribution quickly when reproduction initiates . Future studies will be required to explore these ideas .
Strains used were Bristol N2 , JK3743: fog-1 ( q785 ) I [11] , JK1078: fog-2 ( q71 ) V [6] , XM1012: inx-22 ( tm1661 ) I; fog-2 ( q71 ) V [9] , EM464: C . remanei [36] , BA784: spe-8 ( hc50 ) I [13] , CNQ41: fog-1 ( q785 ) I; ced-1::gfp ( bcIs39 ) V ( obtained by a cross of MD701 [70] and JK3743 ) , CNQ44: ced-3 ( n717 ) IV; fog-2 ( q71 ) V ( obtained by a cross of MT1522 [71] and JK1078 ) , WS4581: rpa-1::yfp ( opIs263 ) [28] and WS2277: hus-1 ( op241 ) I [31] . Strains were maintained as described [72] using E . coli HB101 as a food source . Worms were staged by picking at the L4 stage as identified by visual inspection of vulva shape . Unless otherwise specified , 50 worms were kept per 60 mm plate prior to mating . Mating at “day 0” of adulthood refers to exposure to males for 24 h starting from L4—so that females are exposed to males at the earliest time at which they are sexually mature . More generally , mating at day n of adulthood refers to exposure to males from L4 + n * 24 h to L4 + ( n+1 ) * 24 h . For mating and cell cycle experiments , a sharp increase in cycling intermittency was noted in wild-type selfed hermaphrodites around day 3 adulthood . Therefore , when using day 3 hermaphrodites , special care was taken to use worms at precisely L4 + 72 h . Female mating was performed on 35 mm-diameter plates ( CC7672-3340 , USA Scientific , Ocala , FL ) , at a density of 1 female and 3 young adult males per plate . For C . remanei matings , females were continuously exposed to males , which were refreshed every three days; without this continuous exposure , female reproductive capacity is not exhausted ( progeny count with 24 h male exposure was 288 , n = 20 , compared with 734 , n = 20 , with continuous exposure ) . For brood size scoring , mothers were passaged every day to 35 mm fresh plates , until the end of reproductive activity . Plates on which embryos had been laid were incubated for ~2 days to let progeny hatch and grow in size; progeny were counted before they reached the adult stage . Mothers that crawled off agar were censored from the analysis . To analyze the early response to mating in fine detail , females isolated 3 days after L4 were each plated with 3 males and any progeny removed from the plate every ~3 h for a total of 21 h . After another 24 h , transferred progeny were scored for viability . For cell cycle inhibition experiments , hydroxyurea ( HU ) or the CDK inhibitor Roscotivine ( H8627 and R7772 , respectively Sigma-Aldrich , St . Louis , MO ) were freshly prepared and added to NGM-agar ( cooled to 55°C prior to dispensing ) at a final concentration of 40 mM for HU [73] or 50 μM for Roscovitine . Roscovitine was diluted in DMSO so that the final dilution of the solvent in NGM was 1:1000; control plates were also supplemented with DMSO at 1:1000 . The following day , bacteria were UV-killed by placing open seeded plates in a Spectrolinker XL-1500 ( Spectroline , NY ) for 5 min , and then transferred onto HU , Roscovitine or control plates . Immediately after , 1-day adult virgin females were transferred to HU or control plates for 24 h . They were then returned to regular plates and mated as above . For starvation experiments , females were picked at the late L4 stage , rinsed 5 times in M9 , and starved in complete S-medium [74] . Following starvation , females were transferred to seeded plates , on which they were kept for 24 h prior to mating . For the thermotolerance assay , worms were transferred to HU or control plates at day 1 of adulthood , kept on these plates for 24 h , returned to regular plates , and shifted to 35°C . Every 2 h until all worms had died , plates were removed individually from the incubator and the number of live and dead worms recorded . For the lifespan assay , worms were initially treated as for the thermotolerance assay but without temperature upshift . A count was made of live and dead worms every day until all worms had died . Worms that had desiccated on the side of plates or died due to an exploding vulva were censored from the time of death . Worms were kept on NGM plates until day 1 of adulthood , at which point they were transferred to 500 μl S-medium containing serial dilutions of HB101 at final concentrations of 108 ( low ) , 109 ( medium ) or 1010 ( high ) cells / mL in 24-well tissue culture plates rocked on a nutator . At day 2 of adulthood worms were pulsed for 1 h with EdU resuspended in water and delivered at a final concentration of 0 . 4 mM and processed as detailed below . To label cells with the thymidine analogue EdU ( C10337 , Life Technologies , Grand Island , NY ) , worms were fed EdU-labeled E . coli . To prepare labeled E . coli , strain MG1693 was grown in minimal medium supplemented with glucose [75] and 75 μM EdU for pulse-chase experiments and 7 . 5 μM for continuous labeling experiments . When required , Fluoresbrite fluorescent microspheres ( 19507–5 , Polysciences , Warrington , PA ) were added to bacteria prior to seeding , at a 1:100 dilution . Immediately following seeding , plates were stored at 4°C . Plates were warmed to 20°C prior to use . Worms were kept for 0 . 5–8 h on EdU-labeled bacteria in the dark , returned to non-labeled bacteria for experiments that required a chase , and were fixed and processed as described [76] using 0 . 1 μg/ml DAPI to label DNA and 1:200 anti-PH3 antibody ( 9706 , Cell Signaling , Beverly , MA ) to label M-phase cells , and imaged at ~0 . 3 μm z intervals with LSM 710 or 780 confocal microscopes ( Carl Zeiss MicroImaging , Oberkochen , Germany ) , using a 63x objective . EdU continuous labeling image data were assayed manually for the presence of at least one EdU-positive cell in individual mitotic zones . EdU pulse chase data were analyzed as described in S1 Text . For whole germ line imaging , for counts of apoptotic cells detected with the CED-1::GFP reporter or foci detected with the RPA-1::YFP reporter , and for counts of cells in diplotene or diakinesis extruded germ lines were fixed and stained with DAPI , 0 . 16 μM Alexa 594-conjugated Phalloidin ( A12381 , Life Technologies ) , and 1:1000 anti-GFP ( for CED-1::GFP and RPA-1::YFP; ab5450 , Abcam , Cambridge , MA ) . Image stacks were acquired at 0 . 3–0 . 6 μm z intervals using a 63x objective . In some instances , several panels were imaged for each gonadal arm that were subsequently stitched [77] . RPA-1 foci were analyzed in the proximal meiotic region ( from the beginning of zone 5 and into 6 , as described [78] ) . The dataset was blinded by a user who copied renamed files from all datasets into a single directory . A second user scored each of the renamed image files by selecting 20 random cells in zones 5 and 6 using the DNA channel of the image and recording RPA-1::YFP foci within each of these cells using Parismi [52] . The Wilcoxon rank sum test as implemented by the R project or Matlab was used to test for significance of differences ( two-tailed test ) , unless otherwise stated . A generalization to interval-censored data of the Wilcoxon test implemented by the “interval” R package [79] was used to analyze labeling times in continuous EdU labeling experiments . Bootstrapping was performed using the “boot . ci” function of the “boot” R package [80] . The log-rank test was used for analysis of survival data .
|
Stem cell cycling is expected to be beneficial because it helps delay aging , by ensuring organ self-renewal . Yet stem cell cycling is best used sparingly: cycling likely causes mutation accumulation—increasing the likelihood of cancer—and may eventually cause stem cells to senesce and thus stop contributing to organ self renewal . It is unknown how self-renewing organs make tradeoffs between benefits and drawbacks of stem cell cycling . Here we use the C . elegans reproductive system as a model organ . We characterize benefits and drawbacks of stem cell cycling—which are keeping worms primed for reproduction , and reducing the number of future progeny worms may bear , respectively . We show that , under specific conditions of reproductive inactivity , stem cells switch back and forth between active and dormant states; the timing of these switches , whose genetic control we start delineating , appears random . This randomness may help explain why populations of aging , reproductively-inactive worms experience an increase in the variability of their reproductive capacity . Stochastic stem cell cycling may underlie tradeoffs between self-renewal and senescence in other organs .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"cell",
"physiology",
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"reproductive",
"system",
"senescence",
"gonads",
"caenorhabditis",
"cell",
"cycle",
"and",
"cell",
"division",
"cell",
"processes",
"animals",
"germ",
"cells",
"animal",
"models",
"physiological",
"processes",
"developmental",
"biology",
"caenorhabditis",
"elegans",
"oocytes",
"model",
"organisms",
"organism",
"development",
"sperm",
"research",
"and",
"analysis",
"methods",
"cell",
"cycling",
"animal",
"cells",
"aging",
"anatomy",
"cell",
"biology",
"ova",
"physiology",
"nematoda",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"organisms",
"genital",
"anatomy"
] |
2016
|
Intermittent Stem Cell Cycling Balances Self-Renewal and Senescence of the C. elegans Germ Line
|
The advent of two-photon microscopy now reveals unprecedented , detailed spatio-temporal data on cellular motility and interactions in vivo . Understanding cellular motility patterns is key to gaining insight into the development and possible manipulation of the immune response . Computational simulation has become an established technique for understanding immune processes and evaluating hypotheses in the context of experimental data , and there is clear scope to integrate microscopy-informed motility dynamics . However , determining which motility model best reflects in vivo motility is non-trivial: 3D motility is an intricate process requiring several metrics to characterize . This complicates model selection and parameterization , which must be performed against several metrics simultaneously . Here we evaluate Brownian motion , Lévy walk and several correlated random walks ( CRWs ) against the motility dynamics of neutrophils and lymph node T cells under inflammatory conditions by simultaneously considering cellular translational and turn speeds , and meandering indices . Heterogeneous cells exhibiting a continuum of inherent translational speeds and directionalities comprise both datasets , a feature significantly improving capture of in vivo motility when simulated as a CRW . Furthermore , translational and turn speeds are inversely correlated , and the corresponding CRW simulation again improves capture of our in vivo data , albeit to a lesser extent . In contrast , Brownian motion poorly reflects our data . Lévy walk is competitive in capturing some aspects of neutrophil motility , but T cell directional persistence only , therein highlighting the importance of evaluating models against several motility metrics simultaneously . This we achieve through novel application of multi-objective optimization , wherein each model is independently implemented and then parameterized to identify optimal trade-offs in performance against each metric . The resultant Pareto fronts of optimal solutions are directly contrasted to identify models best capturing in vivo dynamics , a technique that can aid model selection more generally . Our technique robustly determines our cell populations’ motility strategies , and paves the way for simulations that incorporate accurate immune cell motility dynamics .
Cellular motility and interactions underlie many processes in the immune response , including lymphocyte recirculation through blood and lymphoid organs , their interactions with cells presenting specific antigen , and relocation to the specific tissues where they engage in protective immunity [1] . In the last decade , two-photon microscopy has provided unprecedented insight into how immune cells move and interact in vivo [1 , 2] . Parallel to this , computational modeling and simulation techniques have been applied to exploring hypotheses of immune system function [3 , 4] , even simulating the effects of interventions [5 , 6] . Agent-based simulations ( ABS ) , wherein individual immune cells are simulated as discrete entities with their own state in a spatially explicit environment , have found widespread application in immunology , with far-ranging applications including: understanding granuloma development [7] , Payers patch development [8] , the search efficiency of lymphocytes in the lymph node [9 , 10] , the establishment and subsequent recovery from autoimmune disease [5] , and the mechanisms underlying cancer [11] . There is clear scope to combine detailed spatio-temporal two-photon microscopy data with spatially-explicit agent-based simulation to further understanding of how cellular motility integrates with other immune processes to impact health . An established body of research in ecology has demonstrated , however , the complexities of determining which models of motility best characterize a given dataset . In the Lévy walk model , an agent’s motility is described by a sequence of randomly oriented straight line movements drawn from a power-law , long-tailed distribution [12] . Hence , agent motilities are characterized by many relatively short movements punctuated by rare , very long movements . A diverse range of organisms have been described as exhibiting Lévy walk motility , including bacteria , honey bees , fruit flies , albatrosses , spider monkeys , and sharks [13 , 14] . T cells in the brains of Toxoplasma gondii-infected mice have also been shown to perform a Lévy walk [15] . Interest in the Lévy walk is in part due to theoretical work demonstrating it an optimal strategy for finding sparsely , randomly distributed targets [16 , 17] . However , subsequent work has cast doubt on Lévy walk’s apparent pervasiveness in nature , owing to methodological discrepancies in its identification [18 , 19] . The spatial motility of agents in both two- and three-dimensions is an intricate and nuanced phenomenon that cannot be well specified using only one metric . The mean squared displacement over time metric is frequently used to differentiate Lévy walk and Brownian motion characteristics in a dataset , yet models differing in key aspects of motility can produce identical measures [20 , 21] , e . g . , slow moving directionally persistent cells , or fast moving less-directional cells . To accurately simulate the motility dynamics of a biological dataset requires an appropriate motility model assigned appropriate parameter values , and evaluating putative parameter values requires simultaneous consideration of several complementary motility metrics . Here we evaluate several random walk models’ , including Brownian motion , Lévy walk , and several correlated random walks , capacities to capture the motility dynamics of lymph node T cells responding to inflammation and neutrophils responding to sterile laser injury of the ear pinnae . Each model is independently simulated , and those model parameter values that best align simulation and in vivo motility dynamics are determined through novel application of a multi-objective optimization ( MOO ) algorithm: NSGA-II [22] . Parameter estimation is performed through simultaneous consideration of three metrics of cell population motility: the distributions of translational and turn speeds observed across the population , and the distribution of meandering indices . The differences between simulation and in vivo distributions generated under each metric form objectives for the MOO algorithm . The resulting Pareto fronts generated under each model , representing parameter values delivering optimal trade-offs in performance against each metric , are contrasted to ascertain which model best captures the biology . Our random walk models are designed following a detailed analysis of which statistical distributions best fit a cellular population’s translational and turn speed data . Such assessment is complicated by inherent biases in imaging experiments , wherein fast moving and directionally persistent cells rapidly leave the imaging volume . Hence , slower , less directional cells are over-represented in in vivo datasets . It is unclear whether cells observed to differ in directional persistence and translational speed are a result of these biases , or whether these observations represent fundamental differences in cellular motilities . Our novel analytical approach fits a given statistical distribution to a population’s pooled translational ( or turn ) speeds , whilst segregating observations drawn from the distribution into groups that correspond to tracks in the in vivo dataset . This segregation reproduces the imaging experiment biases , therein discounting their confounding influence on the analysis . We find that cells comprising our in vivo datasets are genuinely heterogeneous , differing in their inherent translational speed and directionality . This finding could reflect intrinsic cellular characteristics , or may arise as features of the environment through which they migrate . In subsequent analysis , we find that translational and turn speeds in both in vivo populations are significantly negatively correlated , indicating that cells do not simultaneously perform very fast translational movements and turns . To investigate the significance of these two observations on leukocyte motility we designed four correlated random walk models that differentially include ( or exclude ) each . We then simulate each to evaluate the integrative impact of these features on overall motility dynamics . We determine that Brownian motion poorly reflects both our datasets . Lévy walk competitively captures directional persistence , but performs poorly on translational and turn speed metrics , underscoring the value of considering several motility metrics simultaneously . Interestingly , for neutrophils Lévy walk provides the most even balance of metric trade-offs of any model examined . Both T cell and neutrophil motility dynamics were better captured by CRWs simulating cells as heterogeneous , rather than homogeneous , populations . Capture of T cell dynamics was further improved by negatively correlating simulated cell translational and turn speeds , however this was not as evident for neutrophil data . We have provided here evidence , for the first time , that cells within both T cell and neutrophil populations exhibit a continuum of inherent directionalities and translational speeds . Further , we have shown that cells do not simultaneously perform very fast translational and turn movements . We have developed a novel framework to fit statistical distributions to cell translation and turn speeds whilst accounting for experimental bias . Thereafter , the manner in which these two components of motility combine to impact overall spatial exploration is analysed through a novel coupling of 3D agent-based simulation with multi-objective optimization . This latter framework for the first time calibrates and assesses putative motility models through simultaneous consideration of several motility metrics , accounting for trade-offs in performance against each . These frameworks provide the means to robustly analyse and accurately reproduce cellular motility patterns , as they explicitly reflect the constraints of in vivo data .
We hypothesized that our T cell and neutrophil cellular populations were statistically heterogeneous , comprising cells differing in their inherent directionalities and translational speeds . Accordingly , we observed varying median track translational and turn speeds within both cellular populations , Fig 1A and 1B . These distributions could reflect genuinely heterogeneous features between cells , or could represent statistical sampling artifacts arising from finite cellular observation durations within a finite spatial volume . We quantified this experimental bias , Fig 1C , S3 and S1F Figs . Median track translation speed was strongly negatively correlated with the number of times the cell was observed in the imaging volume , and median track turn speed was strongly positively correlated with number of observations . Together these data indicate that fast cells moving in a highly directional fashion quickly left the imaging volume . We sought to establish whether the perceived heterogeneous cellular characteristics ( Fig 1A and 1B ) represent a genuinely heterogeneous population , or arise from experimental bias , and which statistical distributions best describe these data . We devised a novel statistical approach to address this question ( S4 Fig , Methods and S1 Algorithm ) , wherein observations are drawn from given statistical distribution and grouped . The groups reflect the structure in which the translational ( or turn ) speeds observed in a cellular population come from specific tracks . Hence , we could analyse all observations as one pooled dataset or extract the median values across the groups . This structure exactly matches that of the in vivo dataset being analysed , with the number of groups matching the number of tracks , and the number of observations within each group matching that of each track . Further , we impose similar correlations between the number of observations in each group and the median observation value , therein reflecting the experimental biases present in the in vivo dataset . This is done by establishing the maximum number of observations of any track in the in vivo dataset , and initially populating each group with the same number of observations . Thereafter we iterate through each track in the in vivo dataset , and select a group from which to discard data such that track and group share the same number of observations , and the correlations between median observation data and number of observations are similar . This procedure is used to assess how well a given statistical distribution captures cellular translational ( or turn ) speed data , despite the experimental biases inherent in obtaining it . A successful capture must reproduce both the distribution of all translational ( or turn ) speeds pooled from all tracks ( S1A and S1B Fig ) , and how these are allocated into tracks yielding the distribution of median track characteristics ( Fig 1A and 1B ) . We assess a variety of statistical distributions , depicted in S5 Fig , including uniform , Lévy and Gaussian; these are termed ‘homogeneous’ as the same parameterized distribution is used to populate all groups with observations . We also assess a ‘heterogeneous’ Gaussian , wherein each group is populated by a bespoke Gaussian sub-distribution; hence , these groups are statistically heterogeneous , each is composed of observations drawn from a ( potentially ) unique Gaussian . A given statistical distribution is first fitted to the in vivo dataset’s pooled translational speeds ( or turn speeds , S1A and S1B Fig respectively ) , pooling all groups’ observations when performing the fitting . This is done 5 independent times for statistical rigour . We use each fitted solution to generate 100 datasets using the procedure outlined above , giving 500 datasets in total . For each of these , we contrast the groups’ median observation values with the tracks’ median translational ( or turn ) speed values using the Kolmogorov-Smirnov ( KS ) statistic . This yields 500 KS values for each statistical distribution we examine . The statistical distribution yielding lowest KS values best reflects the in vivo translational ( or turn ) speed dynamics; these are graphed as cumulative distribution functions in Fig 1E to 1H , explored below . Cellular turn dynamics are analysed using the same procedure , but additionally accounting for the maximum discernible rotational velocities for each track as determined by imaging experiment temporal resolution ( Methods , S1 Algorithm and S1 Table ) . T cell and neutrophil translational dynamics are better captured with a statistically heterogeneous Gaussian distribution than a homogeneous Gaussian distribution . When fitting distribution parameters against pooled in vivo translational speed data both statistical distributions performed well , Fig 1D , S6 and S7 Figs: KS values differentiating modeled and in vivo pooled translational speed data were low . However , median track translational speed data were better captured by the heterogeneous Gaussian distribution , Fig 1E and 1F , and S8 Fig . We also evaluated the capacity for Lévy distributions , the foundation of the Lévy walk , to reproduce in vivo translational dynamics . The Lévy distribution was competitive with the Gaussian distributions in capturing pooled translational speed data , Fig 1D , but was inferior in its capture of median track translational speed data , Fig 1E and 1F and S8 Fig . We determined that homogeneous and heterogeneous Gaussian distributions both accurately capture pooled turn speed data , Fig 1D , S9 and S10 Figs , but the heterogeneous Gaussian proved superior in reproducing in vivo median track turn speed distributions , Fig 1G and 1H and S8 Fig . We additionally evaluated a uniform distribution’s capture of turn speed dnymaics , which corresponds with Brownian motion and Lévy walk motility models where successive trajectories are uncorrelated . We determined that the uniform distribution provided a competitive reflection of in vivo pooled turn speed distributions , but was the worst of the three models in reproducing median track turn speed dynamics . We hypothesized that owing to physical constraints on rates of cytoskeletal remodeling cells are unable to perform both very fast translational movements and turns simultaneously . We confirmed this in both our datasets , Fig 2 . The Spearman’s correlation coefficient between cell turn speed and the median of the translational speeds recorded immediately before and after the turn was -0 . 29 and -0 . 27 for T cell and neutrophil datasets respectively . Collectively these data suggest that cells in both our T cell and neutrophil datasets are statistically heterogeneous: the distributions of varying median track translational and turn speeds reflect inherent differences in cellular speed and directionality , rather than sampling artifacts . Further , they suggest a trade-off between fast translational movement and large directional alterations . We next sought to investigate the significance of these observations by designing several correlated random walk models around the statistical distributions explored here , and evaluating their capture of our leukocytes’ spatial exploration through simulation . Through agent-based simulation we have assessed the ability of six random walk models to reproduce the motility dynamics of our T cell and neutrophil datasets ( full details in Methods ) : Brownian motion , Lévy walk and four correlated random walks ( CRW ) . Table 1 details how these random walk models are designed around the statistical distributions explored in the previous section ( see S5 Fig ) . The HomoCRW and IHomoCRW both represent cellular populations as statistically homogeneous: all cells draw translational speeds from the same homogeneous Gaussian distribution , and similarly for turn speeds . The HeteroCRW and IHeteroCRW models instead define bespoke , potentially unique distributions for each individual cell , rendering them statistically heterogeneous in inherent translational speed and directionality . The IHomoCRW and IHeteroCRW models impose an inverse correlation between translational and turn speeds , whereas HomoCRW and HeteroCRW do not . Each model was independently implemented in a 3D simulation , and subsequent calibration identified parameter values that align simulation with in vivo motility dynamics . Calibration was performed through multi-objective optimization [22] , therein simultaneously considering several metrics ( ‘objectives’ ) of cellular motility . A multi-objective approach is necessary as no single metric can fully specify the complexities of 3D motility . Three objectives are employed , each quantifying a specific difference between the motility profiles of the target in vivo dataset and a given model-parameter set simulation dataset respectively . A motility profile constitutes: the distribution of translational speeds observed across all cells at all time points pooled together; similarly for turn speeds; and the distribution of cell meandering indices , defined as a cell’s net displacement divided by its total distance traveled . These distributions are contrasted using the Kolmogorov-Smirnov ( KS ) statistic , therein forming the three calibration objectives . The meandering index was selected over alternatives such as mean squared displacement ( MSD ) for it’s ability to capture a distribution of heterogeneous cellular directionalities , which MSD does not; this choice is revisited in the Discussion . A calibration exercise yields a three-dimensional Pareto front comprising sets of putative model parameter values ( ‘solutions’ , S11 Fig ) . These solutions are Pareto-equivalent: no solutions offer an improvement in any objective without a worsening in another . We evaluate which models best reproduce in vivo motility by contrasting their respective Pareto fronts through three complementary analyses ( additional details in Methods ) . Firstly , the proportion of each Pareto front that is non-dominated by each of the others is ascertained ( S11 Fig ) ; a solution is dominated if there exists another with at least equal performance on all objectives and superior performance on at least one . Secondly , the best ( lowest ) 30 Λ values of each front are contrasted; a low Λ value reflects a solution with low mean and variance in its 3 objective KS scores . The best 30 Λ values represent those in the centre of the front , providing good performances on all objectives simultaneously . Lastly , the distribution of KS scores represented across each Pareto front for each objective are directly contrasted . Each model is independently calibrated three times against each in vivo dataset , the best solutions of which form a Pareto front for subsequent evaluation . Calibration is performed using NSGA-II [22] for 40 generations , comprising between 20 and 100 candidates per generation , a reflection of the number of model parameters and hence the difficulty of the calibration exercise ( see Methods ) . The performances of the very best solutions found for each motility model , that with the lowest Λ value , against in vivo data are shown in S12 to S23 Figs . The Pareto fronts of best calibrated solutions for Lévy walk outperform those of Brownian motion . All Lévy walk solutions are non-dominated by Brownian motion solutions in both datasets , Table 2 . For the T cell dataset all Brownian motion solutions are dominated by those of Lévy walk , and for neutrophil data only 7% of solutions are non-dominated . These patterns are reflected in the superior Λ values that Lévy walk solutions’ offer over those of Brownian motion ( Fig 3 ) . Brownian motion constitutes a particularly poor reflection of our T cell data , with only 7 tracks in the best Λ value solution remaining after applying a 27μm net displacement filter ( applied to reflect in vivo data preprocessing to remove anomalous imaging artifacts , see Methods ) , S12 Fig . Brownian motion and Lévy walk are inferior to all the CRW models in capturing T cell motility; ≥99% of their solutions are dominated in all cases ( Table 2 ) , and they provides the poorest Λ values found in any model ( Fig 3 ) . For the neutrophil dataset , Brownian motion is again suboptimal compared to all CRWs , in terms of both Λ values and non-domination . In contrast to T cell capture , however , Lévy walk solutions are not universally dominated by those of the CRWs , with as much as 72% of the Lévy walk Pareto front being non-dominated by that of HomoCRW . The HeteroCRW and IHeteroCRW models fare better , with >76% of their Pareto fronts non-dominated by that of Lévy walk , versus 32% and 45% vice versa . In terms of Λ value performance Lévy walk completely dominates all other models , Fig 3 . This is somewhat surprising given the competitive non-domination performances , and likely reflects how performances against each objective , explored below , are balanced in Lévy walk solutions; solutions with equal KS measures on each objective score lower Λ values ( Methods ) . Lévy walk accurately and competitively captures the directional persistence of in vivo data , but Brownian motion does not . Brownian motion delivers a narrow distribution of meandering index KS values , far inferior to other models’ values ( Figs 4 and 5 for T cells and neutrophils respectively ) . Lévy walk mirrors the performance of the best CRW model in capturing neutrophil meandering indices , Fig 5 , and is statistically indistinguishable from all CRWs in capturing T cell meandering indices except the IHeteroCRW which offers the best performance , Fig 4 . The meandering index reflects the interplay between cellular translation and orientation adjustments , and it is notable that Lévy walk does not perform particularly well in either of these , despite offering such competitive performance in capturing meandering indices . For turn speed KS values , Lévy walk and Brownian motion represent the poorest fits . Lévy walk offers a poor fit to T cell translational speeds , but exhibits a strangely narrow distribution of KS values for neutrophils; as shown in Fig 5 , Lévy walk’s worst translational KS values are better than the worst of the CRWs , however its best are worse than those of CRWs . Brownian motion poorly reflects the in vivo translational speed dynamics of both datasets . The previous section’s exploration of leukocyte heterogeneity found the Lévy distribution to poorly fit to leukocyte median translational speed data , and similarly , the uniform distribution to poorly fit median turn speed data; it is through these distributions that Lévy walk translational and reorientation adjustments are drawn . We analysed the present Lévy walk simulation’s capture of leukocyte heterogeneity , S24 Fig . Overall , Lévy walk performance is better than might be expected given the previous section’s results . It offers competitive or improved capture when contrasted with the HomoCRW and IHomoCRWs , but generally inferior to the HeteroCRW and IHeteroCRW , with the exception of median neutrophil turn speed data . In summary , Brownian motion is universally poor in capturing both T cell and neutrophil motility . Lévy walk is similarly poor in capturing T cell performance , but for neutrophils the situation is more complex . Though uncompetitive in turn speed capture , moderately so in translational speed capture , and being largely Pareto-dominated by the IHeteroCRW’s Pareto front , Lévy walk offers by far the best performance on Λ values . We theorise that there exists a portion of the neutrophil Lévy walk Pareto front comprising solutions with similar , low mean objective KS values , as this would yield low Λ values despite not dominating in any particular objective alone . We find that CRW models accommodating heterogeneous characteristics between cells better reflect in vivo data ( Table 2 ) . HeteroCRW Pareto fronts for both datasets are almost entirely non-dominated by the HomoCRW Pareto fronts , versus <7% vice versa . A similar trend is found when comparing the inverse CRW class of models , where IHeteroCRW solutions were largely non-dominated by IHomoCRW models . Here , however , the IHomoCRW was was 42% non-dominated on the neutrophil dataset , substantially higher than the 5% of the T cell dataset . The superiority of heterogeneous over homogeneous CRW models was also reflected through the best 30 Λ value distributions , Fig 3 . We find no overlap in Λ value distributions between HeteroCRW and HomoCRW models on either dataset , with the former providing superior values . On the T cell dataset IHeteroCRW is similarly superior , however neutrophil dataset IHeteroCRW and IHomoCRW Λ distributions overlap , with the former providing marginally superior values . To provide an intuition into the magnitude of the separation between heterogeneous and homogeneous CRW models’ Pareto fronts , S25 and S26 Figs provide three-dimensional plots of Pareto front solutions against each objective KS score . The separation between Pareto fronts is particularly large for HeteroCRW and HomoCRW on the neutrophil dataset , and IHeteroCRW and IHomoCRW on the T cell dataset . The better fit of heterogeneous over homogeneous models of cellular populations is reflected in performance on each objective ( Figs 4 and 5 ) . HeteroCRW yields a distribution of KS values statistically significantly lower than HomoCRW models for all three measures of cellular motility on both datasets , with the exception of T cell meandering indices where no statistically significant difference is observed ( Fig 4 ) . We similarly find IHeteroCRW to yield statistically significantly lower values than IHomoCRW on both datasets , with the exception of neutrophil meandering indices where IHomoCRW provides lower values ( Fig 5 ) . Our simulation studies further support our determination that our in vivo cellular populations are statistically heterogeneous , and that observed distributions of median track translational and turn speeds ( Fig 1A and 1B ) are not sampling artifacts . Our simulations impose the same experimental constraints as are present in vivo: finite observations of cells within a finite imaging volume . Despite not being used as criteria for model calibration , the heterogeneous CRW models best captured median track translation and turn characteristics , with the exception of neutrophil median track turn speeds ( S24 Fig ) . Analysis of both in vivo datasets revealed significant negative correlations between cellular translation and turn speeds ( Fig 2 ) . This correlation could impact cellular directionality , and hence meandering indexes , as subsequent fast translational movements would be directionally persistent . As such , we examined whether the inverse CRW formulation , which imposes this quality on simulated cells ( see S27 Fig ) , better reflects the in vivo data than the standard formulation . Inverse CRW models better capture T cell motility dynamics than the standard formulations , but the difference is moderate . 85% of IHomoCRW solutions are non-dominated by HomoCRW models , in contrast to 27% vice versa ( Table 2 ) . A larger disparity is found for IHeteroCRW and HeteroCRW models , with values of 95% and 10% respectively . Fig 3 reveals , however , that the magnitude of this dominance is marginal: the Kolmogorov-Smirnov statistic reveals a difference of 0 . 7 between HomoCRW and and IHomoCRW Λ value distributions , and no statistically significant difference between HeteroCRW and IHeteroCRW models . Given the Pareto-dominance of IHeteroCRW over HeteroCRW models this Λ value finding is surprising , and suggests that the dominance occurs on the periphery of the Pareto fronts; Λ values focus on the centre only . Corresponding analysis of model performances’ on each objective highlights translational speeds as the objective where IHeteroCRW outperforms HeteroCRW; no significant difference is observed on other objectives ( Fig 4 ) . IHomoCRW’s capture of T cell turn speeds is significantly better than HomoCRW’s , but we find no other statistically significant differences across objectives . The lack of statistically significant differences between inverse and standard model formulations on meandering index performance is surprising , given that it was specifically this objective that the inverse formulation was hypothesized to offer improvement in . Rather , the inverse formulation facilitates performance improvement on other objectives whilst maintaining a similar meandering index profile . S28 and S29 Figs provide three-dimensional plots of Pareto front solutions against each objective KS scores for standard versus inverse model formulations . IHomoCRW better captures neutrophil dynamics than HomoCRW , but this finding does not extend to heterogeneous CRW formulations . 98% of IHomoCRW solutions are non-dominated by the HomoCRW model , and 9% vice versa . Conversely , IHeteroCRW and HeteroCRW are largely Pareto-equivalent , where 67% of HeteroCRW solutions are non-dominated in contrast to only 50% of IHeteroCRW models . These findings are supported by Λ value distributions , where IHomoCRW yields substantially better values then HomoCRW , yet no significant difference is found between IHeteroCRW and HeteroCRW models ( Fig 3 ) . We find that the IHomoCRW models offers significantly better meandering index and translational speed values than the HomoCRW model , Fig 5 . HeteroCRW provides superior translational speeds to IHeteroCRW , but otherwise these two models are statistically indistinguishable .
The advent of two-photon in vivo cellular imaging techniques facilitates detailed examination of cellular motility and interaction . The resultant data permits identification of cellular motility strategies , which can be incorporated into broader immune simulations to understand the development and potential manipulation of the immune response . Determining which motility model best fits a biological dataset requires simultaneous consideration of several metrics of motility; three dimensional motility is too intricate a phenomenon to be fully specified in only one metric . Here we have evaluated the capacity of six random walk models , including Brownian motion , Lévy walk and four correlated random walks , to reproduce the motility dynamics of lymph node T cells and neutrophil datasets under inflammatory conditions . Our evaluation is made possible through the development of a novel simulation calibration methodology , where multi-objective optimization identifies parameter values that provide optimal trade-offs for a given model against several metrics of motility . We found that Lévy walk , an optimal strategy for finding sparsely randomly distributed targets [16 , 17] , and identified as the motility pattern of CD8+ T cells in Toxoplasma gondii-infected mouse brains [15] , does not optimally capture our T cell motility dataset . Its performance in capturing neutrophil motility was competitive with other models’ , performing well in some motility measures but poorly in others . We attribute our finding to the simultaneous consideration of multiple motility metrics . Lévy walk’s best meandering index performance is competitive with other models’ , and as such optimization on that metric alone might highlight Lévy walk as an optimal model ( Figs 4 and 5 ) . It is only when performance against this metric balanced with others that Lévy walk’s quality of capture diminishes . It is the micro-level details of leukocyte motility that Lévy walk fails to capture , given their straight-line directional persistence punctuated by uniformly random reorientations of direction . This is supported by our fitting of statistical distributions to cell translational and turn speed data , where Lévy and distributions and uniform distributions ( corresponding with uncorrelated cellular trajectories ) poorly captured the data . We do not discount the possibility that hybrid strategies , where micro-level correlated random walks are subject to macro-level directional persistence captured by Lévy walks [16] , might better reflect in vivo data . We determined our T cell and neutrophil populations to be statistically heterogeneous in their inherent translational speed and directional persistence . We devised a novel approach for fitting statistical distributions to either translational or turn speed data whilst accounting for imaging experiment bias . Our approach ruled out the possibility that these heterogeneous qualities arise as sampling artifacts from observing cells for finite durations within a finite imaging volume; cells that simply happen to be moving fast in a persistent direction as they crossed the imaging volume would give the illusion of being statistically distinct from cells that happened to be moving slowly with little directional persistence at time of observation . We quantified this bias , and found strong negative correlations between median cell track translational speed and observational duration in both datasets . Likewise , we found strong positive correlations between median turn speeds and observational duration . Our statistical distribution fitting approach uses a given distribution to reproduce in vivo data , capturing the same number of tracks , the same observations per track , and imposing similar correlations between median track feature and number of observations . A heterogeneous statistical distribution , wherein each track’s data is generated from a bespoke , potentially unique , Gaussian distribution best reflected our in vivo data in all cases . Homogeneous distributions , wherein the same parameterized distribution was used to model all tracks’ data could not reproduce the heterogeneity observed in vivo , despite accounting for the experimental biases . We confirmed the significance of this cellular heterogeneity through agent-based simulation , which , rather than separately exploring translation and turn dynamics , integrates them to produce 3D tracks . CRW models representing a continuum of heterogeneous qualities within a cellular population proved superior to treating all cells as statistically equivalent . This finding supports the conclusion that leukocytes differ in their inherent rotational and translational speeds . We discount the alternative conclusion that these more complex models capture nuances ( rather than general qualities ) of the training data set , as levels of over-fitting were monitored and deemed acceptable ( see Methods ) . The large sizes of our datasets , 751 T cells & 1017 neutrophils ( see Methods ) , further suggest that these heterogeneous qualities do not result from small sample sizes . Banigan et al . first described a heterogeneous population of CD8+ T cells in uninflamed lymph nodes , characterizing them as two distinct homogeneous sub-populations , 30% of which perform Brownian motion and the remainder a persistent random walk , all of them drawing velocities from the same distribution [21] . In contrast , here we identified an entire continuum of inherent cellular translation and turn characteristics , in both neutrophils in the mouse ear pinnae , and lymph node T cells , both under inflammatory conditions . Analysis of both our T cell and neutrophil datasets revealed strong inverse correlations between cell translational and turn speeds: cells do not simultaneously perform fast translational movements and large reorientations . This has been shown previously for neutrophils [23] , but we are unaware of any such finding in T cells . We again used simulation to evaluate the impact of this characteristic on overall motility , devising CRWs that impose this negative correlation ( ‘inverse’ CRW ) and contrasting their capture of in vivo dynamics with those that do not . We found inverse CRWs to better capture T cell data than standard formulations , in particular improving capture of translational speeds when coupled heterogeneous qualities . In neutrophil data , an inverse homogeneous CRW substantially improves upon standard homogeneous CRW performance , yet inverse and standard heterogeneous CRW models are indistinguishable . This finding could originate from constraints on the cytoskeleton remodeling processes [24] . Alternatively , cellular dynamics can be explained through the configuration of obstacles in the environment [25]; our findings might represent features of the environment rather than the cell , where cells must slow in order to move around an obstacle . We conclude that the inverse heterogeneous CRW models best capture leukocyte motility: their corresponding Pareto fronts are non-dominated by any other model ( Table 2 ) , with one exception where IHeteroCRW and HeteroCRW were indistinguishable . Previous lymphocyte modeling efforts have incorporated explicit cellular arrest phases between periods of fixed speed , straight-line motility [15 , 26] . Our in vivo datasets do not record cells as being stationary , or moving in straight lines ( S1A and S1B Fig ) . As such , we have explored CRW models that explicitly capture distributions of translational and turn speeds . Other work has focused on modeling lymphocytes as point-processes confined to the lymph node reticular network [27] , explicitly modeling cellular morphology [25 , 28] , and conceptualizing cell trajectories as features of environmental obstacles [25] . The possibility of calibrating the configuration of an environment by proxy of the resultant cellular motility is intriguing . Our multi-objective optimization framework is independent of the motility paradigm and could be more broadly applied in these contexts . We opted to employ three objectives in our multi-objective approach , based on the pooled translational speeds of all cells across all time points into a single distribution , similarly for turn speeds , and track meandering indices . We consider this the minimum required to accurately specify motility , capturing how cells move translationally through space , how subsequent trajectories are correlated , and how these two aspects integrate to define overall spatial coverage . Multi-objective optimisation can accommodate more objectives , and hence additional motility metrics could be incorporated ( or substituted ) . In particular , we believe there is merit in studying how recent , more sophisticated motility metrics might be incorporated into our framework [20 , 21] . It is practical , rather than technical , considerations that limit the number of objectives one can use: in our experience the number of Pareto front members tends to increase with each additional objective , and more objectives constitute a more complex problem which can require greater computational effort to solve to a similar extent ( e . g . , as measured through objective KS values ) . Convention in multi-objective optimisation dictates that one choose objectives which are not correlated with one another; to do so increases the complexity of the optimisation problem whilst providing little benefit in capturing better quality solutions . Candidates for additional objectives might include the median track translational or turn speed distributions , however we note that for our favoured motility model , the inverse heterogeneous CRW ( IHeteroCRW ) , these characteristics are well captured despite not being explicit criteria in model calibration ( S17B , S17C , S23B and S23C Figs ) . We consider it essential to include an objective capturing how translational and turn characteristics integrate to dictate spatial coverage . In this regard we employ the meandering index , but possible alternatives include mean squared displacement ( MSD ) or spatial volume explored . Each of these introduces some bias , and hence the decision is somewhat arbitrary . For instance , meandering indices tend towards 1 for short tracks; S2A and S2B Fig quantify this for our in vivo datasets . We note , however , that our simulation approach imposes the same experimental constraints as exist in vivo , and all our data are processed through the exact same analytical pipeline ( see Methods ) . As such the same biases arise in all our data , providing a fair comparison between in vivo and simulation experiments . It is notable that similar correlations and scatter plots occur between track duration and meandering index for our simulation and in vivo datasets ( T cells: S2A , S12H to S17H Figs; neutrophils: S1C Fig ) MSD has been used extensively to discriminate between motility models , however , in addition to the known issues with this metric [20 , 21] , our characterisation of statistically heterogeneous populations prompted our choice of the meandering index , which neatly captures the distribution of cellular directional persistencies and which the MSD does not . We have performed a pilot study substituting MSD in place of the meandering index , calibrating the IHeteroCRW model against neutrophil data ( details in Methods ) . Capture of pooled translational and turn speed data formed the remaining two objectives . S30 Fig contrasts IHeteroCRW’s capture of neutrophil motility under each calibration scenario . As to be expected , the meandering index and MSD metrics were best aligned when used directly as a calibration objective . Pooled translational speed data was best captured using the meandering index , and turn data capture was statistically indistinguishable . Interestingly , median track translational speeds were better captured using the meandering index , and median track turn speeds through MSD ( neither were used in driving calibration ) . The best single solution arising from the MSD calibration is shown in S31 Fig , and can be contrasted with that of meandering index calibration , S23 Fig . Again , the results are remarkably similar , with the exception that using the meandering index better captures median track translational speeds and correctly captures the in vivo MSD , whilst calibrating with MSD poorly captures in vivo meandering indices . The similarities in these data support our belief that both meandering index and MSD capture similar aspects of motility when coupled with metrics of pooled translational and turn speed data , as in the current context . Banigan et al . have proposed metrics capturing displacement probability densities , and displacement autocorrelations for given time intervals [21] . We consider these metrics more statistically robust than either the meandering index or MSD , and have calculated displacement autocorrelation measures for our leukocyte and modeled datasets ( S32 and S33 Figs ) . However , given our focus on cellular heterogeneity , captured in both the data spread at each time interval and how individual cells perform across intervals , it is not clear how to integrate such high dimensional data into an objective to be used in the present calibration framework . This we highlight as meritorious further work . We note that the IHeteroCRW model generally deemed superior by our present methodology also provides a close qualitative alignment with in vivo displacement autocorrelation data . Our novel method for contrasting putative models , and therein parameterizing them , has a valuable role to play in the development of biological simulations . The construction of simulations which demonstrably capture biological systems has received recent attention [29] . This resulted in a process through which assumptions underpinning the abstraction of key biological components and processes into a conceptual model and thereafter a software implementation are explicitly captured [30] . A complementary technique , borrowing from safety critical systems engineering , decomposes a claim such as “This simulation is an adequate representation of the biology” into sub-claims against which evidence is cited [31] . Additionally , statistical analyses quantifying the impact of biological uncertainty on simulation results by highlighting critical parameters and pathways have been developed [32 , 33] . Together these techniques support the development and interpretation of biologically meaningful simulations . The novel multi-objective optimization approach developed here is complementary in helping select between competing abstractions of the biology by providing numerical evidence of improved capture . Whilst there exist established model selection techniques such as the Akaike Information Criterion and Schwarz criterion [34] , it is unclear how to apply them over multiple metrics of biological capture , as in the present case . A strength of both the Akaike Information Criterion and the Schwarz criterion is their consideration of model parameter number when determining the most appropriate model . This feature is currently lacking from our multi-objective approach , and we see value in further work investigating how to reconcile these approaches . In the context of our present simulation work , the model with the most parameters ( inverse heterogeneous CRW ) yielded either the outright or joint best capture of the biology . We note , however , that this model’s parameters and the features they represent are not arbitrary , but are instead biological driven: they were found to be present in both our in vivo datasets . Simulation parameterization presents another challenge in biological simulation . The required biological data do not always exist as the corresponding experiments either have not or cannot be performed , and simulation’s abstractive nature complicates their adoption . Existing parameterization approaches include exhaustive search of all possible parameter value combinations [35 , 36] , maximum-likelihood estimation [15] , various forms of regression [37] , and genetic algorithms [38] . These techniques do not always scale to simulations with many parameters , and none accommodate the simultaneous consideration of several metrics of simulation’s capture of the biology as our present MOO-based approach does . We have developed and demonstrated a technology that more robustly determines which motility strategies best characterize a given biological dataset . Furthermore , it can implicitly embed these motility dynamics in a simulation , therein enabling more accurate simulations of immune response development . The intricate and nuanced motility patterns that our method reproduces are important , as it is at this scale that two nearby cells either contact or not , and these interactions can have a profound downstream influence on the immune response . Our approach can be used to characterise and quantify , in detail , how various factors impact and manipulate cellular motility , such as was done through inhibition of LFA-1 affinity and avidity regulation in T cells [39] .
All procedures involving mice were reviewed and approved by the Garvan/St Vincents Animal Ethics Committee ( AEC ) . The AEC fulfills all the requirements of the National Health and Medical Research Council ( NHMRC ) and the NSW State Government of Australia . Neutrophil data was obtained using in vivo two-photon microscopy of ear pinnae in anesthetized C57/BL6 mice . Neutrophils were recruited in response to sterile needle injury and neutrophil migration was recorded and analyzed following the induction of a small sterile laser injury as was described previously [40] . Neutrophils were visualized with the aid of Lysozyme M fluorescent reporter . The analysis of lymphocyte motility fluorescent lymphoid cells were adoptively transferred and cell migration was visualized 24 hours later in explanted cervical lymph nodes perfused with warmed and oxygenated medium . Inflammation was induced using either S . aureus bioparticles or ovalbumin in Sigma adjuvant . Two-photon imaging was performed using an upright Zeiss 7MP two-photon microscope ( Carl Zeiss ) with a W Plan-Apochromat 20′/1 . 0 DIC ( UV ) Vis-IR water immersion objective . High repetition rate femtosecond pulsed excitation was provided by a Chameleon Vision II Ti:Sa laser ( Coherent Scientific ) with 690-1064nm tuning range . We acquired 3μm z-steps at 512×512 pixels and resolution 0 . 83μm/pixel at a frame rate of 10 fps and dwell time of 1 . 27μs/pixel using bidirectional scanning . Neutrophil dataset z-depths were 180μm , and T cell dataset z-depths ranged from 150 to 220μm . Both datasets were cropped using Imaris software to correct for tissue drift as needed . Raw image files were processed using Imaris ( Bitplane ) software . A Gaussian filter was applied to reduce background noise . Tracking was performed using Imaris spot detection function to locate the centroid of cells and x , y and z coordinates of each spot were exported together with track ID and time interval information . The T cell calibration data is pooled from 9 individual imaging datasets , comprising a total of 751 cells tracked for a total of 20424 spots , yielding a mean of 27 spots per track . The neutrophil dataset comprises data pooled from 6 individual imaging datasets , totaling 1017 cells encompassing 24619 spots , a mean of 24 spots per track . The T cell experiments were conducted for around 30 min with time-series data recorded every 35 seconds , and for 45 min with time samples every 45 seconds for neutrophil data; exact figures are given in S1 Table . Several statistical distributions , graphically depicted in S5 Fig , were independently fitted to a given dataset: either cellular translational or turn speed data . This was performed for both our T cell and neutrophil datasets independently of one another . A graphical overview of our method is given in S4 Fig . We obtain a Lévy distributed random variable as follows , adapted from [41]: L ( α , β ) = β sin ( α X ) cos ( X ) 1 / α cos ( ( 1 - α ) X ) Y ( 1 - α ) / α ( 1 ) Where random variable X has uniform density on the interval [−π/2 , π/2]; Y has unit exponential density , generated as Y = − lnZ where Z is uniformly distributed over [0 , 1]; and β is a scaling factor . L is symmetrical around 0 and hence we take the absolute value , represented as |x| . A ‘homogeneous Gaussian’ distributed variable G ( μ , σ2 ) has mean μ and standard deviation σ2 . It is homogeneous in that the same parameterized Gaussian is used to represent all cells’ translational ( or turn ) values . In contrast , a ‘heterogeneous Gaussian’ distribution comprises a bespoke Gaussian G i ( μ i , σ i 2 ) for each cell i in the dataset . The mean μi and standard deviation σ i 2 of Gi are themselves drawn from Gaussian distributions; this is done once at Gi’s creation , and the values are maintained throughout Gi’s use thereafter . Hence , a heterogeneous Gaussian is formulated as G i ( μ i = G ( μ M , σ M 2 ) , σ i 2 = G ( μ S , σ S 2 ) ) , and has parameters μM , σ M 2 , μS and σ S 2 . U ( λ ) represents a uniformly distributed random variable over the range ( 0 , λ] . The parameters describing each statistical distribution are shown in Table 3: To evaluate the capacity for a given statistical distribution , D , to reproduce an in vivo dataset’s translational data we create an artificial dataset of similar structure . Values are drawn from D , and allocated into groups . There is one group for each track in the in vivo dataset , and initially each group contains as many observations drawn from D as the maximum number of observations found in any in vivo track . S1 Algorithm , in the supplementary data , discards observations from each group such that the number of observations in each group exactly matches the number of observations in a specific in vivo track . The observations to be discarded from each group are chosen such that the correlation between the number of observations in groups and the median observation values of those groups align with the correlations found for in vivo tracks . In this manner , the artificial dataset generated by D reflects the experimental bias inherent in the in vivo data . The pooled observational data , and the median observation values amongst the groups are then extracted , and contrasted with in vivo translation or turn data being analysed as follows . Let T represent the target data , be it either translational or turn speed data from one of our datasets , to which a given statistical distribution is to be fitted . First D is fitted against the pooled data T , that is , all the translation/turn observations pooled into one distribution . Fitting is performed using the python scipy . optimize . minimize method , using the ‘Powell’ solver , on the basis of minimizing the Kolmogorov-Smirnov ( KS ) statistic between pooled T data and pooled data generated using D in S1 Algorithm . This is performed 5 independent times , the results of which are shown in S6 , S7 , S9 and S10 Figs . Upon the conclusion of each fitting exercise , 100 further datasets are generated using the fitted D . We quantify how well each dataset captures the median track data in T using the KS statistic , yielding a total of 500 KS values for each D . Contrasting these 500 KS values reveals which statistical distribution best captures T , with low values indicating a better capture . The best alignment for each model on each in vivo dataset is shown in S8 Fig . We highlight that this procedure does not attempt to reproduce cellular motility in space , which is an emergent product of how translational and turn movements are integrated . Rather , it determines which distributions best capture translational and turn data independently of one another , and assess whether cells are heterogeneous in these characteristics . We design several random walk models based the distributions investigated here , and assess their capture of cellular motility in space through 3D agent-based simulation , as detailed in the Sections that follow . The six random walk models explored in this paper are detailed below . The models are constructed around the statistical distributions described above , and illustrated in S5 Fig . Table 1 summarizes which statistical distributions are employed in each random walk model , and how . The random walk models are simulated over time , and as we adopt the notion Dt to indicate a value drawn from randomly distributed variable D at time t . The random walk models are implemented in a discretized time , three dimensional continuous space agent-based simulation wherein cells are implemented spheres that cannot overlap . Only cells residing within a 412×412×100μm volume are tracked , replicating in vivo experimental conditions . T cell simulation state was updated and recorded for downstream analysis every 30s , and simulation were executed for 30min of simulated time . The corresponding neutrophil figures are 45s and 50min . These values were selected to broadly mirror in vivo experiments , as described in S1 Table . Note that Lévy walk simulation states were updated every 3s instead , owing to the variable cell run-durations of this model , however simulation state was still recorded every 30s and 45s as with other models . Both simulated and in vivo data undergo the same motility analysis , based on time series tracked cell spatial locations sampled every Δt seconds . For each time point ti , the vector describing the movement of a cell to its current location is calculated , and termed di . The displacement and translational speed over vector di are calculated . A cell’s turn speed at time ti is calculated as the angle between vectors di+1 and di divided by Δt . The largest measurable turn angle is 180° , and conversion into turn speeds ( °/min ) depends on the time step . Simulation time steps , 30s for T cells and 45s for neutrophils , correspond with maximum turn speeds of 360 and 240°/min respectively . These figures match the maximum discernible turn speeds for the in vivo datasets . However , the maximum discernible turn speed for each experiment within a dataset will differ with the time step ( see S1 Table ) , and this could represent an artifact for our calibration experiments . Given the majority of recorded turn speeds lie well below the maximum values ( S1B Fig ) we believe the influence of this discrepancy on calibration experiments to be minor , however we acknowledge its existence . A cell’s meandering index is defined as the net displacement from its first to last observed locations divided by its total distance traveled . This yields a value between 0 and 1 , respectively indicating the extremes of a cell finishing where it started or traveling in a straight line . Cells with total displacements <27μm are excluded from the analysis to avoid artifacts introduced by the sessile contaminating cell types such as dendritic cells , or cells that are dead or dying . This same displacement threshold is also applied to simulation data to ensure fair comparisons . The figure of 27μm was derived empirically using Imaris software , and represents an optimal trade-off for removing unwanted artifacts whilst minimizing the exclusion of motile T cells and neutrophils . The motility profile for a dataset , which typically constitutes several replicate experiments , comprises the following metrics . All translational speeds for all cells are pooled together to form one distribution . A similar pool of all cell turn speeds is constructed . All cell meandering indexes are pooled together into one distribution . Only these three metrics are used in simulation-based motility model calibration and evaluation . The following additional metrics are also derived , but not used in calibration or evaluation . We construct distributions of median track translation and turn speeds . Mean squared displacement ( MSD ) over time interval plots are produced . Displacement data for a given time interval is extracted from anywhere in the time-series , i . e . , time intervals are not absolute from time zero . Time intervals of 0 to 25% of the maximum track length are investigated . Slopes for MSD plots are calculated using linear regression . Displacement autocorrelation was calculated as in [21] . Each of the six models is implemented in simulation in turn , and then independently calibrated against each of the in vivo datasets . Calibration is performed using NSGA-II [22] , a multi-objective optimization algorithm based on a genetic algorithm that uses Pareto fronts to track candidate solutions representing the best trade-offs found to date with respect to each objective . NSGA-II is an elitist algorithm , meaning that a subsequent generation’s population is composed of the best solutions found to date: the solutions comprising the Pareto front . If the Pareto front comprises more members than the population size , a subset composed of those Pareto members having the largest fitness differences between their immediate neighbours summed for all objectives is selected , a strategy intended to promote full coverage of the Pareto front . If the Pareto front comprises fewer members than the population size then members of the next front ( those dominated by only one other solution ) are selected in the same manner , and so on until the entire population has been selected . New solutions are generated through blended crossover of their two parents , coupled with Gaussian mutation using the standard normal distribution . These evolutionary operators correspond to the Inspyred python package implementation of NSGA-II . For further details on NSGA-II we refer to the reader to [22] . Candidate solutions represent putative model parameters . Evaluation of a solution entails executing ten replicate simulations with the parameters it represents , and generating a motility profile from the pooled results . This motility profile is contrasted with that of the in vivo dataset: the Kolmogorov-Smirnov ( KS ) difference between the motility profiles’ distributions of cell translation ( S1A Fig ) and turn speeds ( S1B Fig ) , and meandering indices ( S1C Fig ) together form three objectives . A perfect simulation representation of an in vivo data set would yield a KS value of 0 for each objective . In reality , no random walk model , by virtue of being an abstract model , will likely achieve this . Instead , some disparity in at least one metric will exist . The use of Pareto fronts accommodates trade-offs between metrics; two solutions are Pareto-equivalent if neither provides better alignments with in vivo data across all measures . An individual calibration is performed for a maximum of 40 generations of the genetic algorithm , for all models . Calibration is terminated before 40 generations only if over-fitting , as described below , is detected . The number of candidates in each generation is scaled with the number of model parameters , thereby reflecting the complexity of the problem , as shown in Table 4: We avoid over-fitting models , wherein calibrated solutions represent the nuanced stochastic-sampling-derived features of the data rather than its general qualities , by dividing in vivo datasets into training ( 70% of cell tracks ) and validation sets ( 30% ) , as is standard machine learning practice [37] . Each putative model parameter set is independently evaluated against both training and validation datasets , and two Pareto fronts , representing the best solutions found with respect to each , are maintained throughout calibration . Progression of candidate solutions through subsequent generations is determined through performance against the training dataset alone . The over-fittedness of the population is defined as the proportion of training dataset Pareto front solutions that are not also members of the validation dataset Pareto front . Calibration is stopped when either the maximum number of generations have been run , or the over-fitted metric >0 . 8 . The model assessments reported here are made on the basis of validation dataset Pareto front solutions . We note that in no cases were any calibration efforts terminated prematurely on the basis of over-fitting , but over-fitted scores of around 0 . 6 were not uncommon . Calibration produces a Pareto front comprising those parameter values yielding the best reflections of the in vivo dataset . By contrasting Pareto fronts produced by two different models , that which is most capable of reproducing the motility of in vivo cells is ascertained . For a given model and in vivo dataset ( T cell or neutrophil ) , calibration is performed three times . One overarching Pareto front is then generated from the best solutions generated under each exercise , and is used in model evaluation . Three complementary analyses are performed when contrasting two models . First , the proportion of each models’ front that is non-Pareto-dominated by the other is calculated . If two models are exactly equal in their capture of the biology across all objectives , then these values should be 100% for each . If the two values are equal , but not 100% , then the models are still considered equal reflections of the biology overall , but they differ in how well they reflect particular objectives . Pareto front sizes are reported alongside these proportions , to highlight where high or low values simply reflect fronts containing few or many solutions . Second , we contrast the best ( lowest ) 30 Λ values found within a Pareto front using the Kolmogorov-Smirnov statistic ( Fig 3 ) . The Λ function , defined below for a candidate m , delivers low values to solutions having low mean objective KS values with small variance . Hence , it selects those solutions that perform well , and equally well , on all objectives . Λ ( m ) = α · K S ¯ ( m ) 2 + ∑ o ∈ Ω K S o ( c ) - K S ¯ ( m ) 2 ( 8 ) K S ¯ ( m ) represents the mean objective KS score for member m , Ω represents the set of objectives and KSo ( m ) represents the KS scores for member m against objective o . The coefficient α can be used to prioritize mean or variance terms , a problem specific decision; a value of α = 1 is used throughout this manuscript . S35 Fig depicts how Λ values vary in a hypothetical scenario comprising two objectives , under different values of α . Lastly , the distribution of scores for each objective generated under each Pareto front are contrasted , thereby highlighting how well each model captures each motility characteristic . These are shown in Figs 4 and 5 . The distributions are plotted on the left of these figures and are statistically contrasted using the Kolmogorov-Smirnov statistic , the values of which are given in the tables on the right of these figures . Experiments where the meandering index was replaced with mean squared displacement ( MSD ) as an objective for multi-objective optimisation used the same experimental setup as reported above . The MSD calibration objective operates by taking the absolute difference between the MSD linear regression slopes generated for candidate solution and neutrophil dataset as reported above . Two remaining calibration objectives are constructed from KS statistics applied to pooled translational and turn speed data , as reported above . Calibration was performed three independent time using 100 candidates for 40 generations , with an overfitting termination threshold of 0 . 8 . The best solution from the MSD-based calibration exercise , reported in S31 Fig , is that with the lowest sum of objective values . The Λ function described above is inappropriate in this context , as the MSD objective is not based on the KS statistic . Hence , is it nonsensical to take their mean value . The 3-dimensional continuous space simulation is written in Java , using the MASON simulation framework library [42] . We use the Inspyred implementation of NSGA-II , written in Python , to perform calibration . Kolmogorov-Smirnov statistics , and their associated p-values , are determined using Python’s scipy . stats . ks_2samp module . The statistical modeling of cellular translation and turn speed dynamics was performed using python , and its numpy and scipy packages . The 3D agent-based simulation and multi-objective optimisation software we developed for this manuscript is distributed under version 3 of the GNU General Public License in the S1 Software ZIP file ( the third party libraries we employ will need to be acquired separately from their respective sources for licensing reasons ) .
|
Advances in imaging technology allow investigators to monitor the movements and interactions of immune cells in a live animal , processes essential to understanding and manipulating how an immune response is generated . T cells in the brains of Toxoplasma gondii-infected mice have previously been described as performing a Lévy walk , an optimal strategy for locating sparsely , randomly distributed targets . Determining which motility model best characterizes a population of cells is problematic; multiple metrics are required to specify as intricate and nuanced a process as 3D motility , and the tools to evaluate model-parameter combinations have been lacking . We have developed a novel framework to perform this model evaluation through simulation , a popular tool for exploring complex immune system phenomena . We find that Lévy walk offers an inferior capture of our data to another class of motility model , the correlated random walk , and this determination was possible because we are able to explicitly evaluate several motility metrics simultaneously . Further , we find evidence that leukocytes differ in their inherent translational and rotational speeds . These findings facilitate more accurate immune system simulations aimed at unravelling the processes underpinning this critical biological function .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"blood",
"cells",
"cell",
"motility",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"classical",
"mechanics",
"fluid",
"mechanics",
"immunology",
"mathematical",
"models",
"simulation",
"and",
"modeling",
"mathematics",
"brownian",
"motion",
"statistics",
"(mathematics)",
"stochastic",
"processes",
"neutrophils",
"research",
"and",
"analysis",
"methods",
"statistical",
"distributions",
"random",
"walk",
"white",
"blood",
"cells",
"animal",
"cells",
"mathematical",
"and",
"statistical",
"techniques",
"t",
"cells",
"fluid",
"dynamics",
"probability",
"theory",
"continuum",
"mechanics",
"physics",
"cell",
"biology",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"physical",
"sciences",
"statistical",
"data"
] |
2016
|
Leukocyte Motility Models Assessed through Simulation and Multi-objective Optimization-Based Model Selection
|
Subtelomeres are duplication-rich , structurally variable regions of the human genome situated just proximal of telomeres . We report here that the most terminally located human subtelomeric genes encode a previously unrecognized third subclass of the Wiskott-Aldrich Syndrome Protein family , whose known members reorganize the actin cytoskeleton in response to extracellular stimuli . This new subclass , which we call WASH , is evolutionarily conserved in species as diverged as Entamoeba . We demonstrate that WASH is essential in Drosophila . WASH is widely expressed in human tissues , and human WASH protein colocalizes with actin in filopodia and lamellipodia . The VCA domain of human WASH promotes actin polymerization by the Arp2/3 complex in vitro . WASH duplicated to multiple chromosomal ends during primate evolution , with highest copy number reached in humans , whose WASH repertoires vary . Thus , human subtelomeres are not genetic junkyards , and WASH's location in these dynamic regions could have advantageous as well as pathologic consequences .
Human chromosome termini are unusual in their sequence content and frequency of structural rearrangement . Their ends are capped with telomeres that protect chromosomes against degradation and fusion [1] and that can exert a silencing effect on nearby genes [2 , 3] . Patchworks of large DNA segments duplicated on various subsets of chromosomes lie just proximal of telomeres [4] . These subtelomeric regions comprise less than 0 . 1% of the human genome , but account for over 40% of interchromosomal duplicates in the genome assembly that formed since human and chimpanzee diverged . Other lines of evidence further support the notion that subtelomeres are hotspots of DNA breaks and repair [5–7] . While subtelomeric dynamics might disrupt gene function , they can also fuel rapid changes in subtelomeric gene repertoires . Subtelomeres of yeast and the malaria parasite , Plasmodium , harbor genes with important roles in adaptive processes [8 , 9] . Human subtelomeric genes vary in copy number and chromosomal distribution and include odorant and cytokine receptors , homeodomain proteins , secretoglobins , and other genes of unknown function [4] . Demonstrations of expression from more than one subtelomeric location for several human genes [4] suggest that interindividual variation in subtelomeric gene repertoires might underlie some phenotypic differences among humans . Indeed , subtle chromosomal rearrangements involving subtelomeres and neighboring chromosome-specific regions are detected in 3%–5% of individuals with unexplained mental retardation or developmental disorders [10] . However , no essential gene has yet been identified within human subtelomeres . Here , we focus on the most telomerically duplicated human genes ( currently arbitrarily named MGC52000 ) , which heretofore had unknown function . A truncated copy of MGC52000 was first annotated in the pseudoautosomal region of Xqter/Yqter and called CXYorf1 [11] . A murine ortholog , orf19 , encoding a 475-amino acid protein with similarity to proteins of unknown function from Drosophila melanogaster and Caenorhabditis elegans , was subsequently identified , but no motifs that would disclose its biological function were reported [12] . The C . elegans ortholog ( Y48E1B . 1 ) was recently named ddl-2 for its involvement in life-span regulation and reported to have several proline-rich domains [13] . We now show that MGC52000 encodes a new member of the Wiskott-Aldrich Syndrome Protein ( WASP ) family and is conserved from Entamoeba to human . Known WASP family members participate in cytoskeleton reorganization and signal transduction by acting as effectors of Rho-GTPases and polymerizing actin via the Arp2/3 complex [14 , 15] . They are involved in cell motility , phagocytosis , and cytokinesis in diverse processes such as embryogenesis , angiogenesis , inflammatory immune response , microbial infection , and cancer metastasis [16–18] . Five WASP family members are currently known in mammals , and they fall into two subclasses , WASP/N-WASP and SCAR/WAVE [16 , 18] . We have renamed the human subtelomeric MGC52000 genes WASH , for Wiskott-Aldrich Syndrome Protein and SCAR Homolog . Our comparative and phylogenetic analyses of the WASP family reveal that WASH proteins define a new subclass in the WASP family . We experimentally confirm that human WASH protein colocalizes with actin in cells and promotes Arp2/3-dependent actin polymerization in vitro and that WASH is transcribed in most tissues . We show that the single Drosophila WASH ortholog is essential for viability . It is therefore remarkable that human genomes harbor multiple functional WASH paralogs in highly dynamic subtelomeric regions .
Details of materials and methods are provided in Text S1 .
Seven copies of WASH are evident in the latest human genome assembly ( March 2006 ) . One is in 2q13–14 at the ancestral telomere–telomere fusion site ( 2qFS ) , and the others lie at different chromosomal ends ( Figure 1 ) . In contrast , the mouse genome contains only one ortholog ( AJ304796 ) at an internal location on chromosome 17 ( which we verified by FISH , not shown ) . One human cDNA in GenBank , BC048328 , contains a full-length 468-aa open reading frame ( ORF ) with 86 . 3% amino-acid identity to the mouse ortholog . Human WASH genes have 11 exons and span 15 kbp , commencing in a 1-kb CpG island and ending within 5 kbp of the telomere array at four locations in the assembly ( Figure 1 ) . Only the 9p copy in the assembly encodes a full-length intact protein , while other copies are prematurely truncated by frame-shifts or in-frame stop codons . Three chromosome termini carry truncated copies lacking the first two exons , owing to past translocation events that transferred only the distal portion of the gene [4] . Spliced transcripts also lacking the proximal portion and containing a shorter ORF ( 264 aa ) have been reported to GenBank ( e . g . , AY217347 , Figure 1 ) . We detect spliced transcripts of human WASH by RT-PCR in all tissues tested using RT-PCR , Northern blot analysis , and a 72-tissue dot-blot ( Figures S1–S3 ) . According to publicly available array data [19] , human WASH is expressed in a variety of tissues , with somewhat higher levels in blood cells and some areas of the brain . In order to deduce the function of the WASH protein , we compared orthologous sequences that we identified from 21 diverse organisms ( Figures 2 , 3 , and S4 ) . The resulting multiple sequence alignments reveal high conservation of this protein among vertebrates . Regions of local high similarity throughout the protein , but especially in the C-terminal portion of the alignment , are apparent between predicted vertebrate WASH proteins and sequences of fly , worm , and even evolutionarily distant Dictyostelium and Entamoeba . Notably , one of the two Entamoeba WASH genes encodes a short protein , containing only the C-terminal portion , and is like the human short form , indicating its potential functional significance . Vertebrate WASH proteins have conserved predicted nuclear localization and export signals and sumoylation sites ( Figure S5 ) suggesting a possible role in the nucleus . The most highly conserved C-terminal regions of predicted WASH proteins show similarity to the actin-binding domain ( WH2 ) of WASP family members ( Figures 2 , 3 , and S4–S9 ) . WASH proteins also have a proline-rich stretch ( P ) , followed by the WH2 domain ( V ) , a central region ( C ) , and an acidic stretch ( A ) at the very C terminus . Together , these domains form the so-called VCA module found in all WASP family members [16] ( Figure 3A ) . The VCA module is the minimal region required for binding and activation of actin polymerization via the Arp2/3 complex [14] . In known WASPs , the WH2 domain contains four conserved residues essential for actin binding [20] , the A region has a conserved tryptophan residue and mediates binding to the Arp2/3 complex [14] , and the amphipathic helical structure and conserved arginine in the C region induce conformational changes in Arp2/3 necessary to stimulate actin nucleation [21] . All these features are well conserved in WASH orthologs ( Figures 2 and S4 ) . Mammalian WASH , WASP , and SCAR proteins separate into distinct clades supported by high bootstrap values in a phylogenetic tree ( Figure 3B ) based on the alignment of the VCA region of all known WASP family members ( Figure S6 ) . Based on their VCA sequences , WASP family orthologs from nonvertebrates do not cluster with statistical support into any one of these three clades . However , WASP family members from these organisms can be grouped together with the corresponding mammalian orthologs based on sequence homology within the N-terminal portion of the proteins ( Figures 3A and S7 ) . Structural divergence of the known WASP family members outside the VCA region leads to variation in their activity and regulation and is used to subdivide them into WASPs and SCARs [15] ( Figure 3A ) . WASP and N-WASP act as specific effectors of the small GTPase Cdc42 and have a WASP homology domain 1 ( WH1 ) and a GTPase-binding domain ( CRIB ) at their N termini [22] . SCAR proteins lack these domains , but possess a specific SCAR-homology domain ( SHD ) and play a major role in Rac-induced actin dynamics [23] . WASH proteins lack N-terminal domains characteristic of either WASP or SCAR proteins , but they possess two evolutionarily conserved regions that appear to be specific to WASH orthologs ( named WASH homology domains 1 and 2 ( WHD1 and −2 ) ) ( Figures 2 and 3A ) . Based on the distinguishing features of the N-terminal portions and our phylogenetic analyses of VCA regions , we conclude that WASH proteins define a new WASP family subclass that is conserved from Entamoeba to human . Human WASH protein colocalizes with actin in vivo , consistent with our prediction that WASH has a role in actin polymerization and cytoskeleton reorganization ( Figure 4 ) . We transiently expressed full-length WASH corresponding to the BC048328 cDNA sequence and carrying GFP at its C terminus in COS-7 cells . This fusion protein colocalizes with actin in filopodia and lamellipodia , which are actin-rich cell surface extensions , but not in actin-rich stress fibers . WASP family proteins are known to stimulate actin polymerization by the Arp2/3 complex . The VCA module of previously characterized WASP family members is necessary and sufficient for this activity [14] . In order to test the effect of WASH on actin polymerization by the Arp2/3 complex , we performed pyrene actin-polymerization assays using the VCA region of the human WASH and N-WASP proteins . Without the Arp2/3 complex , neither the WASH VCA nor the N-WASP VCA has an effect on actin polymerization ( Figure 5B and 5C ) . The combination of WASH VCA and Arp2/3 complex strongly stimulates spontaneous assembly of monomeric actin ( Figure 5A and 5C ) , leading to reduction of the initial lag associated with nuclei formation , and increasing the maximum rate of filament elongation , which is dependent on the concentration of growing ends . WASH's effect on actin polymerization in the presence of Arp2/3 is less robust than that of the N-WASP VCA ( Figure 5A and 5B ) , but is similar to that of other family members like WASP and SCAR , which each induce unique kinetics of actin assembly [24] . We conclude that WASH , like the known WASP family members , is an endogenous activator of de novo actin filament assembly . The single WASH ortholog in Drosophila , which we have named washout ( wash; CG13176 ) , is structurally and phylogenetically distinct from the two WASP family members already characterized in this species ( Figure 3A ) . Homozygous mutations in either Wasp or Scar result in zygotic lethality , although some of the Wasp mutants ( “escapers” ) survive until adulthood and appear morphologically normal , but are lethargic and passive in their behavior [25 , 26] . wash gene products , like those of Wasp and Scar , are provided maternally , and wash transcripts appear to be distributed uniformly throughout the early embryo , based on RNA in situ hybridizations performed by the Berkeley Drosophila Genome Project ( BDGP ) [27] ( http://www . fruitfly . org/cgi-bin/ex/insitu . pl ) . To investigate washout function , we obtained a P-element insertion line , P{EPgy2}CG13176EY15549 ( ref [28]; BDGP , unpublished data ) , with a P-element insertion at the beginning of the washout coding region . These flies are homozygous viable with no apparent phenotype . We generated numerous imprecise excision alleles of this insertion line that are lethal when homozygous . In one of the alleles , washΔ185 , more than half of the coding region ( up to the VCA module ) is deleted and two stop codons at positions 11 and 12 are introduced ( Figure 6A ) ; no flies homozygous for this allele survive to adulthood ( Figure 6B ) . Flies bearing precise excisions of the P-element insertion in the homozygous state were viable ( Figure 6B ) , indicating that the recessive lethality of washΔ185 is due to disruption of washout . To determine the stage at which lethality occurs in washΔ185 homozygous flies , we crossed flies carrying the washΔ185 allele to a balancer chromosome ( a multiply inverted chromosome that suppresses recombination ) carrying GFP , which allows for easy detection and separation of washΔ185 homozygous embryos . Analyses of 200 washΔ185 homozygous embryos revealed that the majority of them ( 98% ) hatched , although none of them produced an adult fly . Further analysis showed that these washΔ185 animals die at the transition from 3rd larval instar to prepupal stage . While flies heterozygous for the washΔ185 allele have wild-type pupal morphology , washΔ185 homozygotes fail to contract their body and evert their spiracles prior to secreting their pupal cuticle , resulting in an elongated appearance ( Figure 6C ) . While a single-copy gene encodes WASH in most species analyzed , WASH genes experienced extensive duplication and dispersal to multiple chromosome ends during primate evolution . Our FISH assays detect many chromosomal sites , as well as extensive variation in the copy number and location of WASH genes between primate species and among human individuals ( Figure 7 ) . The total number of WASH copies , excluding partial duplications , ranges from >15 to >20 in the six analyzed human genomes . Collectively , these copies are found in 16 different sites and represent both WASH pseudogenes and intact WASH genes ( see below ) . Our FISH analyses also support the conclusion that termini of 16p , Xq , and Yq typically carry a truncated WASH . Nonhuman primates have far fewer copies of WASH than humans . WASH was detected at only eight or nine sites in the two chimpanzees analyzed by FISH and four sites in gorilla ( Figure 7 ) . WASH appears to be single-copy in orangutan and rhesus macaque , where it resides on 12p . Since the 12p location is shared among primates , it is likely to be the ancestral location of WASH before it duplicated during hominid diversification . Given the large number of human WASH loci detected by FISH and the relative paucity of sequenced WASH paralogs in public databases , we characterized the protein-coding potential of additional copies . We sequenced long-range PCR products encompassing coding exons 2–10 from three unrelated individuals . We have so far identified up to five potentially functional WASH variants and multiple pseudogenes per genome , although deeper sampling would be required to account for all copies detected by FISH in these individuals . Our sequence analyses also reveal that two of the eight human WASH copies captured in a panel of monochromosomal hybrid cell lines are full-length , intact ORFs . Intact and null WASH alleles are segregating in the human population at some chromosomal loci . For example , the hybrid panel's chromosome 20 carries the two full-length ORFs , one mapping to each end by FISH , but the presence of WASH at 20qter is polymorphic ( Figure 7 ) . Furthermore , the chromosome-9 allele captured in a panel of monochromosomal hybrid cell lines contains a stop codon in coding exon 5 , but is 99 . 9% identical to the sequenced 9pter allele in the genome assembly , which appears intact . Our survey found a total of 12 different intact WASH ORFs , in addition to BC048328 , assigned by our analyses to chromosome 20 , and the 9p copy in the assembly . These 14 ORFs differ from each other across exons 2–10 by up to18 amino acids ( ≥95 . 8% identity ) and one indel of three amino acids in the P region ( Table S1 ) . It is possible that some of these variants have slightly different functions , as they exhibit up to 11 nonconservative amino acid differences , although we detect no evidence for positive selection in mammalian WASH genes by PAML or K-estimator analyses ( Text S1 ) . Intact , and possibly more divergent , copies of WASH might be found at other chromosomal ends in other individuals , since subtelomeres undergo interchromosomal sequence exchange [4] . Indeed , by evaluating the sequences of WASH copies in the current assembly , we detect two apparent exchange events that resulted in the transfer of different segments of several hundred base pairs from one chromosome to another ( Figure S10 ) . Thus , WASH genes evolve through a combination of acquired mutation , recombination between alleles and/or paralogs , and selection , and WASH repertoires can be expected to vary extensively among individuals .
This study shows that the most telomerically located human gene family encodes evolutionarily conserved proteins with orthologs in vertebrates , flies , worms , slime mold , and entamoeba . Human WASH colocalizes with actin filaments in lamellipodia and filopodia and stimulates actin polymerization in vitro , corroborating our sequence-based predictions that WASH orthologs are new members of the WASP family . Our sequence comparisons show that these newly identified WASH proteins form a previously unrecognized subclass in the WASP family , distinct from the known N-WASP/WASP and SCAR/WAVE proteins . Known WASP family members have been implicated in the formation of lamellipodia and filopodia and in membrane-trafficking processes such as endocytosis , intracellular pathogen motility , and vesicle motility [16 , 18 , 29] . Yeast has only one WASP-related protein; its deletion causes a severe growth phenotype [30] . WASp and SCAR are important for cell-fate decisions and cell morphology , respectively , during Drosophila embryonic development [25 , 26] . Drosophila WASp is also essential for myoblast fusion [31–33] . We show here that the newly identified third WASP family member in Drosophila , washout , is essential , indicating that this gene also has an important role in early development and is not redundant with other WASP family members . The Drosophila WASH protein was recently identified as a component of a nuclear complex containing various transcriptional factors and chromatin modifiers [34] . Consistent with this finding , the vertebrate WASH consensus sequence possesses predicted nuclear localization , nuclear export , and sumoylation signals ( Figure S5 ) . Actin and actin-binding proteins are found in the nucleus , where they might be involved in chromatin remodeling , RNA processing , or gene expression [15 , 35 , 36] . N-WASP has been shown to participate directly in transcriptional regulation [36 , 37] in addition to its role in the cytoplasm . It is therefore likely that WASH also has both nuclear and cytoplasmic functions . We anticipate that loss of WASH gene function in humans will have pathological consequences , particularly since human WASH is ubiquitously expressed , with highest levels in hematopoietic tissues and some brain areas . Knockout of either N-WASP or Scar2 , both of which are also ubiquitously expressed , is embryonic lethal in mice [38 , 39] . Scar1 is most highly expressed in brain , and , accordingly , Scar1-null mice exhibit defects in brain function [40] . The WASP gene defective in human Wiskott-Aldrich syndrome is expressed in hematopoietic cells [41]; its loss causes eczema , thrombocytopenia , and immunodeficiency [42] . Human WASH genes are at heightened risk for deletion and rearrangement , since subtelomeres are hotspots of meiotic interchromosomal sequence transfers [4] . Furthermore , somatic variation in subtelomeric organization and/or telomere length could influence WASH gene expression . WASH was reported to be overexpressed in a breast cancer cell line [43] and might , like overexpression of N-WASP and the SCARs , contribute to metastasis [17] . On the other hand , subtelomeric dynamics might contribute to normal human phenotypic variation and , more generally , to diversification of the WASP family . Subtle intra-species phenotypic variation might result from variation in WASH repertoires ( gene number , location , and/or sequence ) . Intact and defective/missing WASH alleles segregate in the human population at some loci ( e . g . , 9pter , 20qter ) . We have shown that subtelomeric exchanges can create WASH loci that appear to combine sequence differences accrued by copies on different chromosomes . Finally , the identified human sequences with intact long ORFs differ from one another by as many as 11 nonconservative amino acid changes , raising the possibility that the encoded proteins might have slightly different functions . Remarkably , the genes encoding this evolutionarily conserved protein multiplied within subtelomeric regions during primate evolution and thus might contribute to interspecies phenotypic differences . Maintenance of subtelomeric segmental duplications in yeast requires selective pressure [44] . Although genetic drift cannot be excluded as the explanation for the rapid recent expansion of WASH copies in the human lineage , WASH expansion might be important for fast evolving processes such as immune response and/or brain function . The presence of WASH orthologs in Entamoeba also raises the possibility of WASH involvement in pathogenic infection . A number of unrelated pathogens hijack actin-polymerization pathways in host cells to facilitate their infection [45 , 46] . Rho GTPases and WASP family members are implicated in these pathogenic processes [45–47] . It is possible that the primate-specific subtelomeric expansion of WASH genes is associated with host response to pathogen infection . Given the critical roles of the WASP family in diverse cellular processes , elucidation of the specific role ( s ) of this newly identified and evolutionarily conserved WASH subclass should provide important insights into actin dynamics in response to external signals . The location of WASH in highly dynamic human subtelomeric regions predisposes it to duplication , deletion , and rearrangement . Certain rearrangements could have pathological consequences if WASH plays as important a role in humans as it does in Drosophila , where it is essential . Further characterization of WASH protein function ( s ) should shed light on how WASH genes and their proximity to telomeres contribute to normal human variation as well as pathology .
Genomic sequence variants of human WASH obtained in this study are available from the National Center for Biotechnology Information ( NCBI ) GenBank database ( http://www . ncbi . nlm . nih . gov/sites/gquery ) under accession numbers EU240546–EU240557 .
|
Human subtelomeres are rearrangement-prone regions near chromosome ends . They are concentrations of large , recent interchromosomal duplications . Over half of subtelomeric sequences changed copy number or location since humans and chimpanzee diverged , and subtelomeric content varies greatly among humans . Despite this dynamic activity , subtelomeres contain genes . We report the discovery of genes defining a previously unrecognized third subclass of the Wiskott-Aldrich Syndrome protein ( WASP ) family within human subtelomeres . The known WASP family members reorganize actin structures in cells in response to various signals , thereby causing cells to change shape and/or move . Representatives of this newly identified subclass , called WASH , exist in many other species , even in Entamoeba and slime mold . Like other WASP family members , WASH colocalizes with actin at the cell periphery and promotes actin polymerization in vitro . Flies lacking WASH die before becoming adults , demonstrating that WASH is critical for survival , and its function is distinct from that of the two other WASP subclasses , Wasp and Scar . Identification of the WASH subclass opens the way for future elucidation of WASH's role in the life cycles of diverse organisms , the implications of human variation in WASH copy number , and the consequences of WASH's location in dynamic telomere-adjacent regions .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion",
"Supporting",
"Information"
] |
[
"primates",
"developmental",
"biology",
"infectious",
"diseases",
"cell",
"biology",
"entamoeba",
"caenorhabditis",
"human",
"computational",
"biology",
"evolutionary",
"biology",
"drosophila",
"molecular",
"biology",
"genetics",
"and",
"genomics"
] |
2007
|
Human Subtelomeric WASH Genes Encode a New Subclass of the WASP Family
|
Accurate diagnosis of Loa loa infection is essential to the success of mass drug administration efforts to eliminate onchocerciasis and lymphatic filariasis , due to the risk of fatal encephalopathic reactions to ivermectin occurring among highly microfilaremic Loa-infected individuals living in areas co-endemic for multiple filarial species . From a pool of over 1 , 800 L . loa microfilaria ( mf ) expressed sequence tags , 18 candidate L . loa mf-specific PCR targets were identified . Real-time PCR ( qPCR ) assays were developed for two targets ( LLMF72 and LLMF269 ) . The qPCR assays were highly specific for L . loa compared with related filariae and also highly sensitive , with detection limits of 0 . 1 pg genomic DNA , or 1% of DNA extracted from normal blood spiked with a single L . loa microfilaria . Using various DNA extraction methods with dried blood spots obtained from Cameroonian subjects with parasitologically proven loiasis , the LLMF72 qPCR assay successfully estimated mf burden in 65 of 68 samples ( 50–96 , 000 mf/mL by microscopy ) , including all 12 samples subjected to a simple 10-minute boiling extraction . Additionally , the assay detected low-level microfilaremia among 5 of 16 samples from patients thought to be amicrofilaremic by microscopy . This novel , rapid , highly sensitive and specific qPCR assay is an important step forward in the laboratory diagnosis of L . loa infection .
Loa loa is a filarial nematode that infects between 3 and 13 million people in Central and Western Africa . Although infected individuals may exhibit a range of symptoms , indigenous inhabitants of endemic areas are typically clinically asymptomatic even in the setting of high parasite burden . Loiasis is nonetheless an important public health concern due to the occurrence of greater than 1 , 000 severe adverse reactions ( including fatal encephalopathies ) in Loa-infected individuals receiving ivermectin through mass drug administration ( MDA ) programs aimed at the elimination of onchocerciasis and lymphatic filariasis . Consequently , disruption of MDA has occurred in certain communities where these filarial diseases are co-endemic [1] , [2] . The mechanism of Loa-related post-ivermectin encephalopathy is unclear but risk appears to be greatest with microfilaremia above 8 , 000 organisms/mL [3] , [4] . Preventing disruption of MDA therefore necessitates identification of persons with high-level L . loa microfilaremia . Microscopic examination of mid-day blood samples is currently the only diagnostic method routinely used in endemic areas . Due to the need for examiner expertise in parasite morphology and the effort required to process large numbers of samples , this method is impractical for use as a widespread screening tool . Several alternative methods for the definitive diagnosis of loiasis have been evaluated in research settings , but none is used clinically in endemic areas . Most Loa-infected individuals will exhibit a positive IgG antibody response to crude protein extracts of Brugia malayi , but this assay is cross-reactive among all filarial pathogens as well as some intestinal helminthes [5] . Newer serologic assays based on the L . loa SXP-1 antigen and others reliably distinguish loiasis from other filarial and helminthic infections but cannot discern between active and prior infection or quantify microfilaremia [6] , [7] . Loa-specific conventional polymerase chain reaction ( PCR ) assays have been developed , but these are time-consuming , not quantitative , and not generally available for use in clinical settings [8]–[12] . Molecular methods for quantitative detection of parasitic agents [13]–[16] require selection of appropriate target sequences . We therefore sought to identify L . loa molecular target sequences suitable for incorporation into a rapid , sensitive , and high-throughput quantitative PCR assay that might allow reliable , species-specific determination of L . loa microfilaremia without the need for formal training in conventional parasitologic methods . To this end , we used novel bioinformatics strategies to parse large numbers of B . malayi , Wuchereria bancrofti , Onchocerca volvulus , and L . loa transcripts based on expressed sequence tags ( ESTs ) . We report here a set of L . loa microfilaria ( mf ) -specific target sequences with incorporation of two into real-time PCR assays that allow for enumeration of microfilaremia .
All samples were acquired under registered protocols approved by the Institutional Review Board of NIAID ( NCT00001230 ) , the Cameroon Ethical Committee and the Cameroon Ministry of Health with written informed consent obtained from all subjects . L . loa , W . bancrofti , and Mansonella perstans: mf were purified from the blood of patients seen by the NIH/NIAID Clinical Parasitology Unit . O . volvulus: adult worms were obtained from excised onchocercomas of Guatemalan patients . B . malayi: mf and adult worms were obtained from the Filariasis Research Reagent Repository Center ( Athens , GA , USA ) . L . loa ( 5 million mf ) and O . volvulus ( 50 adult female worms ) were digested overnight at 56°C in buffer G2 ( Qiagen ) with 20 mg/mL proteinase K . Genomic DNA was extracted using Genomic tip-100/G and Genomic DNA buffer set ( Qiagen , Valencia , CA , USA ) . The remaining filarial organisms were digested as described above , and genomic DNA was extracted with phenol/chloroform . One million L . loa mf and 500 , 000 B . malayi mf were frozen under liquid nitrogen and disrupted by a stainless steel piston/mortar apparatus . Total RNA was extracted using the RNeasy Kit ( Qiagen , Valencia , CA , USA ) , and poly-A RNA was isolated with the Oligotex mRNA Mini Kit ( Qiagen , Valencia , CA , USA ) . cDNA was synthesized from 1 µg of L . loa or B . malayi total RNA in 50 µL reactions containing 160 units MultiScribe reverse transcriptase , 5 . 5 mM MgCl2 , 2mM dNTP mix , 2 . 5 mM random hexamers , 20 units RNAse inhibitor , and 1X RT buffer ( Applied Biosystems , Foster City , CA , USA ) . A cDNA library was created in pTriplEx2 using the SMART cDNA Library Construction Kit ( Clontech , Mountain View , CA , USA ) . The library was screened by PCR amplification of individual plaques using primers specific to the pTriplEx2 phagemid insertion site ( table 1 ) and sequencing at the NIAID Rocky Mountain Laboratories Genomics Unit ( Hamilton , MT , USA ) . L . loa mf ESTs were assembled into contigs using the Desktop cDNA Annotation System ( dCAS 1 . 4 . 3 ) software package [17] . Contigs were selected for further evaluation as candidate assay targets based on the number of ESTs comprising the contig ( abundance ) , length of at least 200 bp with a predicted open reading frame ( ORF ) , and lack of sequence homology to i ) the non-redundant protein database ( nr ) , ii ) ESTs from related filarial pathogens , and iii ) L . loa L3 larval stage ESTs ( D . L . Fink et al . , unpublished ) . RT-PCR was performed on L . loa total RNA using the OneStep RT-PCR kit ( Qiagen , Valencia , CA , USA ) and primers specific to the 5′ and 3′ ends of each target transcript ( table 1 ) . Quantification of specific PCR product was accomplished using a 2100 Bioanalyzer instrument and 2100 Expert software ( Agilent Technologies , Waldbronn , Germany ) . L . loa mf were spiked into 200 µL aliquots of whole blood obtained from a healthy volunteer with no history of exposure to filaria-endemic regions . Following zinc BB disruption [18] , DNA was extracted using the QiaAmp DNA blood and kit ( Qiagen , Valencia , CA , USA ) . Duplicate sets of spiked whole blood samples were created by adding intact L . loa mf as described above to 50 µL aliquots of whole blood . After addition of 150 µL distilled water , the samples were vortexed briefly then boiled for 10-30 minutes . Following removal of 2 µL aliquots , samples were spun in a bench-top centrifuge at maximum speed for 5 minutes and supernatants recovered . Mid-day venous blood samples were obtained by fingerprick from Cameroonian volunteers living in a region endemic for L . loa as part of a study on Loa-associated ophthalmologic , cardiac , and renal impact . Fifty µL of each collected blood sample was examined microscopically , while an additional 50-100 µL was spotted onto filter paper . Blood spots were partitioned into 6 mm circular sections ( 10 µL dried blood each ) using disposable sterile biopsy punch tools ( Acuderm , Inc . , Ft . Lauderdale , FL , USA ) . A set of 36 punched blood spots ( 5 punches per sample ) were submerged in 200 µL phosphate buffered saline ( PBS ) and subjected to DNA extraction by the zinc BB/Qiagen spin column method described above . A second set of 36 punched blood spots ( 2 punches per sample ) was transferred into sterile tubes containing 2 ml easyMAG lysis buffer ( BioMerieux , Durham , NC , USA ) , pulse vortexed for 15 seconds , and then incubated for 10 minutes at room temperature . Samples were extracted into 50 µL easyMAG elution buffer according to manufacturer's recommendations for off-board lysis . A third set of 12 punched blood spots ( 4 punches per sample ) was immersed in 200 µL distilled water and boiled for 10–30 minutes at 99°C while shaking . qPCR was performed in an ABI 7900 sequence detection system using Taqman fast chemistry reagents ( Applied Biosystems , Carlsbad , CA , USA ) and primer/probe sets described in table 1 . Amplification conditions were 20 seconds at 95°C followed by 40 cycles of 1 second at 95°C and 20 seconds at 60°C . Quality of template was confirmed for all samples using a control primer/probe set targeting a conserved region of the eukaryotic 18S ribosomal RNA gene ( Applied Biosystems , Carlsbad , CA , USA ) . All statistical analyses were performed using GraphPad Prism 5 . 0 ( GraphPad Software , Inc . , San Diego , CA , USA ) . For each qPCR assay of a clinical sample , the number of mf present in the template was extrapolated from a standard curve derived from blood samples spiked with limiting dilutions of mf , using SDS 2 . 2 . 2 software ( Applied Biosytems , Carlsbad , CA , USA ) . These extrapolated values were adjusted by the ratio of the percentage of blood spot material used as template ( 1–10% ) compared to the percentage of spiked blood sample used as template ( 1% ) to estimate the number of mf present in each blood spot . The corrected estimates were then divided by the volume of blood processed for DNA extraction ( 50 µL for boiled samples vs . 20 µL for all other samples ) to determine the final estimates of mf concentration for each blood spot sample .
A screen of our L . loa mf-stage cDNA library produced sequence information for 1 , 882 ESTs , which were assembled into 518 unique contigs by dCAS analysis . From these , 18 potential PCR targets were identified by virtue of having limited similarity to all publicly available nematode ESTs , to the non-redundant protein database ( nr ) , and to clustered ESTs derived from L . loa L3 larvae . Each candidate transcript included a start codon and stop codon separated by at least 200 base pairs , indicating a potential ORF ( table 2 ) . Amplification of the 18 candidate transcripts was confirmed by RT-PCR using primers corresponding to the 5′ and 3′ ends of predicted ORFs . All but one of the RT-PCR reactions produced specific PCR products ( data not shown ) . Two targets ( predicted ORFs from contigs LLMF72 and LLMF269 ) were chosen for further evaluation based on abundance of specific RT-PCR product and lack of other nonspecific products . Using limiting dilutions of L . loa total RNA as template , RT-PCR detected as little as 3 . 2 pg ( LLMF72 ) or 0 . 64 pg ( LLMF269 ) total RNA , corresponding to the RNA present in a fraction of 1 mf ( figure 1 ) . For both targets the final concentration of specific PCR product was directly proportional to the log amount of RNA template used . To shorten the running time of the assay and possibly gain sensitivity , Taqman real-time PCR ( qPCR ) primers and probes were designed for the LLMF72 and LLMF269 targets . Following reverse transcription , qPCR detected limiting dilutions of L . loa total RNA from 20 ng to 2 pg , with a linear relationship between the log amount of RNA used and the number of reaction cycles needed to detect a signal above baseline ( figure 2a ) . Neither qPCR assay detected cDNA prepared from B . malayi mf RNA ( 10-fold dilutions from 20 ng to 2 pg ) , although from these same B . malayi samples a conserved region of the 18S ribosomal RNA gene could be detected using a primer/probe set targeting this sequence ( data not shown ) . Similar to the situation with total RNA , both qPCR assays detected limiting dilutions of L . loa genomic DNA from 10 ng to 0 . 1 pg . A linear relationship was again observed between the amount of genomic DNA used as template and the number of reaction cycles needed to detect signal above baseline ( figure 2b ) . Neither qPCR assay detected genomic DNA from B . malayi , O . volvulus , W . bancrofti , or M . perstans ( 10-fold dilutions from 10 ng to 0 . 1 pg ) , although all samples were detected in a linear fashion using the 18S rRNA primer/probe set ( data not shown ) . When both LLMF72 and LLMF269 primer/probe sets were used together in a single assay , there was no reduction compared with the LLMF72 assay alone in the number of reaction cycles at which any amount of DNA was detected ( data not shown ) . The qPCR assays were next evaluated with DNA extracted from whole blood samples spiked with limiting dilutions of intact L . loa mf ( 1 to 10 , 000 organisms ) . Using 1% of the total extracted DNA from each sample as template , there was once more a linear relationship between the log number of mf spiked and the number of reaction cycles needed to detect signal above baseline ( figure 2c ) . The lower limit of detection for both assays was 1% of DNA extracted from a single L . loa mf . Combining both LLMF72 and LLMF269 primer/probe sets into a single assay did not increase the sensitivity of the assay beyond that seen with the LLMF72 assay alone ( data not shown ) . Recognizing a slight advantage in fewer reaction cycles to positive ( i . e . , higher sensitivity ) , the LLMF72 assay was selected for further evaluation with blood samples from a cohort of Cameroonian study subjects with well defined L . loa microfilaremia . DNA was extracted from a portion of the dried blood spots using our standard DNA extraction method prior to performing LLMF72 qPCR . Using the spiked blood samples as a standard curve to estimate the concentration of L . loa organisms present in each blood spot , there was a significant positive correlation between the extent of microfilaremia predicted by qPCR and the level confirmed previously by microscopy ( figure 3a , Spearman r = 0 . 74; P<0 . 0001 ) . Among 36 blood spots evaluated , only one ( 40 mf/mL by microscopy ) was negative by qPCR . A second set of 36 blood spots was subjected to automated DNA extraction by easyMAG , and concentration of mf was again estimated by LLMF72 qPCR . Among this set , there was a strong linear correlation between predicted and observed microfilaremia ( figure 3b , r2 = 0 . 88; P<0 . 0001 ) . There were two samples positive by microscopy ( 20 mf/mL and 200 mf/mL ) but negative by qPCR . Sixteen of the blood spots in this set were collected from individuals who were apparently amicrofilaremic by microscopy . Five of these samples , however , contained detectable L . loa genomic DNA by qPCR ( predicted organism burden 1–7 mf/mL ) . To investigate whether time and effort of DNA extraction could be reduced , DNA was extracted from another set of spiked whole blood samples by boiling for 10–30 minutes . Using a standard curve derived from the previously spiked blood samples subjected to zinc BB/spin column extraction , the LLMF72 qPCR assay was positive for all boiled blood samples except the sample spiked with a single organism ( figure 4a ) . There was a linear relationship between number of organisms spiked and number estimated by qPCR , with no increase in DNA extraction efficiency observed with longer boiling time . Centrifugation of boiled samples and use of the supernatants as template did not have any effect on qPCR assay results ( data not shown ) . Efficiency of DNA extraction by boiling was also evaluated with a set of 12 dried blood spots . Boiling resulted in detection of L . loa DNA by LLMF72 qPCR in all 12 samples . Furthermore , there was a significant positive correlation between organism burden as assessed by microscopy and organism burden as estimated by qPCR ( figure 4b , Spearman r = 0 . 71; P = 0 . 009 ) . Overall , our qPCR assay was positive for 65 of 68 samples with parasitologically proven L . loa microfilaremia ( sensitivity 96% ) following various methods of DNA extraction . Performance of qPCR compared to microscopy for each extraction method is summarized in table 3 .
Improved molecular diagnostics for L . loa infection are needed not only in endemic areas where onchocerciasis and lymphatic filariasis elimination efforts are disrupted but also in clinical laboratories of resource-rich countries where the relative infrequency of filarial infections limits the utility of conventional parasitology methods to confirm infection in immigrants and travelers . In this study , we undertook bioinformatics analysis of filarial ESTs to identify candidate targets for Loa-specific qPCR assays . From an initial set of 18 candidates , we selected two target sequences ( LLMF72 and LLMF269 ) to develop working real-time PCR assays and defined the performance parameters of one ( LLMF72 ) using blood samples from individuals with known L . loa infection status . Our search for assay targets initially focused on mf-specific transcripts , based on the assumption that such targets would be most abundant in the setting of microfilaremia . Once it became clear that the assay performed equally well with genomic DNA compared with RNA , the advantages of a DNA template ( greater stability , no need for extra time and effort of reverse transcription ) became readily apparent . Abandoning the transcriptional specificity of an mf-specific RNA target was not a concern , as the mf stage is the only one in which organisms are found in the bloodstream . Furthermore , free DNA probably does not exist in the absence of organisms due to the relative impermeability of the microfilarial sheath . Nonetheless , detection of RNA by real-time PCR following reverse transcription quantitatively assesses expression in human blood and could be applied to estimating transmission potential in vector populations [19] , [20] should L . loa elimination campaigns be contemplated . Subsequent to the development of our assays , collaboration with the Broad Institute at Harvard/MIT resulted in sequence assembly and initial annotation of the genomes of L . loa , O . volvulus , and W . bancrofti ( http://www . broadinstitute . org/annotation/genome/filarial_worms/MultiHome . html ) . Preliminary analysis of the LLMF72 and LLMF269 target sequences indicate that they reside within single-copy genes encoding hypothetical proteins . Both target sequences have similarity to single regions of the B . malayi , W . bancrofti , Schistosoma mansoni , and Caenorhabditis elegans genomes , although there is no evidence of gene expression among ESTs of these other organisms . Species specificity of the targets is conferred by a lack of sequence similarity at the primer/probe binding sites . Consequently , our assays are negative with as much as 10 ng of purified genomic DNA ( equivalent to 104–105 mf ) from B . malayi , W . bancrofti , O . volvulus , or M . perstans . This level of specificity is extremely important , as L . loa may be co-endemic with both W . bancrofti and M . perstans , whose life cycles also include bloodstream mf . Perhaps the most notable aspect of our qPCR assays is their sensitivity . We achieved a lower limit of detection equal to 2 pg reverse-transcribed RNA , 0 . 1 pg genomic DNA , or 1% of DNA extracted from 200 µL of whole blood spiked with a single L . loa mf . Using DNA extracted from clinical samples as template , the LLMF72 qPCR assay was able to detect a single mf in a 20 µL dried blood spot ( equivalent to a burden of 50 mf/mL ) . Three false negative results were obtained with blood spot samples where mf burden was very low ( 20–200 mf/mL , or up to 4 organisms in a 20 µL blood spot ) . In light of the observed lower limits of detection for purified genomic DNA and spiked whole blood samples , these false negatives were most likely due either to sampling error ( no organisms present in the processed blood spot ) or to issues with the DNA extraction process . Previous reports of L . loa-specific PCR assays have not included explicit determination of analytic sensitivity for purified DNA or microfilaria burden [8] , [12] , though assays developed by Tourre et al targeting the gene for a 15 kD protein antigen identified individuals with parasitologically proven amicrofilaremic or occult loiasis [9]-[11] . In contrast to qPCR , these conventional PCR assays require significantly longer run times ( in particular where nested PCR is used ) and also rely on time-consuming gel-based detection methods . Finally , qPCR has the distinct advantage of being quantitative , allowing for estimation of mf burden . Mf burdens predicted by the LLMF72 qPCR assay were typically 2- to 10-fold lower than microscopic observations , no matter which DNA extraction method was used . These predictions were based on a standard curve derived from spiked whole blood samples , further indicating that extraction from dried blood spots is likely less efficient than extraction from fresh whole blood . Consequently , the qPCR assay will likely be even more sensitive , consistent , and accurate with larger volume ( 100 µL and above ) fresh blood samples . Microfilaremia estimation was most accurate and most consistent with samples subjected to easyMAG extraction . This automated extraction process will therefore be preferred in laboratory settings where such equipment is available . It is nonetheless encouraging that boiling blood spots for 10 min enabled detection of as few as 180 mf/mL ( 7 organisms in a 40 µL blood spot ) . These numbers favor successful detection of microfilaremia well below the threshold for post-ivermectin encephalopathy in endemic areas where conditions may necessitate small sample volumes , delay between sample collection and processing , and simplified extraction of template . Another advantage of using blood spots is that these are already routinely collected for screening purposes ( Ov16 rapid diagnostics , filarial-specific antibodies ) in areas co-endemic for other filarial infections and would facilitate simultaneous evaluation for multiple infections . The LLMF72 qPCR assay also detected low-level microfilaremia in 5 of 16 samples thought to be amicrofilaremic by microscopy . These discordances are unlikely attributable to false-positive assay results , because DNA extraction and assay set-up were conducted under rigorous conditions designed to protect against cross-sample contamination , and all qPCR runs included internal no template controls with verified negative results . Rather , the molecular assay is likely detecting small numbers of microorganisms that were missed by microscopy due to sampling error with the 50 µL aliquots that were examined . Previously published qualitative L . loa PCR assays have yielded positive results on blood samples from individuals with documented subconjunctival adult worms but no detectable microfilaremia by conventional methods [9] , [10] , [21] . Among the samples we tested , all five amicrofilaremic but qPCR positive specimens were also positive by serology for IgG against L . loa SXP-1 [7] , supporting infection with L . loa . Corroboration of this interpretation with additional samples would further demonstrate the utility of our molecular assay for detecting low-level microfilaremia in situations where microscopy may suffer from some degree of subjectivity . Aside from their sensitivity , specificity , and ability to objectively quantify microfilaremia , our qPCR assays offer some advantages compared to conventional parasitology . First , they eliminate the need for specialized training in filarial morphology , thereby making the assay accessible to anyone with general training in PCR techniques . Second , they enable rapid and high throughput sample processing , which would be of great benefit to centralized laboratories conducting population-based screening . Finally , our L . loa assay could be multiplexed with quantitative real-time PCR assays for other pathogens ( e . g W . bancrofti and M . perstans ) , using different fluorescent reporters [22] . Our qPCR assays also have several limitations that must be recognized . Similar to microscopy , optimum PCR performance requires proper timing of blood collection at mid-day when microfilaremia is greatest; however , a microfilaremia kinetics study by Kamgno et al . suggests that smaller numbers of organisms may be detectable outside the window of peak microfilaremia [23] , and a nomogram may be used to extrapolate qPCR results to predict true parasite burden . Another important issue is that the assays in their current form are still not practical for point-of-care use in endemic areas . One possible solution is to adapt both DNA extraction and real-time PCR processes to be performed on a handheld battery-operated microfluidic device [24]–[27] . Alternatively , the targets could be incorporated into loop-mediated isothermal amplification ( LAMP ) assays , which are carried out at a single temperature , can use whole blood as template , and can be interpreted using a simple visual color change reporting system [28] , [29] . Finally , in considering the possibility of target sequence variability , it will be important to demonstrate preservation of sensitivity with clinical samples from endemic regions outside of southern Cameroon . Notably , PCR positivity has already been demonstrated with blood obtained from patients exposed in Nigeria , Benin , and the Central African Republic ( data not shown ) . In summary , we have developed a quantitative real-time PCR assay capable of estimating L . loa mf burden from small volume blood samples . The assay is highly sensitive , species specific , and may be used for rapid , high-throughput screening either with extracted DNA or with boiled whole blood . The assay is ready for immediate use in clinical laboratories where real-time PCR equipment is available and may eventually be adapted for use in resource-poor endemic areas . This assay represents an important step forward in the diagnosis of loiasis and may ultimately be of benefit in global health campaigns to eliminate filarial diseases .
|
Loa loa is a filarial nematode that infects over 10 million people in Africa . Most infections cause no symptoms , but individuals with large numbers of blood-stage microfilariae are at risk for fatal reactions to ivermectin , an antiparasitic agent used to treat and prevent infections with Onchocerca volvulus , a related filarial parasite that may occur alongside L . loa . To address the urgent need for a point-of-care L . loa diagnostic assay , we screened a Loa microfilaria gene expression library and identified 18 Loa-specific DNA targets . From two targets , we developed a novel , rapid quantitative PCR assay for estimating L . loa microfilaria burden . The assay is highly sensitive ( detects a single microfilaria in 20 µL of blood ) and correlates well with microfilaria counts obtained with conventional microscopic techniques . The assay is species-specific for L . loa compared with related filarial parasites ( including O . volvulus ) and can be used in its current form in resource-rich areas as a diagnostic tool for L . loa infection . Although modifications will be required to make point-of-care use feasible , our assay provides a proof of concept for a potentially valuable tool to identify individuals at risk for adverse reactions to ivermectin and to facilitate the implementation of filarial control programs .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"infectious",
"diseases",
"diagnostic",
"medicine",
"global",
"health",
"neglected",
"tropical",
"diseases",
"loiasis"
] |
2011
|
Rapid Molecular Assays for Specific Detection and Quantitation of Loa loa Microfilaremia
|
Persistent bacteremia caused by Staphylococcus aureus ( SA ) , especially methicillin-resistant SA ( MRSA ) , is a significant cause of morbidity and mortality . Despite susceptibility phenotypes in vitro , persistent MRSA strains fail to clear with appropriate anti-MRSA therapy during bacteremia in vivo . Thus , identifying the factors that cause such MRSA persistence is of direct and urgent clinical relevance . To address the dynamics of MRSA persistence in the face of host immunity and typical antibiotic regimens , we developed a mathematical model based on the overarching assumption that phenotypic heterogeneity is a hallmark of MRSA persistence . First , we applied an ensemble modeling approach and obtained parameter sets that satisfied the condition of a minimum inoculum dose to establish infection . Second , by simulating with the selected parameter sets under vancomycin therapy which follows clinical practices , we distinguished the models resulting in resolving or persistent bacteremia , based on the total SA exceeding a detection limit after five days of treatment . Third , to find key determinants that discriminate resolving and persistent bacteremia , we applied a machine learning approach and found that the immune clearance rate of persister cells is a key feature . But , fourth , when relapsing bacteremia was considered , the growth rate of persister cells was also found to be a key feature . Finally , we explored pharmacological strategies for persistent and relapsing bacteremia and found that a persister killer , but not a persister formation inhibitor , could provide for an effective cure the persistent bacteremia . Thus , to develop better clinical solutions for MRSA persistence and relapse , our modeling results indicate that we need to better understand the pathogen-host interactions of persister MRSAs in vivo .
Staphylococcus aureus ( SA ) is one of the most common life-threatening human pathogens [1–3] . Methicillin-resistant strains ( MRSA ) exhibit high rates of morbidity and mortality [1–3] . MRSA bacteremia may be treated with anti-MRSA antibiotics , such as vancomycin or daptomycin . However , such treatments fail in about 30–50% of patients , resulting in persistent bacteremia [4 , 5] . Persistent MRSA bacteremia , which is defined as 3–7 days positive blood culture post-therapy [1 , 6 , 7] , is recognized as an urgent public health concern , as increased duration of bacteremia is associated with poor clinical outcomes , such as metastatic and complicated infections [6 , 8 , 9] . There are presently few therapeutic options for treating persistent MRSA bacteremia [1 , 10] . Although anti-MRSA antibiotics are ineffective against persistent MRSA strains in vivo , isolates from such patients have susceptible minimum inhibitory concentration ( MIC ) breakpoints , which is the lowest concentration that inhibits growth when cultured in vitro [1] . Thus , persistence differs from classical resistance , in which isolates exhibit resistance to antibiotics both in vivo and in vitro . Persistent bacteremia ( PB ) is attributed to persistent infection , as bacteria detected in blood derive from infection foci in internal organs , and/or endocarditis lesions [1 , 11–13] . In contrast , in resolving bacteremia ( RB ) , antibiotic treatment leads to a remission of bacteremia within a few days , indicating that MRSA in infection sites was eradicated . Further , when the infection was not cleared , it may then result in relapsing bacteremia when antibiotic treatment is terminated [14 , 15] . Relatively little is known about the specific genotypic or phenotypic characteristics of SA that result in PB . Hence , there is a critical unmet need to understand the bacterial , host , and/or antibiotic factors that contribute to PB . The phenomenon of bacterial persistence was first reported in the 1940’s [16 , 17] and , until now , it has been observed for several pathogens such as SA , Mycobacterium tuberculosis , Salmonella , Escherichia coli ( E . coli ) [18] . When cultured in vitro , most of the bacteria are killed efficiently by antibiotics , but a small portion survives and can regrow after removing antibiotics . As opposed to resistance , the regrown descendants are sensitive to antibiotics [19] . Bacterial persistence may be explained by a phenotypic heterogeneity within the bacterial population in which at least two distinct types of cells , genetically identical normal and persister cells , co-exist . Persister cells are defined as slower growing and having reduced susceptibility to bactericidal antibiotics [19 , 20] . Further , bacteria can switch from one state to another . Balaban et al . proposed a mathematical model of the phenotypic switch and quantitatively analyzed the dynamics of a phenotypic heterogeneity observed in E . coli [21] . Since then , the model and its extensions have been used to understand the dynamics of persistence in in vitro culture [22–25] . In vitro studies have aimed to describe persister cells , which are thought to be related to the phenomenon of small colony variants ( SCVs ) . SCVs are slow-growing phenotypic variants that form small colonies on agar plates [26] . SCVs of SA have been isolated from many clinical situations such as chronic osteomyelitis , endocarditis , and cystic fibrosis and are considered to be relevant for persistent infections [27–30] . SCVs show a poor response to antibiotics , less toxin production , and reduced susceptibility to host defenses [27 , 31] . Although it remains uncertain whether the formation of SCVs is a cause of SA persistence , the phenotypic heterogeneity of SA is thought to be a key feature of persistent infections . In addition , the host defense is thought to be an important factor in allowing a persistent infection . However , there is little understanding of how SA , antibiotics and host immunity work dynamically in the context of the host defense and inherent high variabilities of SA behavior . To gain new insights into the dynamic relationship between MRSA , host immunity and antibiotic efficacy , we developed a mathematical model of persistent infection that recapitulates empirical observations from in vitro and in vivo studies . This model was designed to identify the potential key determinants that may drive persistent versus resolving MRSA bacteremia . In addition , in the simulation , we also followed a clinical practice guideline such as the timing of diagnosis of bacteremia and the treatment periods of vancomycin for RB and PB [32–34] . Interestingly , we found that distinct mechanisms are key to explaining persistence in the presence ( in vivo ) vs . absence ( in vitro ) of immune responses . In vitro , the degree of persistence is critically dependent on the switch rate between the normal and persister states . In vivo , we found that host immunity for persister cells is more important . Further , we explored pharmacological strategies that may influence persistent and relapsing bacteremia outcomes and found that a persister killer , but not a persister formation inhibitor , could provide for an effective cure the persistent bacteremia . Thus , the current modeling results point towards the importance of further experimental studies that address host-pathogen responses of persister cells , rather than antibiotic-bacterial interactions typically studied in vitro .
To develop a mathematical model that contributes to an understanding of the mechanisms governing persistent SA bacteremia , we first developed a bacterial persistence model of SA growth in vitro ( Fig 1 ) [21] . We considered that SA bacteria have two distinct phenotypic states: normal and persister cells . Normal cells are defined as those having normal growth rate and susceptibility to antibiotic and immune clearance , whereas persistent cells are defined as a minor population having relatively slower growth and whose growth may be slowed further by vancomycin but which are not directly killed by vancomycin . Our overarching hypothesis in this respect was that persistent bacteremia is attributed to the phenotypically heterogeneous bacterial populations which have differential susceptibility to antibiotics and the host immune system . The model of in vitro bacterial growth consists of the following two equations: dNdt=−swN2PN+swP2NP+gN ( 1−0 . 95Van ) ( 1−TSAmax ) N−dVanVanN ( 1 ) dPdt=swN2PN−swP2NP+gP ( 1−0 . 95Van ) ( 1−TSAmax ) P ( 2 ) where variables N and P denote the number of normal and persister cells , respectively , in units of colony-forming unit per arbitrary unit of volume ( cfu/a . u . ) . T denotes the total number of SA , a sum of N and P . Switching flux is described by the law of mass action kinetics with swN2P and swP2N being the switch rate from normal to persister cells and from persister to normal cell , respectively . Bacterial growth of normal and persister are governed by growth rate gN and gP , respectively , with a maximum carrying capacity ( SAmax ) leading to growth saturation [24] . The saturable growth was formulated under an assumption that normal and persister cells compete for resources to grow . In in vitro , both normal and persister cells never reached to SAmax in our simulations because of the presence of vancomycin over the period of time . Both growth rates are inhibited in the presence of vancomycin ( Van ) to which a binary value of 0 or 1 , absence or presence , respectively , is assigned . We assumed that vancomycin inhibits 95% of growth . Although vancomycin has potential to kill normal SA , we assumed that persister cells are not susceptible to vancomycin-mediated killing , so the death rate dVan caused by vancomycin , was applied only to normal SA . The values of gN and dVan were taken from in vitro experimental data [35] . gP was assumed to be 1/20 of gN . Table 1 summarizes the parameters used in this model . To investigate how switching rates between normal and persister SA affect the in vitro persistence against antibiotics ( vancomycin ) in the absence of host immunity , we performed a parameter scan in which the values of swP2N and swN2P were varied over at least three logs . The ranges of switching rates were determined based on the published data where in vitro persistence of SA or E . coli to antibiotics was evaluated mathematically [24] . Since differences in the switch rates can affect the proportion of persister cells , a pre-simulation was run for 8 h with an initial condition of N ( 0 ) = 1 cfu/a . u . and P ( 0 ) = 0 cfu/a . u . , then vancomycin was added . As seen in the heatmap ( Fig 1B ) , varying swP2N and swN2P differentially affected the number of normal , persister and total SA . After the pre-simulation without antibiotics ( 0 h ) , normal SA reached a similar level of ~103 . 5 cfu/a . u . in all combinations and was the major subpopulation in the culture . In contrast , the numbers of persister SA differed among the combinations , being correlated with swN2P , but not with swP2N . These simulation results were in concordance with the theoretical understanding [24] . After the addition of antibiotics ( 24 and 120 h ) , the number of normal SA markedly decreased . On the other hand , little change in the number of persister cells was observed for a range of swP2N between ~10−1 . 5 to 10−5 . In this range , the number of total SA over time showed biphasic curves ( Fig 1C ) . At the early phase , most of the total SA was killed as a result of the eradication of normal SA by vancomycin . The ensuing second phase was predominantly comprised of surviving persister cells . The changes of the second slope of total SA illustrate the comparatively slow decline of the persistent subpopulation . Thus , swN2P affected the number of persister cells after the treatment when swP2N was less than ~10−1 . 5 . On the other hand , when swP2N was greater than 10−1 . 5 h-1 , no persistence was observed: Persister cells , as well as normal cells , were substantially killed by vancomycin at any value of swN2P ( Fig 1B at 120 h and Fig 1C ) . These data suggested that higher values of swP2N can lead to a lack of in vitro bacterial persistence and may also affect the pathogenesis of persistent SA bacteremia . To mathematically model in vivo SA growth , we expanded the in vitro model and added terms that describe immune system-mediated clearance ( Im ) . Fig 2A shows the schematic diagram of the model and Table 2 shows the parameters and their values . The ordinary differential equations are as follows: dNdt=-swN2PN+swP2NP+gN ( 1-0 . 95Van ) ( 1-TSAmax ) N-dVanVanN-cN ( 11+aTh ) ImN ( 3 ) dPdt=swN2PN-swP2NP+gP ( 1-0 . 95Van ) ( 1-TSAmax ) P-cP ( 11+aTh ) ImP ( 4 ) where Im is set to 1 in all our simulations of in vivo infections . Both normal ( N [cfu/a . u . ] ) and persister SA ( P [cfu/a . u . ] ) are cleared by the immune system with clearance rates , cN and cP , respectively . Malka et al . reported that the phagocytosis of SA by neutrophils , which has a pivotal role to eliminate SA at the early phase of infection , was saturated with a large number of SA [36] . Worlock et al . also found that to establish an in vivo infection , a minimum inoculum dose was necessary [37] . Hence , we describe the immune system capacity to clear bacteria as being saturable , determined by constants ( a and h ) . The functional formulations of immune clearance used in our model are as follows: cN∙ ( 11+a∙Th ) ∙Im∙N ( 5 ) cP∙ ( 11+a∙Th ) ∙Im∙P . ( 6 ) These formulations are based on the following mathematical equation to represent in vitro saturable SA clearance by neutrophils [36]: killrate=α∙Neu∙Bac ( t ) 1+γ∙Bac ( t ) +η∙Neu ( 7 ) where Neu and Bac ( t ) represent the number of neutrophils and bacteria , respectively . α is the neutrophils’ bacterial killing rate at low concentrations , and γ and η control the saturation in the killing rate as Bac and Neu are increased , respectively . The equation can be also expressed as follows: killrate=α1+η∙Neu∙11+γ1+η∙Neu∙Bac ( t ) ∙Neu∙Bac ( t ) . ( 8 ) In vivo , not only neutrophils but also macrophages play a role in clearing SA , thus Neu was generalized as clearance by immune cells ( Im ) in our model . Further , under the assumption that the number of these innate immune cells is constant , we derive the equation used in our model: killrate=cN∙11+a∙Bac ( t ) ∙Im∙Bac ( t ) ( 9 ) a=γ1+η∙Im ( 10 ) cN=α1+η∙Im . ( 11 ) In our formulation , we also introduce the Hill function to account for how the clearance rate is a function of the number of bacteria . In general , bacterial counts in a patient’s blood is not high , because bacteria in the blood are cleared rapidly within several minutes to 1 hour [38] . Hence , in clinical practice , bacteremia is diagnosed in a qualitative manner , positive or negative , by culturing the patient’s blood on the plate . Hence , due to the lack of quantitative data availability the in vivo model does not include an explicit blood compartment . Instead , we assumed that when the number of SA in infection sites exceeded a threshold number , then , MRSA would be detected in the patient’s blood , since SA detected in blood derives from ongoing infection foci . We explored a range of values for the detection limit of bacteremia and showed how our conclusions remain robust to them ( Fig 3B , S1 Fig ) . To simulate disease using this model we identified two challenges: inherent uncertainties about the quantitative behavior of bacteria and the immune system , and about the appropriate inoculum size . To address both , we performed ensemble modeling to select parameter sets ( ensemble ) which had resulted in informative disease courses . The goal of ensemble modeling is to generate a set of kinetic models ( ensemble ) where each model is described by different sets of kinetic parameters but the same mathematical structure , and outputs match specified target behavior , as for example defined by experimental or clinical data [39] . An advantage of this approach is that it enables accounting for uncertainties related to kinetic parameters which are unknown or difficult to determine experimentally , while the emergent properties of the model must satisfy a set of known biological phenomena . Thus , ensemble modeling is an often used approach in systems biology , especially in metabolic modeling [40] . Here , we applied the ensemble modeling approach to in vivo SA growth model where inherent uncertainties exist about bacterial growth behavior and the effectiveness of the host immune system and selected parameter sets which led to establishing an infection . The workflow of the ensemble modeling is schematized in Fig 2B . First , we generated 60 , 000 sets of parameters for gP , swP2N , swN2P , cN , cP , a , and h . On the basis of the definition of persister cells , the growth rate of persister SA , gP , was randomized to generate values less than those for normal SA . swP2N , swN2P were randomized in ranges shown in Table 2 based on the reported values . cN and cP were randomized so that cP was always less than or equal value of cN . Details are described in Table 2 . Among the model parameters , gN , dVan , and SAmax were fixed in the ensemble modeling . The value of gN was taken from in vitro growth study [35] . Pharmacodynamics effects of vancomycin , dVan and the growth inhibitory effect ( 0 . 95 ) , were fixed because no difference in vancomycin concentration was observed between RB and PB patients [41 , 42] . Further , the value of dVan was taken from the in vitro study and used for the ensemble modeling [35] , because the concentration of vancomycin evaluated in vitro appears similar to that in patients , based on the following observations . The death rate of vancomycin was estimated in vitro at a concentration of 10-fold of MIC . In patients , AUC ( 24h ) / MIC was reported to be 199–426 [42] , indicating , on average , vancomycin plasma concentration was 8 . 3–17 . 8 fold higher than MIC . Since plasma protein binding of vancomycin was reported to be about 40% [43] , we conclude that the clinical vancomycin concentration as an unbound form was around 5–10 fold higher than MIC , which is comparable to in vitro concentration used to estimate the death rate induced by vancomycin . After generating the parameter sets , we selected them which for a given inoculum dose range would establish an infection ( Fig 2B ) . We hypothesized that the minimum inoculum dose was between around 100 and 1 , 000 cfu/a . u and set two criteria to extract parameter sets: an inoculum dose of ~1 , 000 cfu/a . u . will establish a productive infection in the absence of antibiotic treatment , and an inoculum dose of ~100 cfu/a . u . will not ( Fig 2B ) . In the 1st regime , we conducted simulations for 48 h at an inoculum dose of 1 , 000 cfu/a . u . of normal SA and 1 cfu/a . u . of persister SA in the absence of vancomycin ( Van = 0 ) ; Since persister cells were a minor population of SA , we set the proportion of persister cells to 0 . 1% of normal SA . In the in vivo simulation , a 2 h of lag-time was introduced so that bacterial growth begins at 2 h post-inoculation . Then we performed simulations with the parameter sets , evaluated the number of normal SA at 48 h and extracted parameter sets which exceed 1x104 cfu/a . u . for the subsequent criteria . In the 2nd regime , we ran simulations at an inoculum dose of 100 cfu/a . u . and 0 . 1 cfu/a . u . of normal and persister cells , respectively , using the extracted parameter sets from the previous step , then selected parameter sets which met criteria in which the number of normal SA at 48 h was less than 0 . 1 cfu/a . u . Of 60 , 000 initial parameter sets , we obtained 4 , 614 of parameter sets whose minimum inoculum dose to establish infection was between around 100 and 1 , 000 cfu/a . u , and used them for further simulations . The inoculum dose chosen was 1 , 000 cfu/a . u . of normal SA and 1 cfu/a . u . of persister SA ( 0 . 1% of normal SA ) , unless otherwise noted , at which all the parameter sets established a productive infection in the absence of vancomycin . With the goal of classifying the selected parameter sets into RB or PB , we performed the in vivo SA growth simulation with each of the parameter sets under standard vancomycin therapy ( Fig 3A ) . In clinical practice , PB is defined by positive blood cultures on 3–7 days post-therapy [1 , 6 , 7] . In our simulation , we followed a similar diagnostic process by introducing a detection limit of bacteremia to the number of total SA ( T ) at 5 days post-treatment by vancomycin ( Fig 3A ) . Two central assumptions were that SA detected in blood derive from ongoing infective foci , and when the number of SA in infection sites exceeds a certain number ( detection limit of bacteremia ) , then , MRSA can be detected in the patient’s blood . First , we explain the algorithm to classify the models , then we describe how we set the detection limit of bacteremia . In the in vivo simulation under vancomycin therapy ( Fig 3A ) , the simulations were initiated at a dose of 1 , 000 cfu/a . u . of normal SA and 1 cfu/a . u . of persister SA without vancomycin , with bacterial growth starting at 2h post-inoculation . When normal SA reached the threshold of 1x104 cfu/a . u . , vancomycin was administered ( Van = 1 ) . As the trigger for the onset of treatment , we employed not total SA but normal SA , because persister cells are potentially less virulent [28] . Hence , we assumed that persister SAs do not provoke symptoms of infections , such as fever , and are therefore not a trigger of treatment . At 5 days post-treatment , the models were judged as PB if the number of total SA ( T ) exceeded a detection limit , and judged as RB if the number of total SA were below the detection limit . Based on clinical guidelines for MRSA bacteremia [32–34] , different treatment periods of vancomycin were applied for RB and PB: When a model was classified at 5 days of treatment as RB , treatment time was terminated at 2 weeks , but when it was classified as PB , the treatment time was extended to 4 weeks . Further , RB and PB could be sub-classified based on whether they relapsed or not ( Fig 3A , bottom ) ; In case of non-relapsing case , vancomycin therapy ( in conjunction with the immune system ) effectively eradicated both normal and persister SA . In contrast , in cases of relapse , vancomycin eradicated normal SA , but treatment is not sufficient to eradicate persister SA . The surviving persister cells are able to proliferate and switch to normal SA , accelerating bacterial growth until they reach a level of clinical presentation ( >104 cfu/a . u . ) that then trigger a subsequent round of treatment . In this study , relapse bacteremia was defined as at least two cycles of vancomycin therapy over the 180 days observation period . Here , we found that there were 4 types of simulations: RB no relapse , RB with relapse , PB no relapse , and PB with relapse; typical simulation results are shown in the bottom of Fig 3A . Next , to set an appropriate value for the detection limit of bacteremia , we explored the value under the assumption that relapse of bacteremia is rarely observed in the RB scenario . We found that when the detection limit is set to 0 . 3 cfu/a . u . or lower , very few models that satisfy the RB condition would lead to relapsing bacteremia ( Fig 3B ) . In this range , interestingly , the number of models that satisfied the RB condition did not change very much ( from approximately 2 , 500 to 2 , 000 ) over a >300-fold range of the detection limit . For further study , we selected 0 . 1 cfu/a . u . as a detection limit of bacteremia . By applying the detection limit of 0 . 1 cfu/a . u . to the algorithm shown in Fig 3A , we found that of 4614 models , 2 , 479 models ( 53 . 7% of total ) and 2 , 135 models ( 46 . 3% of total ) were classified into RB and PB , respectively . Further , of the 2 , 479 RB models , 2 , 427 models led to clearance ( RB no relapse , 52 . 6% of total ) and 52 to relapsing bacteremia ( RB with relapse , 1 . 1% of total ) . In contrast , of the 2 , 135 PB models , 940 led to clearance ( PB no relapse , 20 . 4% of total ) , and 1 , 195 led to relapsing bacteremia ( PB with relapse , 25 . 9% of total ) . To examine the cases defined by these models in further detail we graphed the timecourses of the simulations . For the 2 , 479 models classified as RB , bacterial numbers indeed dropped rapidly upon vancomycin administration , and only very few models showed relapse ( Fig 4A ) . In contrast , for 2 , 135 models classified as PB , bacterial numbers dropped more slowly , and only partially in the majority of outcomes ( Fig 4B ) . The residual bacteria were due to a substantial persister SA population that is less susceptible to vancomycin treatment in the in vivo model . Indeed , normal SA was largely eradicated by treatment , but only recurred after treatment cessation , presumably driven by the substantial persister population . Because switch rates between persister and normal SA were key determinants of persistence in vitro ( Fig 1B and 1C ) , we examined whether these parameters were also important for determining in vivo persistence . We plotted swP2N and swN2P values for above-defined in vivo RB and PB models ( Fig 4C ) . Remarkably , we saw no effect when the switch rate from persister to normal cells ( swP2N ) or the converse rate ( normal to persister , swN2P ) were altered , except at very high values of swP2N , akin to what we observed in in vitro simulations where a lack of persistence was observed when swP2N was greater than around 10−1 . 5 h-1 ( Fig 1B and 1C ) . These data implied that the switch rates were not strong determinants of PB and RB in vivo . Examining all parameter values in the selected models , we found that only the cP parameter , which determines the susceptibility of persister SA to immune clearance , demonstrated clearly different distributions between RB and PB outcomes ( Fig 4D ) . This implied that cP might be a key determinant of PB vs . RB . Further , we found that cN , cP , a , and h exhibited non-uniform distributions and narrower ranges than the ranges we randomized . These data indicated that these parameters , which are related to clearance of SA by immune systems , were sensitive to the minimum inoculum dose used as criteria in the ensemble modeling . Thus , although the parameter ranges we considered may be narrower than the ranges of other parameters ( Table 2 ) , they are sufficient to fully explore the emergent properties of the model within the criteria applied in the ensemble modeling approach . To identify the key determinants of RB vs . PB outcomes in an unbiased manner , we pursued a machine learning approach . First , we applied the quadratic programming feature selection ( QPFS ) method [44] , which assigns relative influence weights to given features in a manner that minimizes redundancy while maximizing relevance ( see Methods ) . Here , relevance was calculated as Pearson correlation coefficients between each feature and the binary target response ( Fig 5A ) . Graphing the weights of each parameter obtained by QPFS ( Fig 5B ) demonstrated that cP was selected as a sole feature for the subsequent classification . To evaluate the ability of the selected features to discern RB and PB , we used logistic regression to model the log-odds ratio of the probability of PB as a linear combination of the features . The logistic regression models for each combination of the features were evaluated by the Receiver-Operator Characteristic ( ROC ) curve and the area under the ROC curve ( AUC , Fig 5C ) . Compared with the AUC of a model that used all features , the model consisting only of cP showed an equivalent performance . This result indicates that the sole feature of cP identified by QPFS is sufficient to construct an accurate classification model with the best accuracy of 93 . 2% , which was comparable to the model encompassing all features ( 97 . 0% ) . A two-dimensional plot of cP and swN2P , which showed the 2nd highest correlation with target response ( Fig 5A ) , indicates that cP is clearly able to distinguish between RB and PB ( Fig 5D ) . Thus , cP was identified as a sole and key determinant differentiating RB from PB outcomes in this in silico model . Having identified the determinants of resolving vs persistent bacteremia ( RB vs PB ) , we next explored the key determinants of relapsing bacteremia , which resulted from roughly half the models that showed persistence at 5 days of treatment . As described ( Fig 3A ) , almost all ( 2 , 427 vs 52 ) of the RB models did not result in relapse , whereas 1 , 195 vs 940 PB models did result in relapsing bacteremia ( Fig 6A ) . The timecourse data shows that in the large majority of relapsing outcomes , substantial bacteremia occurs shortly after the 4 weeks treatment is terminated , and typically by 60 days of the timecourse . Even following repeated cycles of vancomycin treatment , the bacteremia relapses without fail . Examining the parameter distributions for the three distinct scenarios revealed that PB with relapsing bacteremia is associated with different parameter distributions in cP and gP than those of RB or PB that do not relapse ( Fig 6B ) . To identify the key determinants to distinguish relapsing bacteremia , we first performed feature selection by QPFS as before , then applied the key features to build a multinomial logistic regression model , an extension of the logistic regression for a multi-class problem . By QPFS , not only cP but also gP , the growth rate of persister cells , were ranked as relevant features to classify resolving , persistence and relapsing outcomes ( Fig 6C , upper left ) . These two features were able to show comparable overall accuracy to the all-feature model ( Fig 6C , upper right ) . The AUC for each class indicated that for the class of RB without relapse , an only-cP model was able to show comparable performance to an all-feature model , whereas both features cP and gP were necessary to distinguish the classes of PB with or without relapse ( Fig 6C , bottom ) . A two-dimensional plot of cP and gP is clearly able to discern these three classes ( Fig 6D ) , indicating that cP and gP were important factors that discriminate relapsing from non-relapsing PB . This finding suggests that while immune clearance mechanisms targeting persister cells critically determine resolving and persistent bacteremia , the growth rate of persister cells further determines whether persistent bacteremia is relapsing or not . Having identified that the key determinants of SA pathogenesis were related to persister cells and their interactions with the host , we investigated three possible types of drug strategies affecting persister bacteria . We termed these “persister killer” , “persister reverter” and “persister formation inhibitor” ( Fig 7A ) . The pharmacological strength of persister killer and persister reverter regimens were expressed as kinetic rates and simulated by mass action kinetics . During the simulation , persister killer and reverter regimens were administrated in conjunction with vancomycin which was administered for 2 or 4 weeks for RB and PB , respectively , when normal SA reached to 1x104 cfu/a . u , the defined trigger for the onset of treatment . Persister killer regimens were able to achieve complete remission of PB at 0 . 1 h-1 , which is one-third of the killing rate of normal SA by vancomycin ( Fig 7B , top ) . The reverter regimen could also clear the PB by 0 . 1 h-1 ( Fig 7B , middle ) . Because this rate is comparable to the reported highest value of switch rates from persister to normal for SA and E . coli [24] , the reverter regimen may achieve the pharmacological effect if it can maximally activate the switch rate from persister to normal . The persister formation inhibitor inhibits the switch from normal to persister cells and its strength was expressed as a percent of inhibition . Unlike the killer and reverter regimens , the inhibitor was administered from the beginning of the simulation to investigate the maximum pharmacological effects with different initial numbers of persister cells , 1 or 0 cfu/a . u . When the initial number of persister cells is 1 cfu/a . u . , which is the same initial condition applied for other analysis , the persister formation inhibitor could not achieve complete remission , even in the 100% inhibition of the switch process ( Fig 7B , bottom ) . On the other hand , when the initial number of persister cells was set to 0 , complete remission was achieved by 100% inhibition , but not by 99 . 99% prevention . Hence , persister formation inhibitor regimens may not have the potential to cure PB . Thus , these data indicated that persister killer may be the most promising and robust therapeutic strategy to resolve PB . Note that , in this analysis , complete remission achieved by the hypothetical regimens does not suggest a complete response in all patients . However , it does indicate that persister killer is a robust pharmacological strategy under any condition ( parameter values ) because the parameter sets were generated by merely randomizing values within plausible ranges without knowledge of the actual parameter distributions .
Persistent infection by MRSA , in which antibiotics fail to clear the infection , can be a cause of life-threatening persistent bacteremia . However , MRSA isolates from PB cases remain susceptible to antibiotics in vitro [1] . As a plausible mechanism of in vivo persistent infection , a phenotypic heterogeneity within the bacterial population with co-existing normal and persister cells has been considered [21] . However , little is known about the dynamic relationships between SA , antibiotics and host immunity to explain in vivo persistent infections . Thus , the factors or mechanisms that determine whether antibiotic treatment results in RB or PB have been elusive . An understanding of the determinants governing PB which is invoked uniquely in the context of in vivo infection will facilitate the identification of biomarkers to predict the outcomes and the development of novel therapeutic interventions to avoid persistent infection . In this study , we addressed the dynamic interactions between SA , vancomycin treatment and host immunity by developing a mathematical model consisting of two different bacterial phenotypes—normal and persister cells—that can switch from one to another . Persister cells were defined as slow growing and lacking susceptibility to direct killing by antibiotics . Because inherent uncertainties exist about bacterial growth behavior and the effectiveness of the host immune system , we applied an ensemble modeling approach to identify a single critical parameter that discerns RB and PB: the clearance rate of persister cells by the immune system ( cP ) . The results clearly indicated that the survival of persister cells played a critical role in establishing PB ( Fig 5D ) . To our surprise , unlike in vitro persistence ( Fig 1 ) , the switch rates between normal and persister cells were not critical to establishing PB ( Fig 4C ) . In E . coli , the formation of persister cells has been well-investigated and found to be linked to toxin-antitoxin modules and/or to ATP-dependent mechanisms [19 , 45] . In SA , negligible effects of toxin-antitoxin modules are observed and ATP levels ( e . g . energetics or formation of small colony variants ) may have roles in persistence [19 , 46] . While important , these prior investigations do not address how persistent infections emerge , or their natural history in vivo . Thus , studies investigating the intersection among persister cells , host immunity and antibiotic therapy will give us new insights for understanding in vivo persistent infections . Intracellular SA may be a promising explanation for the persister cell properties that cause persistent infections in vivo , as exploitation of the intracellular compartment can facilitate immune evasion and reduced susceptibility to antibiotics . Until recently , SA was thought to be an exclusively extracellular pathogen [26] . However , there is now accumulating evidence that SA is able to survive within host cells , both professional and non-professional phagocytes . Kubica et al . reported that SA remained within macrophages for 5 days post-infection [47] . Here , phagocytosed SA localized in acidic subcellular compartments and were able to replicate inside the cells before lysing the cells to exit [48] . The cycle of lysis and re-uptake may maintain a pool of viable SA over time [49] . In addition , SA is also able to live and proliferate inside non-professional phagocytes , such as endothelial cells and epithelial cells [49 , 50] , after entering these cells via specific receptor-ligand interaction , which facilitates adhesion and can prompt internalization [51] . Specifically , phenotypic variants ( e . g . SCVs ) are also able to reside in the intracellular compartments of phagocytic and non-phagocytic cells [52 , 53] . Intracellular SCVs have been recovered from patients with chronic osteomyelitis and cystic fibrosis years after the initial infection [28 , 30] . Because methods to detect SCVs experimentally still have difficulties due to the slow growth rate and the reversion to normal SA , it is unclear whether SCVs are an exclusive phenotype of intracellular SA . Although our mathematical model did not contain intracellular SA explicitly , intracellular SA appears to have the same characteristics as the persister cells in our model: relatively slow growth rate , and reduced susceptibility to antibiotics and immune clearance [26 , 49 , 50 , 52] . Hence , we hypothesize that persister and normal cells represent intracellular and extracellular SA in vivo , respectively . Thus , our conclusion that the clearance of persister cells by the immune system ( cP ) was identified as a key determinant of PB vs . RB outcomes , may be directly relevant to the phenomenon of intracellular persistence . In this respect , key determinants may be the interaction between intracellular SA and infected host cells: how infected cells kill intracellular SA and how intracellular SA evade this killing . Indeed , these processes may be affected by bacterial and/or host factors , both genetic and non-genetic . On the other hand , our work suggests that the switching processes between normal and persister cells are not critical to persistence . In this respect , the entering/exiting process whereby extracellular SA and intracellular SA undergo phenotypic switching may not be as important to the establishment of a persistent infection in vivo . Our modeling predicted that a persister killer strategy is the most promising and robust drug strategy to cure PB ( Fig 7 ) . Recently , as an agent to target intracellular SA , antibody-antibiotics conjugate ( AAC ) has been applied to the in vivo intracellular MRSA model , showing good efficacy compared to vancomycin [54] . Kim et al . also reported a new class of synthetic retinoid antibiotics which kill MRSA persister strains [55] . Being able to kill the persister cell directly , therapeutics such as these may hold promise to resolve the persistent infection . In addition to directly killing persister cells , inhibiting the evasion mechanism of persister cells and enhancing host defense activity against persister cells may also improve outcomes . On the other hand , we found that preventing persister emergence may not be efficacious . For example , our modeling results indicate that inhibiting the entry process via blocking receptor-ligand interactions via antibodies may not be as promising an anti-infective strategy against intracellular persistence [56 , 57] . For the purposes of identifying key parameters to distinguish PB and RB , we employed machine learning approaches , instead of global parameter sensitivity analyses , such as partial rank correlation coefficient ( PRCC ) , which explore the impact of each parameter value [58] . In PRCC , multiple parameter sets are generated by randomizing the values , then simulations are performed with the parameter sets . To find the key sensitive parameters , correlations between parameter values and the simulated data of interest are evaluated . However , in our analysis , global parameter sensitivity analyses are not appropriate , because certain parameter sets may show no bacterial growth in vivo even in the absence of vancomycin and may thus lead to the misleading conclusion that such parameter is a key determinant of RB . Further , while parameter sensitivity analysis is a powerful way to evaluate the impact of each parameter , it is difficult to address the effects of multiple parameters , as observed in Fig 6 . Thus , the combination of simulations with the parameter sets and machine learning classification may be a powerful approach to identify the key parameter combinations in complex dynamical systems and is applicable to identify biomarkers of pathogenesis and treatment efficacy . In line with expectations , in our model of in vitro MRSA growth , the fraction of persister cells was a function of the switch rates , especially the switch rate from normal to persister cells ( swN2P ) as shown in Fig 1C . The initial fraction of persister cells at swN2P of 10−5 , 10−3 , 10−2 h-1 was 9 . 0–0 . 8 x 10−4 , 8 . 9–9 . 7 x 10−2 , and 8 . 4–9 . 2 x 10−1% , respectively , when swP2N was in the range of 10−2 to 10−1 . Mulchahy et al . reported the initial fraction of persister cells of Pseudomonas aeruginosa isolated from cystic fibrosis patients who succumb to a chronic untreatable Pseudomonas aeruginosa infection to be around 10−4 to 10−3%[59] . On the other hand , longitudinal isolates showed approximately 100-fold higher fractions , which then often contained the hip mutation , which increases persister formation , i . e . the switch rate from normal to persister cells swN2P . In these cystic fibrosis patients , this increased switch rate caused by the hip mutation only arose after the persistent infection had been established . This may imply that the switch rate is not a key determinant to establish persistence , similarly to the modeling results pertaining to MRSA persistent bacteremia discussed in this study . Instead , the host-pathogen interaction of persister cells may have an essential role in establishing the persistent infection in cystic fibrosis . Previous studies modeled the interplay between bacteria , antibiotics , and the immune system to explore optimal antibiotic treatment regimens and the likelihood of emergence of resistance [60 , 61] . While the models also included subpopulations which acquired antibiotic resistance , the clinical phenomenon of persistence and the factors that may contribute to it , which is the focus of our work , was not explored by these studies . In this study , while the ensemble approach allowed us to survey a large parameter space , key assumptions were made , and some parameter values or conditions were fixed in an arbitrary manner . To investigate sensitivities of such values to our main conclusion about the key determinant of PB , we repeated the analysis using different values of detection limits of bacteremia ( S1 Fig ) , growth inhibition and killing rate by vancomycin ( S2 Fig ) , initial number of persister cells ( S3 Fig ) , SAmax ( S4 Fig ) and growth rate of normal SA ( S5 Fig ) . Further , in this analysis , we did not explicitly specify a cut-off value to define the extinction of SA , which may affect relapse bacteremia , so we also explored the impact of the values ( S6 Fig ) . Even using these different values , the model reached the same conclusion: the key determinant of PB is the clearance rate of persister cell by immune factors . In summary , based on these modeling outcomes , we surmise that persistent MRSA bacteremia results from relatively ineffective killing or slow clearance of persister SA by host immunity , even in the context of potent antibiotics . Despite in vitro results to the contrary , the clinical significance of switch rates between normal and persister cells is likely to be negligible . For a better understanding of the disease , our modeling results indicate that we need to better understand the pathogen-host interactions of persister MRSAs in vivo .
Mathematical model used in the analyses are described in Results section . All the analyses were done in MATLAB version 2017b . The equations were solved numerically by ode15s . All the models selected were classified into RB or PB and binarized into 0 or 1 , respectively , for the subsequent supervised classification modeling . To identify key parameters which discern the two-type of infection , we applied a feature selection method , QPFS . Note that , for classification , we refer to the seven random valued parameters as features . The Pearson correlation statistic was calculated between each of the features and the target class ( 0 or 1 ) . The same Pearson correlation statistic was also calculated between every pair of features . The absolute values of these correlation statistics yielded a feature relevance vector F and redundancy , or cross-correlation matrix Q , respectively . From these , a weight vector x was calculated that minimizes the quadratic program minx{12 ( 1-α ) xTQx-αFTx} ( 12 ) where a balancing factor α was calculated by dividing the mean of F by the mean of Q to balance the linear ( relevance ) and quadratic ( redundancy ) terms . Once the optimal vector of weights , x , was identified , we selected top-ranked features to incorporate into a logistic regression model . For the comparison , all parameters and all parameters except for top-ranked features were also applied to the model . Specifically , we modeled the log-odds ratio of the probability of persistent versus resolving as a linear combination of the features with cross-validation . Maximum likelihood estimates of the regression coefficients were achieved using the fitglm function in MATLAB , with the link function "logit" . Receiver Operating Characteristic ( ROC ) curves were obtained by iterating the log-odds ratio threshold over its full range of possible values , and at each value calculating the true and false positive rates , the ratio of true to total positives , and the ratio of false to total negatives , respectively . As an index of the importance of features incorporated into the model , areas under the ROC curve were calculated for each curve . The models classified into RB or PB were further divided into with or without relapse bacteremia . Since the number of models for RB with relapse is few , other three classes , RB without relapse , PB without relapse and PB with relapse , were used for the subsequent classification modeling . To select key features of the three classes , QPFS was applied . The top-ranked features were incorporated into a multinomial logistic regression model with cross-validation . For the comparison , all parameters and all parameters except for top-ranked features were also applied to the model . Since imbalance of class size affects the classification , especially for the classification using irrelevant features , an equal number of data were used by reducing the number of data in major classes to the same number of minor class . As an index of the importance of features incorporated into the model , areas under the ROC curve were calculated for each class .
|
Staphylococcus aureus causes potentially lethal infections of the bloodstream and target organs when able to enter the body , often via skin trauma or catheterization . Methicillin-resistant Staphylococcus aureus ( MRSA ) resist common antibiotics , but are often successfully treated with vancomycin . However , in some MRSA patients , vancomycin is less effective . This results in persistent bacteremia , even though the isolates can be effectively killed in vitro . MRSA bacteria are thought to switch between two forms , normal and persister cells , that are genetically identical . Persisters , the minor subpopulation , are slow-growing and show lower susceptibility to vancomycin than normal MRSA . To understand the dynamic interplay between the two bacterial populations when challenged by host immunity and vancomycin treatment , we developed a mathematical model and analyzed it in simulations of clinically relevant scenarios . Our work suggests that the immune clearance rate of persister MRSA rather than the MRSA switch rate is a key determinant to establish persistent bacteremia . The model also suggests that increasing killing rate of persisters is a promising therapeutic strategy . Our findings emphasize the need to better understand the interactions of persister MRSAs with host cells and immune responses in vivo .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"antimicrobials",
"cell",
"death",
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"drugs",
"immunology",
"cell",
"processes",
"microbiology",
"mathematical",
"models",
"vancomycin",
"staphylococcus",
"aureus",
"bacterial",
"diseases",
"methicillin-resistant",
"staphylococcus",
"aureus",
"antibiotics",
"pharmacology",
"bacteria",
"bacterial",
"pathogens",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"staphylococcus",
"medical",
"microbiology",
"mathematical",
"and",
"statistical",
"techniques",
"microbial",
"pathogens",
"immune",
"system",
"bacteremia",
"blood",
"cell",
"biology",
"anatomy",
"physiology",
"microbial",
"control",
"biology",
"and",
"life",
"sciences",
"organisms"
] |
2019
|
Identifying determinants of persistent MRSA bacteremia using mathematical modeling
|
Mayaro virus ( MAYV ) of the genus alphavirus is a mosquito-transmitted emerging infectious disease that causes an acute febrile illness , rash , headaches , and nausea that may turn into incapacitating , persistent arthralgias in some victims . Since its discovery in Trinidad in 1954 , cases of MAYV infection have largely been confined there and to the northern countries of South America , but recently , MAYV cases have been reported in some island nations in the Caribbean Sea . Accompanying these reports is evidence that new vectors , including Aedes spp . mosquitos , recently implicated in the global spread of Zika and chikungunya viruses , are competent for MAYV transmission , which , if true , could facilitate the spread of MAYV beyond its current range . Despite its status as an emerging virus , there are no licensed vaccines to prevent MAYV infection nor therapeutics to treat it . Here , we describe the development and testing of a novel DNA vaccine , scMAYV-E , that encodes a synthetically-designed consensus MAYV envelope sequence . In vivo electroporation-enhanced immunization of mice with this vaccine induced potent humoral responses including neutralizing antibodies as well as robust T-cell responses to multiple epitopes in the MAYV envelope . Importantly , these scMAYV-E-induced immune responses protected susceptible mice from morbidity and mortality following a MAYV challenge .
Mayaro virus ( MAYV ) is an alphavirus in the Togaviridae family originally identified on the island of Trinidad in 1954 . MAYV infection can result in an acute febrile illness lasting 3 to 5 days with symptoms including rash , headache , nausea , vomiting , and diarrhea . Similar to chikungunya virus ( CHIKV ) infection , approximately 50% of MAYV-infected individuals develop painful recurrent arthralgia that can last for months after acute illness has cleared . Since its discovery , only sporadic cases of MAYV infection have been reported , mostly in tropical areas of South America [1 , 2] . Serosurveys suggest that it may also be circulating in Central American countries [1 , 3] . In 2015 , the first case of MAYV infection outside of South America was reported on the island of Haiti , highlighting the potential for an expansion of the MAYV range to include island nations of the Caribbean Sea [2] . Alphaviruses are arthropod-borne viruses ( arboviruses ) transmitted between animal reservoirs and hosts via mosquitoes and other vectors . The primary vectors for MAYV are Haemagogus spp . mosquitos , which primarily reside in rural regions [2 , 4 , 5] . Most cases of MAYV infection have been in individuals that have entered forest environments due to work or travel . Recently , vector competence studies have shown that Aedes aegypti mosquitos have the capacity to transmit MAYV , sparking fears that the virus may spread beyond current endemic regions to possibly worldwide given the wider geographical distribution of Aedes aegypti [2 , 6 , 7] . In recent years , Aedes spp . mosquitos have been responsible for causing worldwide outbreaks of flaviviruses , including dengue and Zika [8–10] . With higher than expected seroprevalence in Central America [1 , 3 , 11] , it is evident that human infection with MAYV is becoming more common and widespread . As of currently , there are no approved vaccine or therapeutics available for blocking or treating MAYV infection , and current containment measures have focused on minimizing human-mosquito contacts through vector controls . Alphaviruses like MAYV have single-stranded , positive sense RNA genomes that encode four nonstructural proteins ( nsP1-4 ) and structural polyproteins; capsid , E3 , E2 , 6K and E1[12] . Along with E2 and E1 , the envelope mRNA also encodes for a 6K polypeptide , which contributes to the processing and membrane insertion of E1 and precursor E2 ( pE2 ) viral envelope glycoproteins that are cleaved during translation and get incorporated into mature virions [13 , 14] . Occasionally , a frameshift during translation can lead to production of a transframe ( TF ) polypeptide that consists of the 6K polypeptide with additional amino acids on its C terminus [15] . Both 6K and TF appear to be involved in efficient virus budding [12 , 16] . The E2 and E1 glycoproteins form heterodimers shortly after translation , and these heterodimers associate into trimers when virions assemble . The E2 glycoprotein is primarily involved in virus attachment to host cells while the E1 protein mediates the fusion of the virus and host cell . Both E1 and E2 on the surface of virions are targets of anti-MAYV antibody responses [12 , 17 , 18] . A protective vaccine targeting MAYV would be an important tool for impeding its spread as well as reducing or eliminating disease caused after infection . To date , two MAYV vaccines have been developed and shown to be immunogenic in mice models: ( 1 ) an inactivated virus vaccine [19] and ( 2 ) a live-attenuated virus vaccine [20] . Here , we describe the development of a novel synthetic DNA-based vaccine targeting MAYV . Our lab has previously used this platform to develop potent vaccines against diverse infectious agents , including HIV-1 , Ebola virus , Middle East respiratory syndrome coronavirus ( MERS-CoV ) , and Zika virus ( ZIKV ) [21–23] . DNA-based vaccines are cheaper to design , manufacture , and deploy than conventional vaccine platforms , and they are capable of inducing both humoral and cellular responses with virtually no risk of causing disease themselves [21 , 22 , 24] . Importantly , DNA vectors are non-immunogenic , thus there is no reduction in potency after multiple administrations . While first generation DNA vaccines exhibited poor and inconsistent immunity , improvements in plasmid and antigen designs combined with the addition of electroporation ( EP ) -enhanced vaccine delivery have greatly improved the immunogenicity of these vaccines . A recent phase I clinical trial of a novel synthetic ZIKV vaccine , GLS-5700 , found that it induced both cellular and humoral responses , including neutralizing antibodies , in the vast majority of study volunteers , and passive transfer of post-vaccination sera from volunteers completely protected mice from morbidity and mortality following ZIKV challenge [22] . A phase II clinical trial of a therapeutic DNA vaccine ( VGX-3100 ) encoding consensus sequences of human papilloma virus ( HPV ) E6 and E7 proteins induced antigen-specific humoral and cellular responses in volunteers after EP-enhanced delivery , and these responses could mediate viral clearance and clinical regression of CIN2/3 cervical dysplasia in volunteers [24] . The vaccine described here , scMAYV-E , encodes a synthetically designed , consensus full-length MAYV envelope antigen sequence . EP-enhanced delivery of scMAYV-E into immunocompetent mice induced high levels of cellular responses to multiple MAYV-E epitopes along with robust antibody responses that could neutralize MAYV infection in vitro . Immunization of interferon α/β receptor knockout mice ( IFNAR-/-; A129 ) with scMAYV-E protected the mice from morbidity and mortality following MAYV challenge , where the protection in this model was primarily due to vaccine-induced humoral responses . The robust immunogenicity of the scMAYV-E vaccine demonstrated here supports the need for further testing of this vaccine as a viable means to halt the spread of this virus and protect individuals from MAYV disease .
Human embryonic kidney 293T ( HEK293T; ATCC-CLR-N268 ) and Vero CCL-81 ( ATCC #CCL-81 ) ( ATCC , Manassas , VA , USA ) cells were cultured in D10 media: Dulbecco Modified Eagle's Medium ( Invitrogen Life Science Technologies , San Diego , CA , USA ) supplemented with 10% heat-inactivated fetal calf serum ( FCS ) , 3 mM glutamine , 100 U/ml penicillin , and 100 U/ml streptomycin [23] . Mouse splenocytes were cultured in R10 media: ( RPMI1640 , Invitrogen Life Science Technologies , San Diego , CA , USA ) supplemented with 10% heat-inactivated FCS , 3 mM glutamine , 100 U/ml penicillin , and 100 U/ml streptomycin . All cell types were cultured in incubators set to 37°C and 5% CO2 . The synthetic MAYV vaccine DNA construct encodes a full-length MAYV envelope sequence . The consensus gene insert was computationally optimized for improved expression . The construct was synthesized commercially ( Genscript , NJ , USA ) and then sub-cloned into a modified pVax1 vaccine expression vector under the control of the cytomegalovirus immediate-early promoter as described previously [23] . HEK293T cells were plated in six-well plates at 6x105 cells/well and transfected 24 hours later with scMAYV-E and pVax1 empty vector control plasmids using GeneJammer transfection reagent ( Agilent Technologies , Santa Clara , CA , USA ) according to the manufacturer's instructions . The transfection was carried out in Opti-MEM medium ( Invitrogen ) . The transfected supernatants and cell lysates were collected 48 hours post transfection , and antigen expression was confirmed by western blot analysis . Cells were washed with phosphate-buffered saline ( PBS ) and lysed with lysis buffer containing 50 mM HCl , 150 mM NaCl , 1% Nonidet P-40 , 1% Triton X-100 , 0 . 1% sodium dodecyl sulfate , and a cocktail of protease inhibitors ( Roche , Basel , Switzerland ) on ice for 30 minutes with intermediate vortexing . After 10 minutes of centrifugation at 13 , 000 rpm , the supernatant was collected and analyzed by sodium dodecyl sulfate-12% polyacrylamide gel electrophoresis and transferred to a nitrocellulose membrane for immunoblotting with antisera ( 1:100 dilution ) against scMAYV-E . Secondary antibodies coupled to horseradish peroxidase ( HRP ) were used at a dilution of 1:5 , 000 . Next , the membrane was stripped using NewBlot Nitrocellulose 5x Stripping Buffer ( Li-Cor , Nebraska , USA ) then probed with β-actin rabbit monoclonal antibody ( 1:1 , 000 ) ( Li-Cor , USA ) as a loading control . For Immunofluorescence analysis , cells were seeded on top of coverslips in a 6-well cell culture plate . After washing three times with PBS , the cells were incubated for an hour at 37°C with a Fluorescein isothiocyanate ( FITC ) -conjugated goat anti-human IgG ( Santa Cruz Biotechnology Inc . , USA ) . The nucleus was stained with 4′ , 6-diamidino-2-phenylindole ( DAPI ) at room temperature for 20 minutes . PBS washes were performed after each incubation step . The samples were subsequently mounted onto glass slides using Fluoroshield Mounting Medium ( Abcam , USA ) and were viewed under a confocal microscope ( LSM710; Carl Zeiss ) . The resulting images were analyzed using the Zen software [21 , 25] . Five- to eight-week old female C57BL/6 mice ( The Jackson Laboratory , Bar Harbor , ME , USA ) were housed and vaccinated in a light-cycled , temperature- and humidity-controlled animal facility . Four- to six-week old mice of C57BL/6 background deficient in the interferon-α/β receptors ( IFNAR-/-; A129 ) were purchased from The Jackson Laboratory ( MMRRC Repository-The Jackson Laboratory , USA ) and established a breeding colony approved by the Wistar Institutional Animal Care and Use Committee ( IACUC# 112842X_0 ) . Animals were bred and housed in a barrier animal facility at the Wistar Institute . All murine studies were performed in accordance with the recommendations from the National Institutes of Health ( NIH ) and the Wistar Institute Institutional Animal Care and Use Committee . For DNA immunization , five- to eight-week-old female C57BL/6 mice and four- to six-week old IFNAR-/- mice of mixed sex were delivered 25 μg of DNA in a total volume of 30 μl of sterile water by a syringe into the anterior tibialis ( TA ) muscle . The same site is immediately electroporated by the CELLECTRA adaptive constant current enhanced electroporation ( EP ) delivery device ( Inovio Pharmaceuticals , PA , USA ) , where a three-pronged minimally invasive device is inserted 2 mm into the TA muscle . Each prong consists of 26-gauge , solid stainless-steel electrode , and triangulated square-wave pulses of 0 . 1 Amps are delivered at 52 msec/pulse twice with a 1 second delay at the insertion site . Further details of the EP usage have been previously described in detail [26 , 27] . Blood was collected by the submandibular method preceding the DNA injection and EP procedure . All mice were anesthetized with 2–5% isoflurane ( Phoenix , Clipper , MO , USA ) during procedures . Each group received one , two , or three immunizations at 2-week intervals , and mice were euthanized one week following the last immunization . Spleens were dissected and individually crushed with the use of a Stomacher device ( Seward , UK ) . Splenocytes were strained with a 40 μm cell strainer ( ThermoFisher , USA ) and treated 5 minutes with Ammonium-Chloride-Potassium ( ACK ) lysis buffer ( Quality Biologicals , MD , USA ) to lyse red blood cells . The splenocytes were resuspended in R10 and used in the Mouse IFN-γ ELISpot PLUS assay ( Mabtech , USA ) according to the manufacturer’s instructions . Briefly , 2x105 splenocytes from the scMAYV-E or pVax1 control immunized mice were added to each well and incubated for 18 hours at 37°C in 5% CO2 , either in the presence of media alone ( negative control ) , media with Cell Activation Cocktail ( BioLegend , USA ) containing pre-mixed phorbol 12-myristate-13-acetate ( PMA ) and ionomycin ( positive control ) , or media with peptide pools ( 1 μg/ml ) consisting of linearly pooled 20 individual peptides that are 15-mers overlapping by 9 amino acids spanning the length of the MAYV envelope protein . Spots were formed by the addition of 5-bromo-4-chloro-3-indolyl-phosphate/nitro blue tetrazolium ( BCIP/NBT ) color development substrate ( R&D Systems , USA ) . Spot forming units ( SFU ) were quantified by an automated ELISpot reader ( CTL Limited , USA ) . The average number of SFU from the media alone wells was subtracted from each stimulated well , and the data was adjusted to SFU per 106 splenocytes [22] . MaxiSorp high-binding 96-well ELISA plates ( ThermoFisher , USA ) were coated with commercial recombinant MAYV E1 Envelop Glycoprotein ( Alpha Diagnostic , San Antonio , TX , USA; MAYV11-R-100 ) and MAYV E2 Envelop Glycoprotein ( Alpha Diagnostic , USA; MAYV21-R-100 ) at a concentration of 0 . 5 μg/mL in coating buffer ( 0 . 012 M Na2CO3 , 0 . 038 M NaHCO3 , pH 9 . 6 ) at 4°C overnight . Plates were washed 5 times with PBS buffer solution containing 0 . 01% Tween-20 ( PBST ) ( ThermoFisher , USA ) , and blocked with 10% FBS in PBS at 37°C for 1 hour . Serum samples were serially diluted ( starting 1:50 , dilution factor 3 . 16 ) in PBS containing 1% FBS , and 100 μl was added to each well . After incubation at 37°C for 2 hours , the plates were washed 5 times with PBST and then incubated with HRP-labeled goat anti-mouse IgG ( Sigma-Aldrich , USA ) , at 37°C for 1 hour . After the final wash , 100 μl of fresh 3 , 3’5 , 5’-Tetramethylbenzidine ( TMB ) Substrate ( Sigma-Aldrich ) was added per well and incubated for 10 minutes . The reaction was stopped by adding 50 μl of 2 M H2SO4 , and the optical density of the plate was measured at 450 nm by Biotek ELISA plate reader ( Biotek , USA ) . The antibody endpoint titer was defined as the highest dilution of a serum sample with OD values > ( mean + 3SD ) of pVax1 vaccinated mice . Samples with a titer <50 were given an endpoint titer of 1 . All assays were done in triplicate . For mouse IgG subtyping , Pierce Rapid Antibody Isotyping Kit ( ThermoFisher , USA ) was used with 1:100 dilution of pVax1 mouse sera or scMAYV-E immune sera from C57BL/6 background . 2x106 single-cell suspended mouse splenocytes were added per well to a U-bottom 96-well plate ( ThermoFisher ) . Cells were stimulated for 5 hours at 37°C in 5% CO2 , either in the presence of media alone ( negative control ) , media with Cell Activation Cocktail ( BioLegend ) containing pre-mixed PMA and ionomycin ( positive control ) , or media with MAYV envelope peptides ( 1μg/ml ) spanning the length of the entire protein , where all of the samples contained a protein transport inhibitor cocktail ( eBioscience , San Diego , CA , USA ) . Upon completed stimulation , the cells are washed with FACS buffer ( PBS containing 0 . 1% sodium azide and 1% FBS ) . Cells were stained for the surface proteins using fluorochrome-conjugated antibodies per the manufacturer’s instructions ( BD Biosciences , San Diego , CA , USA ) . The cells were washed again with FACS buffer , then fixed and permeabilized using BD Cytofix/Cytoperm ( BD Biosciences ) per the manufacturer’s protocol before the intracellular cytokines were stained using fluorchrome-conjugated antibodies ( BD Biosciences ) . The following antibodies were used for surface staining: LIVE/DEAD Fixable Violet Dead Cell stain kit ( Invitrogen ) ; CD19 ( V450; clone 1D3; BD Biosciences ) ; CD4 ( FITC; clone RM4-5; eBioscience ) ; CD8α ( APC-Cy7; clone 53–6 . 7; BD Biosciences ) ; CD44 ( A700; clone IM7; BioLegend ) . For intracellular staining the following antibodies were used: IFN-γ ( APC; clone XMG1 . 2; Biolegend ) ; TNF-α ( PE; clone MP6-XT22; eBioscience ) ; CD3ε ( PerCP/Cy5 . 5; clone 145-2C11; Biolegend ) ; IL-2 ( PeCy7; clone JES6-SH4; eBioscience ) . The LSRII flow cytometer was outfitted with the following lasers and bandpass filters: ( i ) violet ( 405 nm ) – 450/50 , 525/50 , 560/40 , 585/42 , 605/40 , 660/40 , 705/70 , 780/60; ( ii ) blue ( 488 nm ) – 530/30 , 695/40; ( iii ) green ( 532 nm ) – 575/25 , 610/20 , 660/20 , 710/50 , 780/60; and ( iv ) red ( 640nm ) – 670/30 , 710/50 , 780/60 . All data was collected using an LSRII flow cytometer ( BD Biosciences ) and analyzed using FlowJo software ( Tree Star , Ashland , OR , USA ) and SPICE v5 . Boolean gating was performed using FlowJo software to examine the polyfunctionality of the T cells from vaccinated animals [21 , 23 , 24] . The Trinidad Regional Virus Laboratory ( TRVL ) 15537 strain of MAYV was obtained from ATCC ( ATCC VR-1863 ) , passaged once through Vero cell culture , and quantified as previously described [20 , 28] . One week after the second immunization , 10 mice from either scMAYV-E vaccinated or pVax1 vaccinated groups were challenged with 102 plaque-forming units ( PFU ) of MAYV diluted in 100 μl of sterile PBS delivered by a gradual intraperitoneal ( i . p . ) inoculation . Mice were weighed daily and evaluated for clinical signs of infection as follows: ( 1 ) decreased mobility; ( 2 ) hunched posture; ( 3 ) footpad swelling; ( 4 ) decreased grip strength of the hindlimb; ( 5 ) paralysis of hindlimb ( s ) ; ( 6 ) moribund . The Wistar Institute IACUC does not approve death as an endpoint . Mice were euthanized if ( 1 ) weight loss was sustained for 3 days or more and total weight loss reaches 20% of the body weight , ( 2 ) mice exhibited 3 or more signs of clinical symptoms as listed above concurrently for over 3 days , or ( 3 ) mice were moribund . All mice that exhibited signs and symptoms of MAYV infection lasted no more than 7 days during the experiment ( s ) . Whole blood was collected on day 6 post challenge for Mayaro viral titer quantification as described previously [20 , 28] , and the footpad swelling of individual mouse was measured with a caliper on the same day . Eight days post challenge , the number of surviving mice was noted and humanely euthanized . Four- to six-week old IFNAR-/- mice were immunized twice at a two-week interval . One week after the second immunization , immune sera were isolated from whole blood and combined into a single pool per group . Passive transfer of immune sera was performed by intraperitoneal ( i . p . ) injection at 200 μl per mouse into four- to six-week old naive IFNAR-/- mice . Mice receiving immune sera from the pVax1 immunized group or PBS served as negative controls . All groups were challenged with 102 PFU of wild-type TRVL 15537 strain of MAYV and monitored daily as described above . Four- to six-week old IFNAR-/- mice were immunized twice at a two-week interval . One week after the second immunization , all mice were euthanized , and spleens were collected and processed as single-cell suspensions . The cell viability was examined by Trypan Blue dye exclusion staining using a Countess II Automated Cell Counter ( ThermoFisher ) . Adoptive transfer was performed by i . p . inoculation ( 2x106 cells/200 μl ) into four- to six-week old naive IFNAR-/- mice . A group receiving splenocytes i . p . ( 200 μl ) from the pVax1 immunized mice or PBS served as negative controls . All groups were challenged with 102 PFU of wild-type TRVL 15537 strain of MAYV and monitored daily as described above . PRNT assay was carried out to detect and quantify the presence of neutralizing antibodies in the immunized mouse serum samples as previously described [25 , 27 , 29] . Heat-inactivated ( 56°C , 30 minutes ) immune sera were diluted serially , and 150 μl of each diluted sample was mixed with an equal volume of 102 PFU of wild-type TRVL 15537 strain of MAYV , followed by incubation at 37°C for 1 . 5 hour for a virus-antibody neutralization reaction . 100 μl of the virus and serum mixture was then used to inoculate a monolayer of Vero cells in a 12-well plate followed by an incubation at 37°C for 1 . 5 hour with rocking every 15 minutes . Next , the supernatant was removed from each well , and a layer of 2% methyl cellulose was added . After further incubation at 37°C with 5% CO2 for 3 days , the cells were fixed , stained with Crystal Violet ( ThermoFisher ) , and plaque numbers were recorded as described [28] . MAYV alone without serum incubation served as negative control . After washing the stained cells with distilled water and air-drying the plates , the number of foci per well were counted using a stereomicroscope . The percentage of infectivity was calculated as: % reduction in infection = {1- ( number of plaques from serum samples / number of plaques from negative control ) } x100 . Purified CD14+ human monocytes were obtained from the University of Pennsylvania Immunology Core ( Philadelphia , USA ) . Human monocyte-derived macrophages ( MDM; 1x106/well ) were cultured in a 6-well plate in Macrophage Base Medium DXF ( PromoCell GmbH , Germany ) supplemented with 60 ng/mL of granulocyte-macrophage colony-stimulating factor ( GM-CSF ) recombinant protein ( R&D Systems , USA ) . The culture was incubated without disturbance at 37°C with 5% CO2 for 3 days . and MDMs were washed once with PBS prior to infection . Cells were infected with 102 PFU of TRVL 15537 strain of MAYV that was preincubated for 1 hour at 37°C with either pVax1 immune sera or a 100-fold dilution of pooled day 35 immune sera from scMAYV-E vaccinated mice . The same media containing virus-pVax1 sera or virus-scMAYV-E sera mixtures were added to washed MDMs in a 6-well plate that were then kept for 1 hour at 37°C with a rocking interval of 15 minutes . The supernatant was removed following incubation , and Macrophage Base Medium DXF ( PromoCell GmbH , Germany ) was added to each well and further incubated at 37°C with 5% CO2 for 48 hours . Uninfected and infected macrophages were stained with Live Cell Labeling Kit- Green Fluorescence-Cytopainter ( Abcam , Cambridge , MA , USA ) according to the manufacturer’s instructions . Stained macrophages were imaged on a microscope ( EVOS Cell Imaging Systems; Life Technologies ) and % live cells ( i . e . , Labeling Dye Green+ ) were assessed by visual inspection of images from six different reviewers assessing the inhibited infection relative to base-line by 90% [22] . A monolayer of Vero CCL-81 cells plated on 12-well plates were inoculated with 200 μl of supernatants from MAYV-infected MDMs that were previously pre-incubated MAYV with either pVax1 or scMAYV-E sera . After 36 hours of incubation , viability of the Vero CCL-81 cells was examined by Trypan Blue dye exclusion staining using a Countess II Automated Cell Counter ( ThermoFisher ) . The assays were done in triplicates , and each dot represents the cell viability from a single well +/- SEM . The experiment was repeated twice . The Wistar Institute Animal Care and Use Committee ( IACUC ) approved the animal experiments under the protocol #112770 in accordance with the Guide for the Care and Use of Laboratory Animals by the National Research Council of the National Academies ( 8th Edition , 2011 ) . The Wistar Institute IACUC and the Animal Facility comply with all applicable federal statues and state regulatory provisions that include but are not limited to the following: the Animal Welfare Act of the U . S . Department of Agriculture ( USDA ) , the U . S . Government Principles for the Utilization and Care of Vertebrate Animals Used in Testing , Research , and Training developed by the Interagency Research Animal Committee ( IRAC ) and the Public Health Service ( PHS ) Policy on Humane Care and Use of Laboratory Animal of the National Institute of Health ( NIH ) . Appropriate practices and procedures as defined in the Biosafety in microbiological and biomedical laboratories ( US Dept . of Health and Human Services ) were used in sample handling . Samples were stored at -80°C in a bio safety level-2 ( BSL-2 ) facility at the Vaccine & Immunotherapy Center , The Wistar Institute , PA , USA . All results are representative of those from at least two independent experiments with similar results . Graphs , standard curves , and pie charts were made using GraphPad Prism ( version 4 . 0 ) software . IC50 values were calculated using a non-linear regression of the reciprocal of the serum dilution compared to the control . The survival data for mouse experiments were graphed using Kaplan-Meier survival curves . Two-tailed p values were calculated by log-rank ( Mantel-Cox ) test for nonparametric data using GraphPad Prism ( version 4 . 0 ) software .
We employed bioinformatics and synthetic DNA technologies to create a novel DNA vaccine encoding a full-length MAYV envelope gene sequence comprised of the E1 , E2 , and E3 glycoprotein domains as well as the 6K/TF polypeptides . The synthetic DNA insert was created by aligning full-length envelope genomic sequences from multiple MAYV strains deposited in the GenBank-NCBI ( National Center for Biotechnology Information ) database and determining the most common conserved amino acid at each position . Consensus antigen sequences account for genetic variability that occurs over time in a sequence and thus mapped at the phylogenetic midpoint ( Fig 1A ) . Studies show that synthetic consensus sequences can focus immune responses against conserved sites as well as broaden T cell immunity [23 , 30] . To improve the transcription and translation of the vaccine inserts , modifications to the insert sequences were made prior to cloning into the modified pVax1 vaccine expression vector namely the addition of an immunoglobulin E ( IgE ) leader sequence to the N-terminus ( Fig 1B ) along with codon and RNA optimization of the sequences [21] . Reference models of the scMAYV-E antigen made using Discovery Studio 4 . 5 . software indicate that its predicted structure matches that of a wild-type MAYV envelope with the fusion loop at the end of domain E1 tucked into a fold in domain E2 ( Fig 1C and 1D ) , thus preserving important envelope functional sites . Expression of the scMAYV-E antigen was confirmed in vitro through western analyses of 293T cell lysates transfected with scMAYV-E vaccine ( Fig 1E ) . The immunogenicity of the scMAYV-E vaccine was evaluated in C57BL/6 mice . Initially , groups of mice were immunized three times , two weeks apart , with 25 μg of either scMAYV-E or an pVax1 empty vector plasmid using EP-enhanced intramuscular ( i . m . ) delivery [31] . Immunized mice were bled on day 0 and one week after each injection to obtain sera , which were assayed for the presence of antibodies to MAYV envelope using ELISA with commercially available MAYV E1 and E2 glycoproteins . The results show that all mice develop anti-MAYV E1 antibodies after a single immunization ( Fig 2A ) , and the anti-MAYV E2 IgG response was boosted by both a second and third immunization ( Fig 2B ) . The presence of anti-E1 and anti-E2 responses in murine sera post third immunization ( day 35 ) was confirmed via western blot analysis using commercially available E1 and E2 glycoproteins as loading antigens and pVax1-transfected 293T cell lysates as a negative control ( Fig 2C ) . Multiple immunizations also enhanced the affinity of the vaccine-induced anti-MAYV responses against E1 and E2 glycoproteins as evidenced by increasing endpoint titers after the second and third immunizations ( Fig 2D and 2E ) . There is a comparable increase in total IgG1 , IgG2a , IgG2b , and IgG3 subtypes after the third immunization ( Fig 2F ) . Both Vero-CCL81 ( Fig 2G ) and U-87 MG neuronal cells ( Fig 2H ) infected with the wild-type MAYV could be identified by indirect immunofluorescence assay using pooled day 35 sera from scMAYV-E immunized mice but not when using pooled day 35 pVax1 sera . Studies on related alphaviruses including CHIKV have established anti-viral antibodies as a primary correlate of protection [27 , 32 , 33] . We next evaluated whether the antibody response elicited by our scMAYV-E vaccine in mice could neutralize MAYV infection in vitro . A plaque reduction neutralization test ( PRNT50 ) was performed on pooled day 35 sera from scMAYV-E immunized , pVax1 immunized , or uninfected control mice . Antibodies in scMAYV-E vaccinated mice neutralized MAYV infection of Vero-CCL81 cells with a high neutralizing titer ( IC50 = 789 . 8 ) while pVax1 immune sera or uninfected control sera were not able to neutralize the virus at the highest serum dilution ( 1:10 ) ( Fig 2I ) . These results indicate that scMAYV-E induces robust , MAYV-specific humoral responses in mice that are capable of blocking MAYV infection in vitro . Multiple alphaviruses are known to infect macrophage cells , which are believed to play a role in alphavirus-induced arthritis [18 , 34] . To assess the potential of scMAYV-E immune sera to protect against macrophage infection , an in vitro infection assay was performed . Addition of wild-type MAYV TRVL 15337 to human monocyte-derived macrophages ( MDMs ) progressively decreased the cell viability over 48 hours post infection ( Fig 3A ) . Importantly , preincubation of MAYV TRVL 15337 with pooled sera from scMAYV-E immunized mice significantly increased the viability of MDMs whereas MDMs incubated with MAYV and pVax1 sera demonstrated high levels of cell death , which were observed in all visual fields using a phase-contrast and a fluorescent microscope ( Fig 3B ) . Percent of live cells labeled with Labelling Dye Green seen as fluorescent green cells were evaluated by six independent reviewers ( Fig 3C ) . These results led to the conclusion that while MAYV was likely inducing cell death , the sera from scMAYV-E immunized mice can significantly increase MDMs viability in the presence of MAYV . To further investigate this observation , we co-cultured for 36 hours Vero CCL-81 cells in a 6-well plate with supernatants from the infected MDMs preincubated with scMAYV-E sera or pVax1 sera . Vero cells maintained in the presence of supernatant from MDMs incubated with MAYV+scMAYV-E immune sera had a cell viability over 60% when compared to a 40% cell viability of Vero cell cultured in the presence of MDMs supernatant incubated with MAYV+pVax1 sera ( Fig 3D ) . Taken together , these results suggest that scMAYV-E immune sera are capable of inhibiting viral replication in MDMs . We next evaluated anti-MAYV cellular immunity in splenocytes collected from the scMAYV-E immunized C57BL/6 mice . One week after the third immunization ( day 35 ) , pVax1 control and scMAYV-E immunized mice were euthanized , and bulk splenocytes were obtained for ELISpot assay . Briefly , harvested splenocytes from mice were ex vivo stimulated with various peptide pools encompassing the full-length MAYV envelope protein ( i . e . , glycoprotein E1 , E2 , and E3 ) . The identity of each peptide pool is shown in Table 1 . The antigen-specific production of interferon gamma ( IFN-γ ) by the cells is reported as spot forming units ( SFUs ) per million cells . Mice immunized ( 3x ) with scMAYV-E exhibit a robust cellular response to multiple peptide pools throughout the MAYV envelope glycoprotein domains ( Fig 4A ) . A similar , robust cellular response to multiple MAYV peptide pools was also observed in a separate cohort of C57BL/6 mice that received a single immunization of scMAYV-E vaccine ( 1x ) ( Fig 4A ) . To better define the dominant epitope ( s ) of scMAYV-E that elicit cellular responses in the C57BL/6 mouse model , an ELISpot mapping analysis was performed on bulk splenocytes from the mice that received three immunizations of scMAYV-E . Thirty-five matrix peptide pools encompassing the entire MAYV envelope protein , each comprised of individual 15-mer peptides that overlap by 9 amino acids , were created ( Table 1 ) and ex vivo stimulated splenocytes for the IFN-γ ELISpot assay as previously described . Several matrix pools from different regions of the MAYV envelope glycoprotein domains were identified by SFU/106 cells , with matrix pools 7 and 13 eliciting the highest responses in the E1 region as well as 21 and 28 in the E3+E2 regions ( Fig 4B and 4C ) . Subsequent mapping analysis identified dominant epitopes within the MAYV E1 glycoprotein as ‘GRSVIHFSTASAAPS’ ( Fig 4B ) and within the MAYV-E3+E2 glycoproteins as ‘LAKCPPGEVISVSFV’ ( Fig 4C ) . The amino acid sequences of the dominant epitopes determined from the ELISpot mapping analysis were confirmed using the immune epitope database analysis resources tools ( http://tools . iedb . org ) , substantiating the effective antigen processing of scMAYV-E vaccine in this strain of mice . To further characterize the cellular response induced by the scMAYV-E vaccine , splenocytes collected from the C57BL/6 mice receiving three immunizations of scMAYV-E or pVax1 control were evaluated by polychromatic flow cytometry . A panel of fluorophore-tagged antibodies was created and used to characterize helper ( CD4+ ) and cytotoxic ( CD8+ ) T cells production of the activated-state cytokines such as IFN-γ , tumor necrosis factor-α ( TNF-α ) , and interleukin 2 ( IL-2 ) post ex vivo stimulation of bulk splenocytes with peptides comprising MAYV full-length envelope . CD4+ and CD8+ T cells isolated only from scMAYV-E vaccinated mice were able to produce each intracellular cytokine upon stimulation with MAYV peptides ( Fig 5A and 5B ) . scMAYV-E vaccination also induced polyfunctional responses in both T cell subsets ( i . e . , production of multiple activated-state cytokines from a single T cell ) and indicated the presence of triple-positive T cells in both subsets ( IFN-γ+IL2+TNF-α+% as red ) ( Fig 5C ) . Combined with the ELISpot results , scMAYV-E DNA vaccine induces both cellular immunity to MAYV and polyfunctionality of antigen-specific T cells . We next evaluated whether the vaccine-induced MAYV-specific responses could protect against MAYV infection or disease in an animal challenge model . Previous studies showed that older immunocompetent mouse models do not exhibit arthritogenic signs of disease upon alphavirus challenge [29 , 35] . We therefore chose to use the interferon α/β receptor knockout mouse ( IFNAR-/-; A129 ) model , which has a defective innate immune response to pathogens . We established that a dose of 102 PFU of MAYV administered i . p . produced measurable clinical signs of disease including weight loss , foot swelling , and criteria for euthanasia . First , the cellular and humoral immunogenicity of scMAYV-E vaccinated IFNAR-/- mice were evaluated as previously described with C67BL/6 mice to confirm that they mount similar adaptive immune responses . IFN-γ ELISpot responses ( Fig 6A ) and antigen-specific IgG endpoint titers ( Fig 6B ) were comparable to those observed in the C57BL/6 mice . For the challenge studies , cohorts of 10 four- to six-week old IFNAR-/- mice were similarly immunized twice , two weeks apart , with either 25 μg scMAYV-E vaccine or pVax1 empty vector plasmid as a control . Animals were challenged on day 21 , one week after the second immunization , with 102 PFU of wild-type MAYV and were checked daily for 8 days for clinical signs of infection . All mice receiving pVax1 empty vector plasmid exhibited significant and progressive weight loss ( Fig 6C ) . In contrast , mice vaccinated with scMAYV-E initially had minor weight loss over the first 4 days of challenge but exhibited slight weight gain after day 5 post challenge ( Fig 6C ) . Percent body weight change on Day 7 comparing scMAYV-E and pVax1 groups had a p-value of 0 . 0115 . Importantly , 100% of vaccinated mice survived the challenge , while all control mice met euthanasia criteria by 6–7 days post challenge ( Fig 6D ) . pVax1 group also had significant footpad swelling at day 6 post challenge ( Fig 6E and 6F ) . Quantification of MAYV in sera collected 6 days post challenge showed that scMAYV-E vaccinated mice had a significant reduction in circulating virus compared to pVax1 injected mice ( Fig 6G ) . Combined , these data demonstrate that immune responses induced by the scMAYV-E vaccine provide protection from morbidity and viral load following MAYV challenge in this murine challenge model . We next evaluated the relative contribution of the scMAYV-E vaccine induced humoral and cellular responses in an in vivo passive and adaptive transfers of MAYV challenge model . In this investigation , a cohort of 6 four- to six-week-old IFNAR-/- mice were immunized twice at a two-week interval with 25 μg of scMAYV-E or pVax1 . One week following the final immunization , the mice were euthanized and blood and bulk splenocytes were collected from each animal . Sera and splenocytes from each group was pooled . Cohorts of 6 naive , four- to six-week old IFNAR-/- mice were injected with either PBS , 200 μl of pooled immune sera , or 2x106 pooled splenocytes containing T cells , then subsequently challenged with 102 PFU of MAYV i . p one hour post passive and adaptive transfers . Challenged mice were monitored for up to 8 days for clinical signs of disease . All mice receiving PBS prior to challenge progressively lost weight and were eventually euthanized due to severe disease as expected between day 5 and 7 . Adoptive transfer of T cells from immunized mice provided some protection from weight loss ( Fig 7A ) and partial protection from disease ( Fig 7B ) . Significantly , 100% of mice receiving scMAYV-E immune sera exhibited no weight loss ( Fig 7A ) and all survived the challenge ( Fig 7B ) . Combined , these data establish the scMAYV-E induced humoral response as the main driver of its protective efficacy in this murine MAYV challenge model .
Mayaro virus is an emerging infectious disease agent endemic in tropical regions of South America , but recent evidence suggests that its range may be expanding into Central America and island nations of the Caribbean Sea [2 , 11 , 35] . The virus causes an acute febrile illness with symptoms including rash , headache , nausea , and diarrhea that can turn into a debilitating , long-term arthralgia in some patients after acute infection has cleared [36 , 37] . There are currently no approved vaccines or therapeutics to combat MAYV disease and spread . Here , we report on the generation and immunogenicity of scMAYV-E , a synthetic , enhanced DNA vaccine encoding a novel consensus-designed sequence of the MAYV-envelope glycoproteins . Immunization of mice with scMAYV-E using enhanced EP delivery induced robust , MAYV-specific humoral and cellular responses . Importantly , these responses can neutralize MAYV infection in vitro and can fully protect susceptible mice from morbidity and mortality following MAYV challenge . The results show that scMAYV-E is a highly immunogenic vaccine candidate that warrants further testing in additional systems and animal models for developing countermeasures against MAYV infection and diseases . The precise correlates of protection for MAYV have not been defined . A recent one-year longitudinal study of confirmed MAYV-infected individuals in Peru found that infection elicited robust anti-viral immune responses including strong neutralizing antibody responses and secretion of pro-inflammatory immune cytokines including IL-13 , IL-7 , and VEGF [28] . They also report that the strong neutralizing antibody response was not sufficient to prevent long-term negative outcomes of MAYV infection; however , these humoral responses developed post infection . Studies on related alphaviruses , including CHIKV , strongly suggest that a potent , neutralizing antibody response primarily mediates protection from infection , but non-neutralizing antibodies may contribute to protection as well through alternative effector functions [38] . Post infection , there is likely an important role for cellular immunity that may complement the humoral responses . Further studies will be needed to address this important issue . The scMAYV-E DNA vaccine elicits both humoral and cellular responses against MAYV and consequently be an important tool to provide comprehensive protection from MAYV infection and disease . Antibodies to MAYV are generated after the initial priming immunization with scMAYV-E , and these responses increase after both one and two boosts in terms of binding capacity and affinity to rE1 and rE2 . Immune sera from vaccinated mice was able to detect full-length MAYV envelope in scMAYV-E transfected cells as well as MAYV infected cells . scMAYV-E vaccination of mice was able to induce neutralizing antibodies that can block viral entry and inhibit cell death induced by MAYV infection in human MDMs . Passive transfer of immune sera from scMAYV-E vaccinated mice to susceptible naive IFNAR-/- mice prior to MAYV challenge completely protected animals from illness , further confirming the importance of a strong humoral response for conferring protection from alphavirus infection . Although the anti-MAYV T cell response appears less important for an immediate protection against MAYV infection , it may still be essential for the prevention of chronic disease by eliminating virus-infected cells . The cellular components induced by the scMAYV-E DNA vaccine target multiple epitopes along the full-length MAYV envelope glycoprotein . The strongest cellular responses were directed to epitopes in the E3+E2 domains of the envelope , whereas the responses to epitopes in the E1 glycoprotein were less robust . The epitope mapping studies using ELISpot assays identified two immunodominant epitopes , ‘LAKCPPGEVISVSFV’ in the E3+E2 domain and ‘GRSVIHFSTASAAPS’ within the E1 domain , providing important and useful reagents for studies of the T cell immune response in this haplotype . Interestingly , adaptive transfer of splenocytes from scMAYV-E immunized mice to susceptible naive IFNAR-/- mice prior to MAYV challenge provided partial protection from weight loss and clinical symptoms of MAYV disease , suggesting that MAYV-specific cellular responses do contribute to protection . In this adoptive transfer experiment , MAYV-specific T cells were not purified or enriched from bulk splenocytes prior to transfer , thus it is possible that the partial protection observed here could be enhanced with a larger dose of antigen-specific T cells . The immunogenicity of the scMAYV-E DNA vaccine mirrors what we observed in a previous DNA vaccine candidate targeting chikungunya virus ( CHIKV-E ) which encodes a synthetic consensus sequence of the full-length chikungunya envelope protein [27] . The CHIKV-E vaccine was similarly able to generate humoral and cellular responses directed towards the CHIKV envelope protein , and these responses could protect mice from morbidity and mortality following a CHIKV challenge [27] . The synthetic DNA vaccines have practical advantages for development including ease of manufacture and stability at warmer temperatures , likely reducing the requirement for a cold chain . They are non-live and non-replicating and do not integrate , thus providing conceptual safety advantages as well . Since DNA vaccine vectors do not induce anti-vector serology , they can be administered multiple times with no loss in potency and without interfering with other vaccine protocols . Such logistical and safety advantages warrant further studies of this vaccine approach , especially pertaining to diseases prevalent in resource-poor tropical settings where MAYV is the most prevalent . To the best of our knowledge , scMAYV-E is the third vaccine candidate for MAYV developed . The first vaccine was an inactivated Mayaro virus , and the second candidate reported was a live-attenuated MAYV vaccine [19 , 20] . Both prior vaccines were shown to induce anti-MAYV humoral responses that could protect mice from MAYV challenge , but neither study reported on the induction of cellular responses to MAYV . The synthetic scMAYV-E DNA vaccine described here generates MAYV-specific humoral and cellular responses without viral replication , which is likely important for immune-challenged , young , pregnant , and elderly populations of potential travelers and residents in endemic areas in need of vaccine-induced immune protection . Additional studies of this vaccine approach using the DNA platform will provide further insight into the relative merits of such methods in the field .
|
Mayaro virus ( MAYV ) is a mosquito-transmitted virus that causes fever , headache , chills , nausea and joint pain that can last for months after infection . The rising number of cases , due to increased mosquito circulation , and the threat of an epidemic emphasize its importance as an emerging virus , but there are no licensed vaccines to prevent Mayaro infection nor therapeutics to treat it . In this study , we gathered fundamental knowledge on how the immune system responds to MAYV infection , and we evaluated the efficacy of a novel , synthetic DNA envelope vaccine ( scMAYV-E ) in mice . Analysis of immune responses in mice demonstrated that this vaccine induces potent T cell immunity and antibodies . Mice who receive the vaccine and then are challenged with live MAYV are protected against Mayaro disease . This data provides evidence that the DNA-based MAYV vaccine may be able to prevent Mayaro disease . Thus , the scMAYV-E vaccine is a promising step forward for MAYV vaccine development . Future testing will evaluate whether this vaccine is a viable means to halt the spread of MAYV and protect individuals from Mayaro disease .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"blood",
"cells",
"humoral",
"immunity",
"blood",
"serum",
"medicine",
"and",
"health",
"sciences",
"enzyme-linked",
"immunoassays",
"immune",
"physiology",
"body",
"fluids",
"immune",
"cells",
"immunology",
"vaccines",
"animal",
"models",
"model",
"organisms",
"experimental",
"organism",
"systems",
"infectious",
"disease",
"control",
"antibodies",
"immunologic",
"techniques",
"research",
"and",
"analysis",
"methods",
"immune",
"system",
"proteins",
"infectious",
"diseases",
"white",
"blood",
"cells",
"animal",
"studies",
"proteins",
"animal",
"cells",
"immunoassays",
"t",
"cells",
"mouse",
"models",
"immune",
"response",
"biochemistry",
"blood",
"anatomy",
"cell",
"biology",
"immunity",
"physiology",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"immune",
"serum",
"humoral",
"immune",
"response"
] |
2019
|
Protective immunity by an engineered DNA vaccine for Mayaro virus
|
Neuronal ceroid lipofuscinosis ( NCL ) is a progressive neurodegenerative disease characterized by brain and retinal atrophy and the intracellular accumulation of autofluorescent lysosomal storage bodies resembling lipofuscin in neurons and other cells . Tibetan terriers show a late-onset lethal form of NCL manifesting first visible signs at 5–7 years of age . Genome-wide association analyses for 12 Tibetan-terrier-NCL-cases and 7 Tibetan-terrier controls using the 127K canine Affymetrix SNP chip and mixed model analysis mapped NCL to dog chromosome ( CFA ) 2 at 83 . 71–84 . 72 Mb . Multipoint linkage and association analyses in 376 Tibetan terriers confirmed this genomic region on CFA2 . A mutation analysis for 14 positional candidate genes in two NCL-cases and one control revealed a strongly associated single nucleotide polymorphism ( SNP ) in the MAPK PM20/PM21 gene and a perfectly with NCL associated single base pair deletion ( c . 1620delG ) within exon 16 of the ATP13A2 gene . The c . 1620delG mutation in ATP13A2 causes skipping of exon 16 presumably due to a broken exonic splicing enhancer motif . As a result of this mutation , ATP13A2 lacks 69 amino acids . All known 24 NCL cases were homozygous for this deletion and all obligate 35 NCL-carriers were heterozygous . In a sample of 144 dogs from eleven other breeds , the c . 1620delG mutation could not be found . Knowledge of the causative mutation for late-onset NCL in Tibetan terrier allows genetic testing of these dogs to avoid matings of carrier animals . ATP13A2 mutations have been described in familial Parkinson syndrome ( PARK9 ) . Tibetan terriers with these mutations provide a valuable model for a PARK9-linked disease and possibly for manganese toxicity in synucleinopathies .
Neuronal ceroid lipofuscinosis ( NCL ) is a progressive neurodegenerative diseases characterized by brain and retinal atrophy and the intracellular accumulation of autofluorescent lysosomal storage bodies resembling lipofuscin in neurons and other cells . NCL has been reported in several domestic animal species including cattle , goat , sheep , cat and dog [1] . To this date eight forms of NCL have been classified by clinical criteria , age of onset and presence of lysosomal storage material in humans . Causative mutations in nine genes ( PPT1 , TPP1 , CLN3 , CLN5 , CLN6 , CLN7 , CLN8 , CTSD and CLCN6 ) have been identified [2]–[7] . All these mutations lead to an early-onset NCL in human , sheep or dogs . Tibetan terriers show a late-onset lethal form of NCL manifesting at 5–7 years [8] . The disease starts most frequently with blindness in twilight and disorientation . Affected dogs often appear nervous or anxious and the lack of motor coordination becomes more severe with disease progression . Affected dogs often have problems to jump up from the ground floor , or have problems by going upstairs . At the final stages of the disease mild to severe seizures have been observed . There are no treatment options for affected dogs . Due to the late-onset of NCL in Tibetan terriers some of the NCL susceptible dogs are bred and have progeny . Tibetan terriers provide a valuable model for late-onset NCL as no causative mutation for late onset NCL is known in human and other species . NCL-genes known from human , sheep or other dog breeds were evaluated and excluded for late-onset NCL in Tibetan terriers in previous studies [9]–[13] . Parallel with our analysis , an ATP13A2 frameshift mutation was identified as responsible for adult-onset neuronal ceroid lipofuscinosis in Tibetan terriers [14] . Here , we describe the same association of the canine chromosome ( CFA ) 2 region which harbors a splice-variant mutation in canine ATP13A2 with late-onset NCL and analyse the mRNA of ATP13A2 from NCL-affected Tibetan terriers to characterize the possible effect of this mutation on the resulting protein .
We collected samples from 24 NCL-affected Tibetan terriers and 1 , 347 Tibetan terriers showing no clinical signs of NCL . Due to the late-onset of NCL , non-affected Tibetan terriers aged <4–7 years can manifest signs of NCL later in life . All 24 NCL-affected Tibetan terriers showed clinical signs of NCL at an average of six years . Necropsy was performed on seven of the NCL-affected Tibetan terriers . Histological examination of retina , cerebellum , cerebrum and spinal cord revealed gold-brown intracytoplasmic pigment within retinal ganglion cells and neurons , respectively , associated with degeneration and glioses using hematoxylin eosin staining . In addition , cytoplasmic accumulations show a positive signal using periodic acidic-Schiff reaction ( PAS ) and luxol fast blue staining ( LBF ) ( Figure 1 ) . Brain iron accumulation using Turnbull blue staining could not be found in affected dogs . The phenotype of controls has been assured through reports of the owners and veterinary examinations . All controls had an age of at least 10 years and all their registered offspring was ascertained as NCL-clear . The controls were distantly related to families with NCL-affected dogs using five generation pedigree records . These requirements made it difficult to collect a larger number of controls fulfilling these requirements . A genome-wide association analysis was performed for 12 Tibetan-terrier-NCL-cases and 7 Tibetan-terrier-NCL-controls using the 127K canine Affymetrix SNP chip ( Affymetrix , Santa Clara , CA , USA ) . Controls were either not closely related with the controls nor with the cases what caused an unbalanced design . Chromosomal regions with genome-wide significant associations ( −log10p-values>3 . 2–3 . 5 ) were located on dog chromosomes ( CFA ) 2 at 81 . 3–86 . 2 Mb , on CFA8 at 61 Mb , on CFA12 at 60 . 2–61 . 4 Mb , on CFA18 at 48 . 4–50 . 3 Mb , on CFA22 at 48 Mb and on CFA37 at 29 Mb using PLINK for association testing with adaptive permutations . We were not able to identify one homozygous haplotype shared among all cases in these associated regions . In order to rule out false positively associated regions and to exclude heterogeneity , we used two different approaches . First , we employed multipoint non-parametric linkage analyses and secondly , we enlarged the number of cases and controls and we tested SNPs from significantly associated regions on CFA2 , 8 , 12 and 18 in this larger data set . A total of 28 microsatellites ( Table S1 ) covering the associated and their adjacent chromosomal regions were employed in three families including 107 Tibetan terriers ( Table S2 , Figure S1 ) . Mendelian inheritance and correctness of marker transmission in the pedigrees genotyped was confirmed using Pedstats [15] . A non-parametric multipoint linkage analysis has been employed as this approach does not require assumptions on the mode of inheritance and specification of genetic parameters such as mode of inheritance , penetrance and allele frequencies and so this approach should be useful for traits when the correct mode of inheritance is unknown or assumptions of genetic parameters may lead to ambiguous results or heterogeneity might be of importance . We ruled out linkage for NCL for the chromosomes 8 , 18 , 22 and 37 ( Table S3 ) . Evidence for chromosome-wide significant linkage of the genomic region at 83 . 22–88 . 0 Mb on canine chromosome 2 was affirmed . A total of 13 SNPs ( Table S4 ) was genotyped for the significantly associated regions on CFA2 , 8 , 12 and 18 in 376 Tibetan terriers whereof 24 were NCL-affected and 30 parents or offspring of NCL-affected dogs . These 376 Tibetan terriers also contained the 107 dogs out of the three families for linkage analysis and further 114 Tibetan terriers to increase the number of family members of these three families and in addition , 111 Tibetan terriers belonging to 17 different pedigrees . Genotyping was performed using the Sequenom technology ( Sequenom , Hamburg , Germany ) . A significant association ( −log10p-value = 6 . 7 ) could only be shown for one SNP ( BICF2S23719003 ) located on CFA2 at 83 . 94 Mb . Mixed model analysis ( MMA ) was employed to ensure the associated region on CFA2 and to narrow down the associated region . The advantage of MMA over a simplistic approach without considering any other effects in modelling association can be seen in removing disturbing effects caused by data structure , different levels of relationships among animals and inbreeding . The model included the respective SNP genotypes , sex and inbreeding coefficients as fixed effects and the genomic relationship matrix for the random genetic effect of the animal . Using MMA the NCL-region mapped to 83 . 71–84 . 719 Mb on dog chromosome 2 ( −log10p-value>200; Figure S2 ) . In the region at 81 . 17–87 . 29 Mb on CFA2 , we evaluated 14 genes for late-onset NCL in Tibetan terriers ( Table S5 ) . For sequence analysis cDNA ( Table S6 ) from two NCL-affected Tibetan terriers and one control dog was used . Polymorphisms detected ( Table S7 ) were confirmed by re-sequencing genomic DNA from four NCL-affected and four NCL-unaffected Tibetan terriers . In the 3′UTR of the canine MAPK PM20/PM21 gene one SNP ( XM_846908:c . 766T>C ) was detected showing co-segregation with the NCL-phenotype in Tibetan terriers . To evaluate this mutation the same 376 Tibetan terriers employed for association analysis were genotyped for the c . 766T>C SNP ( Table S8 ) . In this sample the NCL-associated allele C had a frequency of 0 . 265 and the frequency of T/T genotypes was 0 . 559 . The associations for genotypes and alleles were significant at −log10p>15 with χ2-values of 194 . 58 and 94 . 16 . Out of the 24 known NCL-affected dogs , 22 affected could be correctly affirmed and from the 30 NCL-carriers ascertained from the pedigrees , 27 could be correctly identified with the c . 766T>C SNP . A candidate gene in close vicinity to the highest peaks of the association analyses is the canine ATPase type 13A2 ( ATP13A2 ) encoding a lysosomal type 5 ATPase in human known to be involved in Kufor-Rakeb-syndrome , a variant of Parkinson disease . This gene is annotated at 84 . 09–84 . 11 Mb in Ensemble canine genome reference sequence but not in CanFam 2 . 1 of NCBI . Using the mRNA ( Accession number NM_001141973 ) of ATP13A2 and BLAST for canine expressed sequence tags ( ESTs ) , we found five canine ESTs for the ATP13A2 gene ( Accession numbers DN355771 , DN374414 , DN442497 , DN743972 , DN905373 ) and could predict the canine ATP13A2 gene at 84 . 09–84 . 11 Mb on CFA2 . The canine ATP13A2 gene model contains like the human transcript variant ENST00000326735 gene 29 exons with an open reading frame of 3 , 522 bp coding for a protein with 1 , 173 amino acids . The Ensemble canine cDNA reference sequence starts with exon 2 and exon 1 is missing . Re-sequencing of the canine ATP13A2 using cDNA from five NCL-affected and two NCL-unaffected Tibetan terrier confirmed the predicted gene model . Mutation analysis of the amplified sequences revealed exon skipping of exon 16 in all 5 NCL-affected dogs . To verify this mutation exon 16 and its flanking introns were sequenced using genomic DNA from seven NCL-affected and 30 NCL-unaffected Tibetan terrier . In the canine ATP13A2 gene a single base pair deletion within exon 16 ( c . 1620delG , Figure S3 ) was identified that causes skipping of exon 16 ( Figure S4 , Figure 2 ) in NCL-affected dogs . Sequencing of the whole introns 15 and 16 revealed no mutations and thus , an intronic mutation responsible for skipping exon 16 was not found . Exon skipping does not shift the open reading frame ( ORF ) but leads to a protein shortened by 69 amino acids ( Figure S5 ) . The loss of amino acids in the canine ATP13A2 protein destroys the signatures of the P-type cation transporting ATPase superfamily and sodium/potassium-transporting ATPase , the E1-E2 ATPases phosphorylation site and a part of the domain of the haloacid dehalogenase-like hydrolase ( Figure S5 ) . The single base pair deletion in exon 16 alters the restriction site of the restriction enzyme BglI . Restriction fragment length polymorphism ( RFLP ) was developed for screening the Tibetan terrier population . In dogs with the genotype homozygous for G , the 635 base pair ( bp ) PCR product was cut into two fragments with 322 and 313 bp . In NCL-cases carrying the deletion homozygously , the amplicons were not cut and only the 635 bp product could be observed . We screened 168 Tibetan terriers for the c . 1620delG using BglI . All genotyped 24 known NCL-cases were homozygous for the deletion in exon 16 and the 35 known NCL-carriers heterozygous for the deletion ( Figure 3 , Table 1 ) . Screening of the 168 Tibetan terriers revealed ten not yet known NCL-cases and further 35 NCL-carriers . The rest of the Tibetan terriers screened for c . 1620delG showed the wild type genotype . Tibetan terriers expected as free for the c . 1620delG mutation were homozygous G/G like the dog reference sequence . The frequency of the mutated allele was 0 . 309 . To assure the observed results , 144 dogs from eleven different dog breeds were genotyped for the c . 1620delG mutation . All 144 dogs did show the wild type genotype ( Table S10 ) .
A 1-bp deletion within exon 16 of ATP13A2 ( c . 1620delG ) was identified as responsible for NCL in the Tibetan terrier . All NCL-cases reported in Tibetan terriers were homozygous for this mutation and therefore heterogeneity seems unlikely for this disease in Tibetan terriers . This c . 1620delG mutation causes an alternative splicing of exon 16 but not a frameshift mutation with a premature termination codon as previously supposed [14] . As a result of the in-frame loss of exon 16 , the ATP13A2 protein is shortened by 69 amino acids . Therefore , all NCL-affected Tibetan terriers in the present study can synthesize this shortened ATP13A2 protein . In humans , all three isoforms do not lack exon 16 . This new insight on the structure of the mutated protein may explain why Tibetan terriers express only mild neurodegenerative symptoms and the onset of the disease is late in life . Exonic substitutions often create or eliminate short elements that inhibit or activate exon inclusion or splice-site selection [16]–[18] . These splicing silencers or enhancers are present both in exons ( ESSs , ESEs ) and introns ( ISSs , ISEs ) and have been derived by computational and/or experimental approaches [19]–[20] . Searching the mutated canine ATP13A2 sequences using Human Splicing Finder , version 2 . 4 ( http://www . umd . be/HSF/ ) [19] we could affirm that the 206 bp-exon 16 is no longer recognized by splicing sites because the 1-bp deletion is responsible for a broken exonic splicing enhancer motif sequence . However , the donor site at the exon 15 boundary and the acceptor site at the exon 17 boundary are still recognized . Thus , we assume that the splicing process for exon 16 fails since the c . 1620delG mutation destroys the ESE motif ( “CCGCCTG” at cDNA positions 1614–1620 ) within exon 16 . The ATP13A2 gene encodes a member of the P5 subfamily of ATPases which transports inorganic cations as well as other substrates . Mutations in this gene have been identified in Kufor-Rakeb syndrome ( KRS ) patients , a familial form of Parkinson disease ( PARK9 ) . One of the mutation studies described a 22 bp-duplication mutation in exon 16 of the human ATP13A2 gene [21] . A milder form of KRS than that reported for frameshift or truncating mutations was caused by a G504R missense mutation in exon 15 of the ATP13A2 gene , affecting the cytosolic loop close to the predicted catalytic phosphorylation site [22] . The clinical signs of PARK9 of this patient including aggressive behaviors resemble some of the signs of NCL seen in affected Tibetan terriers . Therefore , NCL in Tibetan terriers might be a mild form of KRS with Parkinson's disease-like symptoms since a truncating mutation was not present in Tibetan terriers . The protein-shortening mutation found in Tibetan terriers seems to be the reason why only a partial overlap of disease symptoms among KRS patients due to a truncating mutation and NCL-affected Tibetan terriers is observed . A wide intra-familial clinical variability of PARK9-linked disease seems to exist in humans . In agreement with the NCL-affected Tibetan terriers , there were also cases in man with no evidence for brain iron accumulation [23] . Knock-down of the PARK9 orthologue in C . elegans enhances α-synuclein ( α–syn ) misfolding [24] . This PARK9 functional relationship with the α–syn pathobiology could be confirmed in rat primary midbrain neuron cultures . On the other hand , yeast PARK9 helps to protect cells from manganese toxicity [24] . To assure that the detected deletion within exon 16 is causative for NCL in Tibetan terriers , all NCL-affected and NCL-carrier dogs were genotyped and in addition , random samples of 144 dogs from eleven different dog breeds were tested for the presence of the c . 1620delG mutation . All individuals of these nine different dog breeds were homozygous for the wild type sequence . The c . 1620delG mutation of the ATP13A2 gene was not yet reported as responsible for exon skipping and can therefore be regarded as novel in this regard . In conclusion , we identified the causal mutation for canine late-onset NCL in the Tibetan terrier . This mutation ( c . 1620delG ) is located within exon 16 of ATP13A2 and leads to a shortened protein through an alternative splicing process for exon 16 due to a broken ESE motif . The pathobiology may be mediated through the connection of ATP13A2 to the α–syn network . Tibetan terriers carrying the susceptible genotype for NCL may be a valuable model for unraveling the pathobiology of a PARK9-linked disease and testing the role of manganese toxicity in synucleinopathies . Knowledge of the causative mutation for late-onset NCL in Tibetan terriers allows genetic testing of these dogs to avoid matings of carrier animals . So the breeders are able to eliminate this disease in their breeding lines .
All animal work has been conducted according to the national and international guidelines for animal welfare . All blood-sampling of NCL-affected dogs was done in veterinary clinics for small animals during the routine of diagnosis of NCL . The blood samples of unaffected dogs were provided from veterinarians in veterinary clinics . Samples for RNA isolation were taken in veterinary clinics for small animals after euthanasia . EDTA-blood samples from affected and control dogs were collected from 1 , 371 Tibetan terriers . Thereof 24 Tibetan terriers were NCL-affected , 35 were NCL-carriers because they were parents or offspring of NCL-affected dogs and the other 1 , 312 Tibetan terriers had a clear or unknown phenotype . Except of one NCL-affected Tibetan terrier from Switzerland all NCL-affected Tibetan terriers were from Germany . Within the known NCL-carriers , three dogs were from Denmark , two from Switzerland and each one from Finland and USA . Genomic DNA was extracted from EDTA blood samples through a standard ethanol fractionation with concentrated sodiumchloride ( 6 M NaCl ) and sodium dodecyl sulphate ( 10% SDS ) . Concentration of extracted DNA was determined using the Nanodrop ND-1000 ( Peqlab Biotechnology , Erlangen , Germany ) . DNA concentration of samples for SNP chip analysis was adjusted to 50–70 ng/µl . For cDNA analysis , biopsies from conjunctiva and cerebrum of five NCL-affected dogs during the pathological examination were taken . These samples were taken 15–30 minutes after the dogs were euthanised . Tissue samples were conserved using RNA-later solution ( Qiagen , Hilden , Germany ) . The RNA was extracted from the cerebrum tissues using the RNeasy Lipid Tissue Mini Kit ( Qiagen ) and transcribed into cDNA using SuperScript III Reverse Transcriptase ( Invitrogen , Karlsruhe , Germany ) . Genome-wide association analyses were performed using the 127K canine Affymetrix SNP chip ( Affymetrix , Santa Clara , CA , USA ) . DNA from 12 Tibetan-terrier-NCL-cases and 7 Tibetan-terrier-NCL-controls were used for a genome-wide association analysis . The phenotype of controls has been assured through reports of the owners and veterinary examinations . All controls had an age of at least 10 years and all their registered offspring was ascertained as NCL clear . The controls were preferably unrelated to families with NCL affected dogs using five generation pedigree records . Quality criteria were minor allele frequencies ( MAF ) >2% and SNP genotyping rates >95% for the genome-wide association for NCL of the resulting 111 , 525 SNPs from the canine 125K SNP chip were performed using PLINK , version 1 . 07 ( http://pngu . mgh . harvard . edu/purcell/plink/ ) [25] and Tassel , version 2 . 1 [26] . Genome-wide significance was ascertained through adaptive permutation testing using a maximum of 5 , 000 , 000 permutations . Non-parametric multipoint linkage ( NPL ) analysis was performed for the NCL-affected Tibetan terriers and their relatives from three different families using MERLIN 1 . 1 . 2 [27] . Linkage among NCL and microsatellites was estimated using the proportion of alleles identical by descent ( IBD ) for affected animals . The NPL statistics Zmean and the LOD ( logarithm of the odds ) scores were employed for detection of allele sharing among affected family members . The maximum ( minimum ) achievable Zmean and LOD score were 6 . 83 ( −2 . 10 ) and 2 . 60 ( −0 . 31 ) indicating enough power to detect chromosome-wide significant linkage . In the case of no linkage , Zmean approaches the minimum achievable value due to an equal distribution of alleles among affected relatives . When linkage is present under the alternative hypothesis , the proportion of alleles IBD significantly deviates from the expected IBD proportions of the null hypothesis . We employed multipoint analyses in order to use marker information from the whole chromosome through linked informative markers and to increase power of linkage analysis . For cDNA analysis of genes in the associated genomic region annotated canine mRNA sequences were used . To assure the annotation we searched the dog expressed sequence tag ( EST ) archive ( http://www . ncbi . nlm . nih . gov/genome/seq/CfaBlast . html ) for ESTs by cross-species BLAST searches with the corresponding human reference mRNA sequences . For PCR primer design the PRIMER3 software ( http://frodo . wi . mit . edu/cgi-bin/primer3/primer3_www . cgi ) were used . Primer pairs used for cDNA amplification of the 14 positional candidate genes are given in Tables S6 and S9 . The PCR reactions were performed in a total volume of 50 µl containing 10 ng of genomic DNA as template , 10 pmol of each primer and 1 U Taq polymerase ( MP Biomedicals , Eschwege , Germany ) . Thermocycling was carried out under the following conditions: initial denaturation at 94°C for 4 min was followed by 35 cycles of 94°C for 30 s , 58–61°C for 30 s , 72°C for 1∶20 min and a final step with 72°C for 5 min before cooling at 4°C for 10 min . Primer pairs for amplification the genomic DNA of canine ATP13A2 exon 16 were as follows: ATP13A2_F 5′-GACCTGCCGTAGGGTGAAG-3′ and ATP13A2_R5′-AAGCTTCCTTCCTGGGCTAC-3′ . Amplicons were directly sequenced using an ABI 3700 capillary sequencer ( Life Technologies , Darmstadt , Germany ) . Afterwards sequence data were analyzed using Sequencher software version 4 . 7 ( GeneCodes , Ann Arbor , MI , USA ) . We genotyped 28 microsatellites on CFA2 , 8 , 18 , 22 and 37 . We used 16 markers of the canine minimal screening set 2 [28] and 12 newly developed markers . Microsatellite motifs were identified in the assembled dog sequences ( dog genome assembly 2 . 1 ) using BLAST http://www . ncbi . nlm . nuh . gov/entrez/query . fcgi ) . The criterion for markers to be included in this set was more than 15 repeats of di- tri- tetra- or pentanucleotide motifs , respectively . Prior to primer design using PRIMER3 repetitive sequences were masked employing repeatmasker ( http://www . repeatmasker . org ) . The average marker distance was about 0 . 5 Mb at 81 . 5–88 . 0 Mb on CFA2 . The PCR was carried out with an initial denaturing for 4 min at 94°C followed by 38 cycles with denaturing at 94°C for 30 sec , optimum primer annealing temperature ( Table S1 ) for 30 sec and elongation at 72°C for 45 sec . All PCR reactions were performed in 12 . 0-µl reactions using 10 pmol of each primer , 0 . 2 µl dNTPs ( 100 µM ) and 0 . 1 µl Taq-DNA-Polymerase ( 5 U/µl ) ( Q-Biogen , Heidelberg , Germany ) in the reaction buffer supplied by the manufacturer for 2 µl template DNA . The forward primers were labelled fluorescently with IRD700 or IRD800 . For the analysis of the marker genotypes , PCR products were size-fractionated by gel electrophoresis on an automated sequencer ( LI-COR , Lincoln , NE , USA ) using 4% polyacrylamide denaturing gels ( Rotiphorese Gel40 , Carl Roth , Karlsruhe ) . Allele sizes were determined using an IRD700- and IRD800-labelled DNA ladder . Genotypes were assigned by visual examination . Formalin-fixed , paraffin-embedded tissue samples of the retina , cerebellum , cerebrum and spinal cord were examined by histology . Briefly , 3 µm-thick sections were cut on a microtome and mounted on glass slides . Subsequently slides were stained with hemotoxylin and eosin ( HE ) , periodic acid-Schiff reaction ( PAS ) , Turnbull blue and luxol fast blue staining ( LFB ) .
|
The neuronal ceroid lipofuscinosis ( NCL ) is a neurodegenerative storage diseases characterized by psychomotor retardation , blindness , and premature death . NCL has been reported in several dog breeds . NCL is characterized by progressive brain and retinal atrophy and the intracellular accumulation of autofluorescent lysosomal storage bodies resembling lipofuscin . Tibetan terriers show a late-onset and lethal NCL ( age of onset 5–7 years ) with an autosomal recessive inheritance . The most frequently described first symptom is blindness in twilight . In the disease progress the affected dogs often appear nervous or anxious and the lack of motor coordination becomes more severe . In the final stages of this disease , mild but also severe seizures have been observed by the owner . There are no treatment options for affected dogs . Through a genome-wide association analysis using the 127K canine Affymetrix SNP chip , we found a 1 Mb candidate genomic region and identified ATP13A2 as the most likely candidate for NCL . A 1-base pair deletion mutation within exon 16 of the ATP13A2 gene caused the loss of an exonic splicing enhancer and , consequently , the alternative splicing lead to skipping of exon 16 . This study provides a suitable animal model for PARK9 in man to develop therapeutic approaches .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"animal",
"genetics",
"veterinary",
"neurology",
"animal",
"management",
"animal",
"breeding",
"neuroscience",
"genetics",
"and",
"genomics",
"veterinary",
"science",
"veterinary",
"medicine"
] |
2011
|
A One Base Pair Deletion in the Canine ATP13A2 Gene Causes Exon Skipping and Late-Onset Neuronal Ceroid Lipofuscinosis in the Tibetan Terrier
|
Accumulating evidence suggests that IL-9-mediated immunity plays a fundamental role in control of intestinal nematode infection . Here we report a different impact of Foxp3+ regulatory T cells ( Treg ) in nematode-induced evasion of IL-9-mediated immunity in BALB/c and C57BL/6 mice . Infection with Strongyloides ratti induced Treg expansion with similar kinetics and phenotype in both strains . Strikingly , Treg depletion reduced parasite burden selectively in BALB/c but not in C57BL/6 mice . Treg function was apparent in both strains as Treg depletion increased nematode-specific humoral and cellular Th2 response in BALB/c and C57BL/6 mice to the same extent . Improved resistance in Treg-depleted BALB/c mice was accompanied by increased production of IL-9 and accelerated degranulation of mast cells . In contrast , IL-9 production was not significantly elevated and kinetics of mast cell degranulation were unaffected by Treg depletion in C57BL/6 mice . By in vivo neutralization , we demonstrate that increased IL-9 production during the first days of infection caused accelerated mast cell degranulation and rapid expulsion of S . ratti adults from the small intestine of Treg-depleted BALB/c mice . In genetically mast cell-deficient ( Cpa3-Cre ) BALB/c mice , Treg depletion still resulted in increased IL-9 production but resistance to S . ratti infection was lost , suggesting that IL-9-driven mast cell activation mediated accelerated expulsion of S . ratti in Treg-depleted BALB/c mice . This IL-9-driven mast cell degranulation is a central mechanism of S . ratti expulsion in both , BALB/c and C57BL/6 mice , because IL-9 injection reduced and IL-9 neutralization increased parasite burden in the presence of Treg in both strains . Therefore our results suggest that Foxp3+ Treg suppress sufficient IL-9 production for subsequent mast cell degranulation during S . ratti infection in a non-redundant manner in BALB/c mice , whereas additional regulatory pathways are functional in Treg-depleted C57BL/6 mice .
Helminths are large multicellular parasites that may survive for years within their mammalian hosts despite their potential exposure to the immune system . This is achieved by active suppression of their host's immune response utilizing regulatory pathways that are intrinsic parts of the mammalian immune system [1] , [2] . Thereby , helminths have been shown to secrete analogs of regulatory cytokines , to induce regulatory receptors on their host's leukocytes and to mediate expansion and activation of regulatory cell populations . Among these , regulatory T cells are the most prominent mediators of immunological homeostasis [3] , [4] . Consequently , Treg numbers and suppressive capacity are increased in helminth-infected humans [5] , [6] , [7] , [8] , [9] and mice [10] , [11] , [12] , [13] , [14] , [15] . Depletion or functional inactivation of Treg resulted in increased immune pathology [16] , [17] , [18] , [19] , reduced parasite burden [13] , [14] , and abrogated suppression of immune response to unrelated antigens in several murine helminth infection models [20] , [21] , [22] , [23] , thus suggesting that Treg may promote parasite survival by active immune suppression [24] . Treg can be identified by constitutive expression of the IL-2 receptor alpha chain ( CD25 ) and more precisely by expression of the transcription factor forkhead box p3 ( Foxp3 ) [25] . Treg depletion or inactivation by injection of mAb specific for CD25 may interfere with effector T cell function . To circumvent this problem mouse strains that express the human diphtheria toxin receptor ( DTR ) in Treg have been generated [26] , [27] . Depletion of regulatory T cell ( DEREG ) mice are transgenic for a bacterial artificial chromosome that drives expression of a fusion protein consisting of the DTR and enhanced green fluorescent protein ( eGFP ) under the control of the Foxp3 promoter . Thus Foxp3+ Treg can be monitored by the eGFP expression in these mice and application of DT results in their rapid but transient depletion . Although this model is not feasible for long term or repeated Treg depletion , it allows the specific eradication of Foxp3+ Treg during a short period while activated effector T cells are not affected directly [28] . We use experimental infection of mice with the pathogenic nematode Strongyloides ratti to investigate the role of Foxp3+ Treg during helminth infection . This rodent parasite is a suitable model for geohelminth infections , displaying tissue migrating and intestinal life stages [29] . S . ratti infective third stage larvae ( iL3 ) penetrate the skin of their rodent host and migrate within 2 days to the head . iL3 are swallowed , reach the intestine and molt via a fourth larval stage to parasitic adults by day 5 post infection ( p . i . ) . Parasites live embedded in the mucosa of the small intestine and reproduce by parthenogenesis . Eggs and already hatched first stage larvae ( L1 ) leave the host with the feces by day 5–6 p . i . Female L1 may directly develop into iL3 and invade another host or molt to free living adults that reproduce sexually for one generation . Immune competent mice and rats resolve the infection spontaneously within 2–4 weeks and remain semi-resistant to subsequent infections [30] , [31] . Experimental infection of mice and rats induces a protective type 2 immune response characterized by induction of the cytokines IL-4 , IL-5 and IL-13 as well as production of nematode-specific IgM and IgG1 [32] , [33] . Migrating larvae in the tissue are opsonized by antibodies and complement and eradicated by granulocytes [34] , [35] , [36] , [37] . Final expulsion of parasitic adults from the small intestine is thought to be promoted by mast cells [38] , [39] , [40] although a conclusive demonstration of protective roles of mast cells would require the use of Kit-independent models [41] . Moreover , the role of IL-9-mediated immunity in eradication of migrating larvae or parasitic adults has not been investigated so far . To date DEREG mice have been used to analyze the role of Foxp3+ Treg during infection of mice with two different nematodes , S . ratti and Heligmosomoides polygyrus , an orally transmitted gastrointestinal parasite [14] , [18] . Both , S . ratti and H . polygyrus infection increased Treg numbers but did not change the frequency of Treg within the CD4+ T cell compartment as effector T cells expanded with similar kinetics . Transient Treg depletion either in C57BL/6 DEREG mice during days 4 to 6 of H . polygyrus infection or in BALB/c DEREG mice during days 0 to 2 of S . ratti infection increased the nematode-specific cellular immune response . Improved immune response did not reduce H . polygyrus parasite burden analyzed day 14 p . i . but aggravated pathology in the small intestine [18] . In contrast , improved immune response to S . ratti was correlated with dramatically reduced parasite burden in the small intestine and equally reduced larval output in the feces throughout infection [14] . Treg depletion did not improve eradication of tissue migrating S . ratti larvae but induced rapid expulsion of parasitic adults in the context of accelerated mast cell degranulation . Despite the different final outcome , both studies highlight the general importance of nematode-induced Treg in suppressing the host's immune response to prevent either expulsion and/or induction of immune pathology . The different effect of Treg depletion on resistance may be explained by the different infection models and variations in timing and methodology of Treg depletion [24] . A possible impact of the genetic background of the murine host , however , has not been investigated systematically so far . Here , we compare the effect of Treg depletion during S . ratti infection in BALB/c DEREG and C57BL/6 DEREG mice . Although Treg expanded with comparable kinetics , displayed a comparable phenotype and equally suppressed the nematode-specific humoral and cellular type 2 immune response parasite burdens were selectively reduced in BALB/c but not in C57BL/6 mice upon Treg depletion . We show that improved resistance in Treg-depleted BALB/c mice was mediated by increased IL-9 production and the protection required mast cells . In line with recent data suggesting that IL-9 plays a fundamental role in control of helminth infection and pathology [42] , [43] , we confirm that IL-9 is central for control of S . ratti infection in both , BALB/c and C57BL/6 mice . We demonstrate that Foxp3+ Treg control this axis of IL-9 production leading to mast cell-driven parasite expulsion in BALB/c mice . As IL-9 production was not significantly elevated upon Treg depletion in C57BL/6 mice , we further demonstrate genetically determined differences in control of endogenous IL-9 production .
To compare the role of Foxp3+ Treg in different mouse strains , BALB/c DEREG and C57BL/6 DEREG mice were infected by s . c . injection of S . ratti iL3 in the footpad . Foxp3+ Treg were identified as CD4+GFP+ lymphocytes at different time points during infection ( Figure S1A ) . S . ratti infection led to a rapid but transient increase in Treg numbers , first in the popliteal lymph nodes ( PLN ) that drain the site of infection ( Figure 1A ) and later in the mesenteric lymph nodes ( MLN ) that drain the small intestine where adult parasites are located ( Figure 1B ) . Upon direct comparison to C57BL/6 mice , BALB/c mice displayed an earlier Treg expansion in the MLN but both strains showed maximal Treg numbers by day 14 p . i . Expanding Treg displayed an activated effector-memory phenotype in both strains indicated by expression of the integrin CD103 [44] ( Figure 1A , B and Figure S1B ) . Treg numbers returned to naïve levels once infection was resolved i . e . by day 36 p . i . As the numbers of CD4+GFP− effector T cells expanded and contracted with similar kinetics ( Figure 1A and 1B ) , the frequencies of Treg within the CD4+ T cell compartment remained constant in PLN and MLN in both mouse strains ( data not shown ) . Expression of Helios and Neuropilin ( CD304 ) has been used to distinguish thymus-derived Treg from Treg that were induced in the periphery [45] , [46] , [47] . The frequency of Helios+ and Neuropilin+ Treg in MLN ( Figure 1C and S1C ) and spleen ( data not shown ) of BALB/c and C57BL/6 mice did not change during S . ratti infection thus suggesting no differences in the origin of expanding Treg in both strains . Transient depletion of Foxp3+ Treg during the first days of S . ratti infection was achieved by three consecutive injections of DT in DEREG mice or non-transgenic littermates ( Figure 2A ) . Consistent with our previous results [14] , Treg depletion in BALB/c DEREG mice resulted in reduced parasite burden in the small intestine at day 6 p . i . Strikingly , Treg depletion within the same infection experiments did not change parasite burden in C57BL/6 DEREG mice ( Figure 2B ) . To monitor the course of S . ratti infection , we quantified the released eggs and first stage larvae in the feces by qPCR specific for the Strongyloides 28S RNA gene [33] . Reduced numbers of parasitic adults in the small intestine of Treg-depleted BALB/c DEREG mice were reflected by reduced larval output throughout infection as we had shown before [14] ( Figure 2C ) . We did not record statistically significant differences in the larval output of C57BL/6 mice in the presence or absence of Treg until clearance of infection at day 28 p . i . ( Figure 2D ) . Thus Treg depletion in C57BL/6 mice did not improve host defense at any time point of infection . As C57BL/6 mice were more susceptible to S . ratti infection , we compared the outcome of Treg depletion in C57BL/6 mice that received a lower infection dose to BALB/c mice that received the standard infection dose . Treg depletion did not reduce parasite burden in C57BL/6 mice carrying less than 10 parasitic adults in the small intestine while a mean parasite burden of 20 adults in BALB/c mice was reproducibly reduced by Treg depletion ( Figure S2 ) . Rapid clearance of S . ratti in Treg-depleted BALB/c DEREG mice is not achieved by improved killing of migrating larvae in the tissue but by enhanced expulsion of adult parasites from the small intestine [14] . Studies using Kit-mutant mice suggested that efficient expulsion of S . ratti and S . venezuelensis adults from the small intestine depends on mast cells [38] , [48] , and improved parasite expulsion in Treg-depleted BALB/c mice was correlated to increased and accelerated mast cell degranulation [14] . Therefore , we recorded the concentration of mouse mast cell protease-1 ( MMCP-1 ) that is released upon degranulation of mucosal mast cells [49] in the serum of BALB/c and C57BL/6 mice ( Figure 2E ) . Both mouse strains displayed comparable increases in circulating MMPC-1 in the presence of normal Treg frequencies . Depletion of Foxp3+ Treg in BALB/c DEREG mice reproducibly increased and accelerated this mast cell degranulation , yielding maximal MMCP-1 concentrations in the serum by day 6 p . i . In sharp contrast , Treg depletion had no impact on mast cell degranulation in C57BL/6 DEREG mice . The absent effect of Treg depletion in C57BL/6 DEREG mice was not due to differences in depletion efficacy or repopulation kinetics ( Figure 3 ) . Our transient depletion regime ( Figures 3A ) reduced the frequency of Foxp3+ Treg within the CD4+ T cell population by 90% until day 2 p . i . in both strains ( Figure 3B–E ) . Repopulation of the Treg compartment occurred rapidly in peripheral blood ( Figure 3C ) and spleen ( Figure 3E ) and with slower kinetics in the MLN ( Figure 3D ) . Comparison of Foxp3+ Treg repopulation velocity in BALB/c and C57BL/6 mice revealed no differences ( Figures 3C–E ) . The missing effect of Treg depletion in C57BL/6 DEREG mice also does not reflect different susceptibility of BALB/c and C57BL/6 mice to DT-mediated side effects , since DT treatment had no effect in non-transgenic littermates of both strains . Taken together these results show that despite comparable induction of Foxp3+ Treg numbers during S . ratti infection , Treg depletion was translated into accelerated mast cell degranulation and improved expulsion of parasitic adults selectively in BALB/c mice . We recognize that depletion of 2–5% of leukocytes in vivo may cause effects that are independent of nature and function of the depleted cell type . In this context it was shown that DT-mediated depletion of dendritic cells caused neutrophilia as a side effect [50] . Increased number of neutrophils subsequently improved control of the employed ultra pathogenic Escherichia coli infection model independent of presence or absence of dendritic cells . We observed a two-fold increase in the frequency of Gr1+CD11b+ granulocytes in the peripheral blood of both DT treated BALB/c DEREG and C57BL/6 DEREG mice ( Figure 4A ) . As granulocytes have been shown to kill migrating S . ratti larvae in the tissue [34] and the absolute expansion of granulocytes in the peripheral blood was more pronounced in BALB/c mice compared to C57BL/6 DEREG mice , we wanted to test if the differential mobilization of granulocytes caused the improved resistance in Treg-depleted BALB/c DEREG mice . To this end we depleted Gr1+ cells by anti-Gr1 mAb injection in DT treated BALB/c DEREG and non-transgenic littermates ( Figure 4B ) one day before infection . Depletion of granulocytes increased parasite burden in general , i . e . in the absence and presence of Treg ( Figure 4C ) . This finding most likely reflects the higher number of larvae that reach the small intestine in the absence of attacking granulocytes . Additional Treg depletion in anti-Gr1 treated mice on the background of a higher S . ratti inoculum still resulted in a statistically significant decrease in the number of parasitic adults in the intestine ( Figure 4C ) . Also depletion of Gr1+ cells after the tissue migrating phase , i . e . at day 3 p . i . , did not abrogate the beneficial effect of Treg depletion in BALB/c mice ( Figure S3 ) . Although we cannot formally exclude a contribution of granulocytes to final expulsion of S . ratti from the small intestine in the presence and absence of Treg , these results show that improved resistance to S . ratti infection in Treg-depleted BALB/c DEREG mice did not depend on granulocytes and thus was not caused by additional mobilization of granulocytes in DT treated DEREG mice . Since control of Strongyloides infection was reported to depend on a Th2 immune response [51] , [52] , we compared the S . ratti-specific B and T cell response in BALB/c and C57BL/6 DEREG mice in the presence and absence of Treg . Concentrations of total IgE ( Figure 5A ) in the serum increased in the absence of Treg in both mouse strains . The early S . ratti-specific IgM response had slightly increased titers upon Treg depletion in BALB/c ( p = 0 . 09 ) and C57BL/6 ( p = 0 . 02 ) mice ( Figure 5B ) . Antigen-specific IgG1 was not detectable at day 7 p . i . ( data not shown ) . Antigen-specific stimulation of splenocytes derived from BALB/c and C57BL/6 mice that had been infected for 6 days induced a comparable production of the Th2 signature cytokines IL-13 and IL-4 ( Figure 5C , 5D , upper panel ) . Treg depletion led to a clearly increased production of these cytokines in both mouse strains . Likewise , production of antigen-specific IL-10 increased upon Treg depletion in BALB/c and C57BL/6 DEREG mice to the same extent ( Figure 5E , upper panel ) . Also IL-13 , IL-4 and IL-10 responses to CD3 engagement i . e . polyclonal T cell activation increased in both strains in the absence of Foxp3+ Treg ( Figure 5C–E , lower panel ) . In order to understand the differential impact of Treg depletion on mast cell degranulation in BALB/c and C57BL/6 mice we analyzed the production of IL-3 and IL-9 . IL-3 functions as growth factor for basophils and mast cells and was shown to contribute to expulsion of Strongyloides venezuelensis [53] , [54] . Treg depletion increased IL-3 production in response to S . ratti antigen and to CD3 engagement in BALB/c and C57BL/6 mice to the same extent ( Figure 5F ) , thus providing no explanation for the observed difference in mast cell activation . IL-9 is a cytokine with pleiotropic function that also contributes to mastocytosis and mast cell activation [55] . BALB/c mice responded to Treg depletion with a statistically significant increase in IL-9 production in response to either antigen-specific stimulation or polyclonal T cell stimulation by CD3 engagement ( Figure 5G ) . A weaker but statistically not significant up-regulation of IL-9 was also observed in Treg-depleted C57BL/6 DEREG mice . C57BL/6 mice produced generally less IL-9 than BALB/c mice , and the amount of IL-9 produced in the absence of Treg in C57BL/6 DEREG mice did not exceed the amount of IL-9 produced in BALB/c mice with normal Treg frequencies . The different impact of in vivo Treg depletion on IL-9 secretion was not observed in vitro . Purified BALB/c- and C57BL/6-derived Treg suppressed IL-2 but also IL-9 secretion by splenocytes derived from S . ratti- mice in syngenic and allogenic combination to the same extent ( Figure S4 ) . Thus the differential regulation of IL-9 was , at least in vitro , not Treg-intrinsic but reflected either a limited capacity of C57BL/6 mice to produce IL-9 ( Figure 5G ) or the presence of redundant regulatory pathways in C57BL/6 mice that maintained control of IL-9 production in the absence of Foxp3+ Treg . Taken together , depletion of Foxp3+ Treg in vivo increased most effectors of the adaptive type 2 immune response to S . ratti infection in BALB/c and C57BL/6 mice to the same extent . Among the parameters analyzed in this study , selectively IL-9 production and mast cell degranulation were significantly increased in Treg-depleted BALB/c mice . This leads to the question if the improved resistance in Treg-depleted BALB/c mice was caused by this increase in IL-9 production and the subsequent mast cell degranulation . A recent study demonstrated a fundamental role for IL-9 in control of Nippostrongylus brasiliensis infection in BALB/c mice [42] . However , the importance of IL-9 in control of gastrointestinal nematode infection depends on the nematode species and the genetic background of the host [56] , [57] , [58] , [59] , [60] . As the role of IL-9 in immunity to S . ratti infection has not been investigated so far , we first compared the impact of either recombinant IL-9 treatment or IL-9 neutralization in BALB/c and C57BL/6 mice ( Figure 6 ) . Forced increase of systemic IL-9 concentration by injection of IL-9 ( Figure 6A ) resulted in a significant reduction of parasitic adults in the small intestine of BALB/c ( Figure 6B ) and also C57BL/6 mice ( Figure 6C ) . Neutralization of endogenous IL-9 that was produced during S . ratti infection resulted in a reciprocally increased parasite burden in both strains . ( Figure 6D–F ) . IL-9 neutralization also reduced mucosal mast cell degranulation in BALB/c and C57BL/6 mice , demonstrating that endogenous IL-9 was required for degranulation ( Figure 6G , 6H ) . These results show that IL-9-driven mast cell degranulation was associated with rapid expulsion of S . ratti in BALB/c and C57BL/6 mice . The fact that Treg depletion increased production of IL-9 in BALB/c DEREG mice but much less in C57BL/6 DEREG mice ( Figure 5G ) strongly suggests that this increased IL-9 contributed to improved resistance in the BALB/c strain . To test if enhanced resistance in Treg-depleted BALB/c mice was directly caused by the increased IL-9 production , we neutralized IL-9 and , as a control Th2 cytokine , IL-13 during Treg depletion in S . ratti-infected BALB/c DEREG mice ( Figure 7A ) . As expected , Treg depletion reduced the parasite burden in the small intestine at day 6 p . i . ( Figure 7B , 7C ) and accelerated the degranulation of mast cells in the presence of endogenous IL-9 ( Figure 7D ) . Additional neutralization of IL-9 , but not neutralization of IL-13 , abrogated this enhanced resistance . Numbers of parasitic S . ratti adults in the small intestine were alike in anti-IL-9 treated BALB/c mice in the presence and absence of Foxp3+ Treg ( p = 0 . 97; Figure 7B ) whereas Treg depletion still led to drastically reduced numbers of parasitic adults in anti-IL-13 treated mice ( Figure 7C ) . Neutralization of endogenous IL-9 also abrogated the increased mast cell degranulation in Treg-depleted BALB/c mice until day 5 p . i . One day later , at day 6 p . i . , mast cell degranulation was also increased in Treg-depleted BALB/c mice that received anti-IL-9 treatment , although parasite burdens were not reduced in this group . This finding suggests that mast cell degranulation during the first days of infection was relevant for improved resistance in Treg-depleted BALB/c mice . To further elucidate the kinetics of IL-9 production relevant for the early mast cell degranulation and protection we neutralized IL-9 during S . ratti infection in Treg-depleted BALB/c mice at different time points ( Figure 8A ) . IL-9 neutralization throughout infection , achieved by anti-IL-9 injections at the day of infection and at day 3 p . i . , abrogated the beneficial effect of Treg depletion in BALB/c mice as expected ( Figure 8B ) . Neutralization of endogenous IL-9 in Treg-depleted BALB/c mice at later time points i . e . starting day 3 p . i . with a second injection at day 5 p . i . led to an intermediate phenotype . The numbers of parasitic S . ratti adults in Treg-depleted mice receiving late IL-9 neutralization were significantly lower than parasite burden in Treg-depleted BALB/c mice where IL-9 was neutralized from the beginning of infection . We still observed a trend towards higher parasite burden compared to the low parasite number achieved upon Treg depletion in the presence of endogenous IL-9 , suggesting that IL-9 was important also after day 3 p . i . Kinetics of mast cell degranulation reciprocally reflected the parasite burden ( Figure 8C ) . Treg depletion increased mast cell degranulation and the additional early IL-9 neutralization but not the delayed IL-9 neutralization prevented that rapid mast cell degranulation . To identify the source of IL-9 , we performed intracellular cytokine staining ( Figure S5 ) . In line with a recent report [42] IL-9 was produced by a very low frequency of CD4+ and CD4− cells in the spleen and mesenteric lymph node of nematode- mice and Treg depletion increased the frequency of both , CD4+ and CD4− IL-9-expressing cells . Taken together , these results show that increased production of IL-9 during the first days of infection was central for establishment of enhanced resistance in Treg-depleted BALB/c DEREG mice . To provide a causative link between the observed IL-9 dependent mast cell degranulation ( Figures 6–8 ) and the improved resistance in Treg-depleted BALB/c DEREG mice , we crossed BALB/c DEREG mice to a recently generated mast cell-deficient mouse strain [61] . In Cpa3CRE knockin mice , Cre-recombinase is expressed under control of the carboxypeptidase A3 ( Cpa3 ) promoter which results in constitutive and complete deletion of mucosal and connective tissue mast cells , and mild reduction in splenic basophil numbers , whereas other features of the immune system are normal . This is in contrast to Kit-mutant mouse models that carry significant additional immune aberrations due to the dysfunctional Kit [41] , [62] . Crossing heterozygous BALB/c DEREG to heterozygous BALB/c Cpa3CRE mice yielded four different genotypes in the first generation offspring: BALB/c Cpa3WT that contained normal mast cell numbers and did not express DTR on Treg , BALB/c DEREG Cpa3WT that contained normal mast cell numbers but expressed DTR on Treg , BALB/c Cpa3CRE that were mast cell-deficient but did not express DTR on Treg , and BALB/c DEREG Cpa3CRE that were mast cell-deficient and expressed DTR on Treg . All groups were treated with DT and subsequently infected with S . ratti . Comparison of parasite burden in the presence of Treg in mast cell-competent BALB/c Cpa3WT mice and mast cell-deficient BALB/c Cpa3CRE mice demonstrated that mast cell-deficiency as such increased parasite burden ( Figure 9A ) . While these data are in agreement with older studies that employed Kit-mutant mice as models for mast cell deficiency in S . ratti [38] and S . venezuelensis [48] infection , it was identified a key role for mast cells in protection against intestinal nematode infection , given that several suggested roles for mast cells have not been confirmed in Kit-independent mast cell deficient mice [61] , [63] , [64] , [65] . Treg depletion in the presence of mast cells reduced parasite burden day at 6 p . i . ( Figure 9A ) and increased production of IL-9 ( Figure 9B ) as we had shown before ( Figures 2 , 4 , 7 , 8 ) . Additional mast cell deficiency completely abrogated improved resistance in Treg-depleted BALB/c DEREG Cpa3CRE mice ( Figure 9A ) , while IL-9 production was still increased . Thus increased IL-9 production that was caused by Treg depletion in vivo was only translated into improved resistance in the presence of mast cells . These results strongly suggest that IL-9 activated mast cells represent the central effector cells mediating rapid expulsion of S . ratti in Treg-depleted BALB/c mice .
Comparing the role of Foxp3+ Treg during nematode-induced immune evasion in fully susceptible BALB/c and C57BL/6 mice we report a pronounced difference . Parasite survival in BALB/c mice was strictly dependent on the presence of Foxp3+ Treg whereas in C57BL/6 mice prolonged parasite survival was still established in the absence of Foxp3+ Treg . Different impact of Foxp3+ Treg depletion on parasite burden did not reflect different kinetics of parasite eradication . Analysis of S . ratti larvae in the feces showed that the kinetics of infection and clearance in the presence of Treg were comparable in both strains . While S . ratti larval output was reduced upon Treg depletion in BALB/c mice throughout infection , larval output remained unchanged in Treg-depleted C57BL/6 mice until natural clearance of infection . Different impact of Foxp3+ Treg on host defense was not due to differences in Treg induction or depletion nor to DT-mediated side effects . Expansion of Foxp3+ Treg was comparable in both strains displaying maximal numbers at day 2 p . i . in the lymph nodes draining the site of infection and at day 14 p . i . in the lymph nodes draining the intestine i . e . the environment of parasitic adults . Also contraction of the Foxp3+ Treg compartment to naïve levels after resolution of infection was alike . Thus spatial and temporal induction of Foxp3+ Treg correlated with parasite migration through the host in BALB/c and C57BL/6 mice . Depletion efficacy and repopulation kinetics were also identical in BALB/c and C57BL/6 mice , as we controlled by monitoring Foxp3+CD4+ Treg in peripheral blood and in the lymphatic organs . Expanding Treg displayed an activated CD103+ phenotype [44] in both strains . Neither BALB/c nor C57BL/6 mice showed preferential expansion of natural , thymus-derived Treg since frequencies of Helios+ and Neuropilin+ Treg remained stable during infection . Biologic function of Foxp3+ Treg in vivo was apparent in both strains because Treg depletion led to a distinctive increase in the cellular and humoral type 2 immune responses: i . e . production of IL-4 , IL-13 , IL-3 , IL-10 , IgM and IgE . Although a functional type 2 immune response has been reported to be central for the control of Strongyloides infection [51] , [52] , these effector molecules obviously did not mediate accelerated eradication of S . ratti upon Treg depletion in C57BL/6 mice . One concern upon comparing BALB/c and C57BL/6 mice is that the latter are more susceptible to S . ratti infection , displaying between two and five times more parasitic adults in the small intestine as BALB/c mice infected with the same iL3 batch . Immune control of this high parasite burden may be difficult even in the context of improved immune responses . To address this possibility , we compared the effect of Treg depletion in low dose infected C57BL/6 mice to BALB/c mice receiving the standard infection dose . C57BL/6 mice that carried 10 parasites per mouse still did not respond to Treg depletion with improved parasite expulsion whereas BALB/c mice that carried two times more parasites displayed drastically reduced parasite numbers upon Treg depletion . Thereby , we ruled out the higher parasite burden as such as explanation for the missing effect of Treg depletion in C57BL/6 mice . Furthermore , we show that in principle reduction of a high parasite burden can be achieved because IL-9 treatment reduced the naturally high parasite burden in C57BL/6 mice and depletion of granulocytes increased the naturally lower parasite burden in BALB/c mice more than two-fold but still Treg depletion led to improved S . ratti expulsion . These results strongly suggest that within our model of murine S . ratti infection Treg controlled mast cell degranulation via control of IL-9 that acted on mast cells and not via cellular interaction , as was shown in a model of IgE induced anaphylaxis [66] . We did not observe differential regulation of any other effector relevant for infection control apart from IL-9 production and mast cell activation . Therefore , we argue that these mechanisms are causative for the observed parasite reduction , yet additional mediators involved in the chain of reactions that lead to parasite clearance cannot be excluded . Regarding the potential cellular source ( s ) of this protective IL-9 , Treg [67] , [68] , Th2 , Th9 [55] , [69] and innate lymphoid cells ( ILC ) [70] have been described . We observed both IL-9 producing CD4+ and CD4− cells during infection at very low frequencies . While IL-9 production by Treg themselves is unlikely in our system as we observed increased IL-9 production specifically upon Treg depletion , our results suggest contribution to IL-9 production by Th9 cells and CD4− ILC . In strong support of this reasoning , a novel IL-9 reporter mouse revealed IL-9 production at comparably low frequencies by Th9 cells and ILC during N . brasiliensis infection [42] . By neutralizing IL-9 at different time points post infection we further show that IL-9 neutralization in Treg-depleted BALB/c mice later than day 2 p . i . did not restore the susceptible phenotype of BALB/c mice containing normal Treg frequencies . This finding suggests that the IL-9 that was relevant for rapid mast cell degranulation and parasite expulsion was produced during the first days of infection . As Treg depletion also reduced worm burden only if Treg were depleted during the first days of infection [14] , we suggest that Treg control early IL-9 production in BALB/c mice . Our results consent with the established function for IL-9 in the recruitment of mast cells and promotion of mastocytosis in general [56] , [71] . The particular importance of IL-9 and IL-9-mediated mast cell activation for host defense during gastrointestinal nematode infection , however , depends on the nematode species and the genetic background of the host . N . brasiliensis-infected IL-9-deficient 129×C57BL/6 ( F2 ) mice displayed reduced MMCP-1 activity in the small intestine but still expelled the parasites as efficiently as wildtype mice in one study [56] . A recently generated IL-9-deficient mouse model on BALB/c background displayed reduced mast cell numbers but was additionally more susceptible to N . brasiliensis infection [42] . Moreover , adoptive transfer of Th9 cells reduced N . brasiliensis parasite burden in either IL-9 deficient or T cell- and B cell- deficient RAG−/− mice , thus demonstrating a fundamental role of T cell-derived IL-9 in eradication of intestinal nematode infection . This is in line with earlier studies showing the general importance of IL-9 by any source in control of intestinal nematode infection . Forced increase in systemic IL-9 concentration by injection of a IL-9 transgenic ( tg ) cell line accelerated expulsion of Trichuris muris in C57BL/6 mice [57] and IL-9 tg FVB mice displayed faster clearance of T . muris infection in the context of increased mastocytosis [57] . Reciprocal neutralization of endogenous IL-9 that was achieved by vaccination against IL-9 before nematode infection rendered resistant C57BL/6 mice susceptible for T . muris infection [59] . IL-9 tg mice also expulsed Trichinella spiralis rapidly from the small intestine in a mast cell dependent manner [58] . Increased T . spiralis expulsion in IL-9 treated C57BL/6 mice was correlated to increased muscle contractility and MMCP-1 production in the small intestine [60] . We analyzed the impact of IL-9 during Strongyloides infection and showed that administration of IL-9 reduced parasite burden while neutralization of IL-9 reciprocally increased parasite burden in the small intestine of S . ratti-infected BALB/c and C57BL/6 mice . Increased parasite burden upon IL-9 neutralization was accompanied by reduced mast cell degranulation in both strains , suggesting that endogenous IL-9 triggered mast cell degranulation during S . ratti infection in BALB/c and C57BL/6 mice in the presence of normal Treg frequencies . Employing Kit-independent mast cell-deficient mice , we provided direct evidence that the increase in IL-9 production observed in Treg-depleted BALB/c mice triggered mast cells to mediate rapid expulsion of S . ratti subsequently . We hypothesize that the non-significant increase in IL-9 production observed in C57BL/6 mice upon Treg depletion was not sufficient to increase mast cell degranulation and enhance parasite expulsion . Since treatment with recombinant IL-9 readily reduced parasite burden also in C57BL/6 mice , the function of IL-9 during S . ratti infection apparently is comparable in both strains whereas the control of IL-9 production in vivo is maintained differently . The different impact of in vivo Treg depletion on IL-9 production was not Treg-intrinsic . Purified BALB/c- and C57BL/6-derived Treg suppressed IL-2 but also IL-9 secretion by syngenic and allogenic effector cells to the same extent . Although an older study reported that BALB/c-derived CD25+CD4+ Treg suppressed in vitro proliferation of syngenic effector T cells more efficiently than BALB/c-derived Treg , this difference was tracked down to C57BL/6-derived effector T cells that were more difficult to suppress by both , BALB/c and C57BL/6 Treg [72] . Since we detected no Treg-intrinsic differences in vitro , the differential regulation of IL-9 in BALB/c and C57BL/6 mice that was reproducibly apparent in vivo can be explained either by a limited capacity of C57BL/6 mice to produce IL-9 or by the presence of redundant regulatory elements in C57BL/6 mice that continued to control IL-9 production in the absence of Foxp3+ Treg . Along the lines of this working hypothesis it is tempting to suggest that the increased susceptibility of C57BL/6 mice to S . ratti infection in general was due to reduced availability of IL-9 as a consequence of more stringent regulation . The reduced susceptibility to S . ratti infection displayed by BALB/c mice on the other hand may reflect increased availability of IL-9 due to the less stringent regulation of endogenous IL-9 production and subsequent mast cell degranulation . Genetically determined differences in the impact of Treg on immune regulation during infection as well as strain-specific differences in the impact of IL-9-driven and mast cell-mediated intestinal inflammation have been reported in other models . Comparing expansion and phenotype of Foxp3+ Treg and Foxp3− effector T cells during Litomosoides sigmodontis infection in susceptible BALB/c and resistant C57BL/6 mice Taylor et al . described comparable recruitment and expansion of Foxp3+ Treg in both strains [13] . Nevertheless , in vivo effector T cell proliferation and expression of the costimulatory molecule GITR was more pronounced in semi-permissive C57BL/6 mice . Since depletion of Treg in susceptible BALB/c mice prior to infection reduced the parasite burden at day 60 p . i . , it is conceivable that Treg promoted nematode-induced immune evasion in BALB/c mice and contributed to the susceptibility of this genotype . Also the impact of Treg depletion or Treg transfer on the course of Toxoplasma gondii [73] and Mycobacterium tuberculosis [74] infection was more pronounced in BALB/c than in C57BL/6 mice , again suggesting different levels of Treg redundancy in different mouse strains . IL-9 induced mastocytosis and subsequent increase in intestinal permeability mediates intestinal anaphylaxis in a murine model for food allergy independent of other Th2 associated effectors [75] . This intestinal anaphylaxis can be induced by vaccination with Alum emulsified Ovalbumin prior intra-gastric Ovalbumin challenge selectively in BALB/c mice and not in C57BL/6 mice [76] . It remains to be elucidated if the resistance to IL-9-driven and mast cell-induced intestinal anaphylaxis in C57BL/6 mice reflected a different function of IL-9 and mast cells in the different mouse strains or different control of IL-9 production . We did not identify a putative regulatory element that maintained control of IL-9 production , mast cell degranulation and thus S . ratti survival in C57BL/6 mice in the absence of Foxp3+ Treg so far . The thorough analysis of the different modes of immune regulation induced during nematode infection in inbred mouse models may eventually improve the interpretation of the diverse results observed in studies within the helminth-infected human population .
Animal experimentation was conducted at the animal facility of the Bernhard Nocht Institute for Tropical Medicine in agreement with the German animal protection law under the supervision of a veterinarian . The experimental protocols have been reviewed and approved by the responsible federal health Authorities of the State of Hamburg , Germany , the “Behörde für Gesundheit und Verbraucherschutz” permission number 54/10 and 55/13 . Mice were sacrificed by cervical dislocation under deep CO2 narcosis . DEREG mice were generated by injecting a BAC directly into fertilized C57BL/6 oocytes [26] . The resulting C57BL/6 DEREG founder was backcrossed to BALB/c thus ensuring similar insertion sites of the BAC into the genome in both strains . Cpa3CRE mice were generated by homologous recombination of Cre recombinase into the Cpa3 locus and backcrossed to BALB/c mice for 19 generations [61] . Heterozygous C57BL/6 DEREG mice were mated with wildtype C57BL/6 mice , and heterozygous BALB/c DEREG mice were bred with wildtype BALB/c or intercrossed with heterozygous BALB/c Cpa3CRE mice in the animal facilities of the Bernhard Nocht Institute for Tropical Medicine in order to provide littermates for the experiments . Wistar rats were purchased from Charles River ( Sulzfeld , Germany ) . Animals were kept in individually ventilated cages under specific pathogen-free conditions and used at the age of 6–10 wk ( mice ) or 4–8 wk ( rats ) . The S . ratti cycle was maintained by serial passage of S . ratti through Wistar rats . S . ratti iL3 were purified from charcoal feces cultures as described [29] . Prior to infection , iL3 were stored overnight in PBS supplemented with penicillin ( 100 U/mL ) and streptomycin ( 100 µg/mL ) . BALB/c DEREG , C57BL/6 DEREG and non-transgenic BALB/c and C57BL/6 littermates were infected by s . c . injection of 2000 or 200 purified iL3 in 30 µl PBS into the hind footpad . Groups of mice received 0 . 5 µg DT ( Merck , Darmstadt , Germany ) dissolved in PBS ( pH 7 . 4 ) i . p . on three consecutive days , starting one day prior to S . ratti infection . Treg depletion was routinely controlled by analysis of peripheral blood samples for GFP , Foxp3 , and CD4 expression at day 2 p . i . ( and day 6 p . i . for the experiment shown in Figure 3 ) . Recombinant IL-9 ( eBioscience , San Diego , USA ) was administered i . p . to BALB/c and C57BL/6 ( 200 ng/mouse and time point ) at the indicated time points . For neutralization of IL-9 BALB/c and C57BL/6 mice or BALB/c DEREG and littermate control mice received 100 µg anti–IL-9 mAb ( clone MM9C1 , BioXCell , West Lebanon , USA ) i . p . at the indicated time points . For depletion of granulocytes mice received 300 µg anti-Gr1 ( clone RB6-8C5 ) i . p . either one day before infection or at day 3 p . i . To count the number of adult parasitic females in the gut , the small intestine was flushed slowly with tap water to remove feces , sliced open longitudinally and incubated at 37°C for 3 h in a petri dish with tap water . The released adult females were collected by centrifugation for 5 min at 1200 rpm and counted . To quantify the release of S . ratti larvae by infected mice , the feces of individual mice was collected over 24 h periods and DNA from representative 200 mg samples was extracted as described [77] . 200 ng DNA was used as a template for qPCR specific for S . ratti 28 S ribosomal RNA gene as described [33] . For analysis of cellular responses mice were sacrificed at the indicated time points and spleen and MLN were dissected . A total of 2×105 spleen or MLN cells were cultured in 3–5 replicates 96-well round-bottom plates in RPMI 1640 medium supplemented with 10% FCS , 20 mM HEPES , L-glutamine ( 2 mM ) , and gentamicin ( 50 µg/mL ) at 37°C and 5% CO2 . The cells were stimulated for 72 h with medium , anti-mouse CD3 ( 145-2C11 , 1 µg/mL ) , or S . ratti iL3 lysate ( 20 µg/mL ) that was prepared as described [33] . The supernatant was harvested for analysis of cytokine production by ELISA . The supernatants of spleen cell cultures derived from infected mice incubated with medium did not contain detectable concentrations of IL-3 , IL-4 , IL-9 , IL-10 or IL-13 . For analysis of serum antibodies and mouse mast cell protease-I ( MMCP-I ) , blood was collected from infected mice at the indicated time points and allowed to coagulate for 1 h at room temperature ( RT ) . Serum was collected after centrifugation at 10 , 000× g for 10 min at RT and stored at −20°C for further analysis . Strongyloides-specific Ig in the serum was quantified by ELISA , as described [33] . Briefly , 50 µL/well S . ratti Ag lysate ( 2 . 5 µg/mL ) in PBS was coated overnight at 4°C on Microlon ELISA plates ( Greiner , Frickenhausen , Germany ) . Plates were washed four times with PBS 0 . 05% Tween 20 and blocked by incubation with PBS 1% BSA for 2 h at RT . Serial dilutions of sera in PBS 0 . 1% BSA were incubated in duplicate , adding 50 µL/well overnight at 4°C . Plates were washed five times , and Strongyloides-specific Ig was detected by incubation with 50 µL/well horseradish peroxidase ( HRP ) -conjugated anti-mouse IgM ( Zymed Karlsruhe , Germany ) for 1 h at RT . Plates were washed five times and developed by incubation with 100 µL/well tetramethylbenzidine 0 . 1 mg/ml , 0 . 003% H2O2 in 100 mM NaH2PO4 ( pH 5 . 5 ) for 2 . 5 min . Reaction was stopped by addition of 25 µL/well 2 M H2SO4 , and OD at 450 nm ( OD450 ) was measured . The titer was defined as the highest dilution of serum that led to an OD450 above the doubled background . Background was always below 0 . 15 OD450 . Concentration of IgE was quantified using the IgE ELISA kit ( BD , Heidelberg Germany ) and MMCP-I in serum was detected using the MMCP-I ELISA Ready-SET-Go kit ( eBioscience , San Diego , USA ) both according to the manufacturers recommendations . IL-3 , IL-4 , IL-10 and IL-13 ) in culture supernatants were measured using DuoSet ELISA development kits ( R&D Systems , Wiesbaden , Germany ) , according to the manufacturer's instructions . IL-9 detection was performed by coating with 2 µg/mL anti-IL-9 Ab ( BD , Heidelberg , Germany ) overnight at 4°C . Plates were blocked with 10% FCS/0 . 05%Tween/PBS for 2 h at RT . Samples and recombinant IL-9 standard ( Peprotech , Hamburg , Germany ) were incubated overnight and detection was performed with an anti-IL-9-biotin AB ( BD , Heidelberg , Germany ) for 1 h at RT and subsequent Streptavidin-HRP incubation for 20 min before development with 100 µL/well tetramethylbenzidine 0 . 1 mg/ml , 0 . 003% H2O2 in 100 mM NaH2PO4 ( pH 5 . 5 ) . The reaction was stopped after 10 min by adding 25 µL of 2 M H2SO4 . To prevent unspecific binding of mAbs , all samples were pre incubated with 25 µL of Fc block at 4°C for 10 min . Surface staining was carried out for 20 min at 4°C using Allophycocyanin ( APC ) -labeled anti-CD4 ( clone RM4-5; BD Heidelberg , Germany ) or Brilliant Violet 510-labeled anti-CD4 ( clone RM4-5 , Biolegend ) , Phycoerythrin ( PE ) -labeled anti-CD 304 ( Neuropilin-1 , clone 3E12 , Biolegend ) , PE- or FITC-labeled anti-CD103 ( clone 2E7 , Biolegend ) , Peridinin chlorophyll protein-cyanine5-labeled anti-CD11b ( clone M1/70 , BD , Heidelberg , Germany . ) and PE-labeled anti-Gr-1 ( clone 1A8 , BD ) . For intracellular staining cells were permeabilized with 250 µl fixation/permeabilization buffer for 30 min at 4°C , washed with permeabilization buffer and stained with PE- or Alexa Fluor 700-labeled anti-Foxp3 ( clone FJK-16s , eBiosciences , SanDiego , USA ) and APC–labeled anti-Helios ( clone 22F6 , Biolegend ) . For analysis of Neuropilin and Helios expression cells were also stained for 30 min on ice with LIVE/DEAD Fixable Blue Dead Cell Stain Kit for UV excitation according to the manufacturers recommendation ( Life Technologies , Darmstadt , Germany ) . Samples were measured on a BD FACSCalibur or LSRII and analyzed using FlowJo software . Statistical analysis was performed with GraphPad Prism software ( San Diego ) using the One way ANOVA followed by the Bonferroni post test or the students t test to calculate the significance of differences between multiple or between two groups , respectively . The data are presented as mean ± SEM; p≤0 . 05 was considered statistically significant .
|
Parasitic worms are large multicellular organisms that manage completion of their life cycles despite exposure to their host's immune system . To avoid expulsion , parasitic worms actively suppress their host's immune response . Here we show that the pathogenic nematode Strongyloides ratti induces the expansion of a specialized subset of regulatory immune cells , regulatory T cells ( Treg ) , that counteract effector T cell function . Treg expanded with similar kinetics and suppressed most features of the nematode-specific immune response in two different mouse strains , BALB/c and C57BL/6 . One central factor of this immune response i . e . IL-9-driven rapid degranulation of mast cells , was suppressed by parasite-induced Treg in BALB/c mice non-redundantly . Consequently , Treg depletion elevated IL-9 production , accelerated mast cell degranulation and led to rapid expulsion of S . ratti in BALB/c mice . S . ratti-infected C57BL/6 mice still displayed low IL-9 production and delayed mast cell degranulation in the absence of Treg . Thus S . ratti was able to complete its life cycle in Treg-depleted C57BL/6 mice . This study shows that parasitic worms delay their expulsion by over-activating regulatory elements of their host's immune system such as Treg . The importance of individual regulatory elements during immune evasion depends on their degree of redundancy within the host that is variable in different genetic backgrounds .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"immune",
"cells",
"immunity",
"immunity",
"to",
"infections",
"immunology",
"immunoregulation",
"biology",
"zoology",
"immunomodulation",
"parasitology",
"helminthology"
] |
2014
|
Foxp3+ Regulatory T Cells Delay Expulsion of Intestinal Nematodes by Suppression of IL-9-Driven Mast Cell Activation in BALB/c but Not in C57BL/6 Mice
|
Muscle fatigue is a temporary decline in the force and power capacity of skeletal muscle resulting from muscle activity . Because control of muscle is realized at the level of the motor unit ( MU ) , it seems important to consider the physiological properties of motor units when attempting to understand and predict muscle fatigue . Therefore , we developed a phenomenological model of motor unit fatigue as a tractable means to predict muscle fatigue for a variety of tasks and to illustrate the individual contractile responses of MUs whose collective action determines the trajectory of changes in muscle force capacity during prolonged activity . An existing MU population model was used to simulate MU firing rates and isometric muscle forces and , to that model , we added fatigue-related changes in MU force , contraction time , and firing rate associated with sustained voluntary contractions . The model accurately estimated endurance times for sustained isometric contractions across a wide range of target levels . In addition , simulations were run for situations that have little experimental precedent to demonstrate the potential utility of the model to predict motor unit fatigue for more complicated , real-world applications . Moreover , the model provided insight into the complex orchestration of MU force contributions during fatigue , that would be unattainable with current experimental approaches .
Muscle fatigue is a temporary decline in the force and power capacity of skeletal muscle resulting from muscle activity . Muscle fatigue can adversely affect the lives of workers , athletes , patients , and the elderly—and is a quotidian ( and bothersome ) presence in the lives of most people . Yet , the basic mechanisms underlying muscle fatigue have not been firmly established . In the periphery , muscle fatigue is thought to arise mainly because of impairments in cross-bridge function and excitation-contraction coupling brought about by accumulation of metabolites and alterations in transmembrane ionic concentrations [1–3] . Centrally , muscle fatigue is manifest as an impairment in activation of the motor neurons that drive muscle fibers . There are a number of factors that likely contribute to this impairment , including diminished output of the higher motor centers that operate on motor neurons , increasing synaptic inhibition directed to motor neurons , and intrinsic adaptations in motor neurons that make them progressively less responsive to synaptic excitation during sustained activity [4–7] . Because control of muscle is realized at the level of the motor unit ( a motor neuron and the muscle fibers it innervates ) , it seems important to consider the physiological properties of motor units ( MUs ) when attempting to understand and predict muscle fatigue . Indeed , the few hundred MUs that make up a typical mammalian muscle usually possess wide ranges of contractile properties including force capacities , contractile speeds , and fatigabilities . While convention suggests distinct clustering of MUs ( i . e . MU types ) based on such contractile properties , it is more accurate to represent MU characteristics as residing along broad continua rather than as falling into distinct categories [8] . Control over the diverse population of MUs making up a muscle is enacted in a highly stereotyped way . With few exceptions , MUs appear to be recruited in an orderly sequence , from those that exert the weakest forces toward those that produce the greatest ( see [9] for review ) . Furthermore , there appears to be a tight association between force capacity and fatigability of MUs , such that stronger MUs are more fatigable ( ie . fatigue more rapidly ) than weaker ones [10–12] . In addition , there is a tendency for weak MUs to have slower twitches ( i . e . longer contraction time ) than strong motor units [10] . The neural mechanisms underlying the orderly recruitment of MUs—from weakest , slowest and least fatigable , to strongest , fastest and most fatigable—were largely revealed by Henneman and colleagues and is referred to as the size principle [13–15] . Once recruited , individual MUs increase their firing rate with increased synaptic excitation over a relatively narrow range of values before saturating at levels that appear to be inversely related to the MU’s recruitment threshold [16–18] . As such , during a given contraction , MUs within a muscle can possess a wide range of activities , from those not yet recruited to those that have reached their maximal firing rates . If the contraction is sustained , MUs will fatigue at different paces dictated both by their individual firing rates ( which can vary over time ) and the intrinsic fatigabilities of their innervated muscle fibers . Because of this complexity , it has been difficult to predict the time-course of muscle fatigue , even for relatively simple tasks involving sustained target forces , let alone for tasks in which force levels vary over time and include varying periods of recovery between contractions . Furthermore , when challenged with different tasks , a muscle might eventually accumulate the same level of fatigue ( loss in overall muscle force capacity ) but do so with very different combinations of fatigue within the individual MUs . Therefore , our goal was to develop a phenomenological model of motor unit fatigue , not only as a tractable means to predict the mechanical aspects of muscle fatigue across a wide range of tasks , but also to illustrate the varying responses of the individual MUs whose collective action contributes to the trajectory of changes in muscle force capacity during prolonged activity . As such , this model will provide a framework for better understanding physiological mechanisms contributing to the fatigue of individual muscles , and will have applications in ergonomics , rehabilitation , and exercise . While the present paper simulated MU fatigue associated with sustained , isometric contractions only , this work is the first phase of a more comprehensive model to predict MU fatigue and recovery for any task demand time-history .
Fig 2 shows outputs from the model for a simulated sustained 20% force contraction . There was a progressive force-capacity decline over the course of the trial—necessitating an increase in excitation ( green trace , Fig 2A ) from the initial value of 27 . 9% maximum voluntary excitation ( MVE ) , to 100% MVE at the endurance time of 511 . 5 s . The increase in excitation was realized as a gradual increase in firing rates ( Fig 2B ) up to the assigned maximum rates of those MUs activated from the outset of the contraction ( MUs 1–90 ) and by recruitment and subsequent increase in firing rate of the highest threshold MUs ( MUs 91–120 ) . In experimental studies , that have used protocols similar to what was simulated in Fig 2 ( i . e . ~20% target force ) , MU firing rates also tended to increase with time [20–22] , not unlike the results of our simulations ( Fig 2B ) . Only a few of the highest threshold MUs , simulated in Fig 2 , did not attain their assigned maximal firing rates at the endurance limit because of the countervailing effects of firing rate adaptation , which was assigned to have a more potent effect on high threshold compared to lower threshold MUs . In the absence of firing rate adaptation ( gray line , Fig 2A ) , muscle force at 511 . 5 s would have been about 85% higher than that produced in the presence of firing rate adaptation ( black line , Fig 2A ) and the endurance time would have been extended by about 40 s . Low threshold MUs ( e . g . MUs 1–20 ) initiated their activities close to their maximum firing rates ( Fig 2B ) , which were also close to the normalized firing rates needed for these slow twitch MUs to attain their maximal force ( Fig 1C ) . As such , and because these MUs were fatigue resistant ( i . e . assigned low fatigability values , Fig 1D ) , their force contributions ( Fig 2C ) remained relatively stable throughout the trial . These low threshold MUs were also the weakest ( see Fig 1B ) and , as such , their contribution to overall muscle force was modest . Somewhat higher threshold MUs ( e . g . MU 40–60 ) also initiated their firing at relatively high rates ( Fig 2B ) but these units had slightly higher intrinsic fatigability ( Fig 1D ) and , consequently , their force decreased gradually throughout the trial . The highest threshold MUs , recruited from the outset of the trial ( e . g . MU80 ) , initially had relatively low firing rates ( Fig 2B ) . In addition , because these units had comparatively brief initial contraction times , their initial normalized firing rates were quite low . For example , the initial firing rate of MU80 was about 13 imp/s ( Fig 2B ) and it had an initial contraction time of about 43 ms ( 0 . 043 s ) ( Fig 1B ) . The product of these two values yields a normalized firing rate ( Eq 4 , Methods ) of about 0 . 58 , which placed it quite low on the force-frequency curve ( Fig 1C ) . As firing rate increased , these MUs moved up the steep portion of the force-frequency curve , leading to an initial increase in their force ( Fig 2C ) which partially compensated for the decreasing force from lower threshold MUs to maintain muscle force at the target level of 20% of maximum ( blue trace , Fig 2A ) . However , because these higher threshold units also had reasonably high fatigability ( Fig 1D ) , their force output eventually started to decline ( e . g . at ~ 250 s for MU80 ) and then declined steeply for the remainder of the trial ( Fig 2C ) . Motor units recruited later in the trial started firing at the minimum rate and then gradually increased their firing rates as excitation increased ( e . g . MU100 , Fig 2B ) . The low starting firing rates , combined with the brief contraction times of high threshold MUs , placed these units initially on the far left , linear portion of the force-frequency curve ( Fig 1C ) . Consequently , as excitation increased , firing rates increased in these units , and force initially increased linearly ( Fig 2C , MU 100 ) then eventually transitioned into the steeper portion of the force-frequency curve and , consequently , force then increased more precipitously ( from ~ 340–440 s , Fig 2C ) . Because these MUs were assigned to have high force capacity ( Fig 1B ) , their force contribution was substantial . These MUs , however , were also the most fatigable and , as such , their force output then decreased steeply . In addition , because of the imposed ‘onion-skin’ organization ( i . e . high threshold MUs assigned the lowest maximum firing rates , see Methods ) and the effects of firing rate adaptation , these high threshold MUs increased firing rate over only a relatively modest range . Eventually , the endurance time was reached when no further voluntary increase in any MU's force was possible . Fig 2D shows the force capacity of each MU relative to its initial force . At the endurance limit ( ~512 s ) , MU1 had lost only about 5% of its force whereas MUs 20 , 40 , and 60 had lost ~ 15% , 35% , and 80% of their force , respectively . Interestingly , MUs 66–98 had lost all their force capacity and were essentially exhausted . Fig 3 shows the simulation of a sustained contraction at 50% of maximum force . Excitation ( green trace , Fig 3A ) increased over the trial , at a rate substantially greater than for the 20% force trial , to maintain the 50% target force in the face of the progressively declining total muscle force capacity ( black trace , Fig 3A ) . At 95 . 5 s , muscle force capacity dropped below the target level of 50% , thereby demarking the endurance limit for this trial . If firing rate adaptation was not included in the simulation ( gray trace , Fig 3A ) , muscle force capacity at 95 . 5 s was well above the target force and the endurance time was extended to ~132 s . MUs 1–109 were recruited from the outset of the contraction ( Fig 3B ) . Of these , MUs 1–72 initiated their activities already at their assigned maximum firing rates . As a result , the firing rates of these MUs declined over time due to the influence of firing rate adaptation . Higher threshold MUs in this group exhibited greater degrees of firing rate adaptation than lower threshold MUs ( e . g . compare MU 60 to MU40 , Fig 3B ) . MUs recruited at from the start of the contraction , but with firing rates less than their assigned maximum ( e . g . MU 80 , Fig 3B ) , initially increased firing rates in response to the escalating excitation . However , the rate of firing rate increase was less than the rate of excitation increase because of the competing effect of firing rate adaptation . MUs with high initial firing rates ( e . g . MU 80 ) eventually reached their maximal rates , after which time their firing rates declined due to adaptation . MUs activated from the outset , but with lower initial rates ( e . g . MU 100 ) , gradually increased their firing rates but did not reach their maximal firing rates . This failure to reach maximal firing rates occurred because , over time , the increasing effects of adaptation undercut the effects of increases in excitation . In some cases ( e . g . MU 100 ) , a near balance was struck between these two competing influences leading to a leveling-off in firing rate . As a consequence of the complex interaction between excitation ( tending to drive the pool of MUs as a collective ) and adaptation ( an intrinsic effect that influences the firing rates of individual MUs ) , an array of firing rate profiles was observed . Indeed , some MUs showed progressive decreases in firing rate ( low threshold MUs ) , some showed increases followed by decreases in firing rate , and others showed mainly progressive increases in firing rate . Furthermore , at any point in time , a range of firing rate responses could be observed . For example , at about mid-way through the contraction ( ~50 s ) , some MUs had stable firing rates , some had slowly decreasing firing rates , others had increasing firing rates , and some units were just being recruited . Such disparate firing rate responses across motor units have also been observed in human motor units during fatiguing contractions ( e . g . [20 , 21 , 23–29] ) . Like that for the 20% force contraction , lower threshold MUs ( i . e . MU 1–60 ) showed little drop in force over the duration of the contraction held at 50% force ( Fig 3C ) . These units were assigned low values of fatigability ( Fig 1D ) and , also , exhibited little firing rate adaptation . As such , their force contributions ( although comparatively small ) were relatively stable during this simulation . Higher threshold MUs , that were recruited from the outset and with high initial firing rates ( e . g . MU 80 , Fig 3C ) , showed a progressive decline in force capacity over the course of the contraction due to both relatively high values of assigned fatigabilities and greater degrees of adaptation , compared to lower threshold MUs . MUs recruited from the outset at low rates ( e . g . MU 100 ) , and those units recruited after the onset of the contraction , initially increased their force contribution due to increasing firing rates . Eventually , however , when diminishing intrinsic excitability ( associated with firing rate adaptation ) matched or exceeded the degree of increasing extrinsic excitation , firing rates leveled off or started to decrease ( see Fig 3B ) . Consequently , force then dropped steeply for these high threshold units that were assigned the highest fatigability values ( Fig 1D ) . At the endurance limit , lower threshold MUs ( e . g . MUs 1–60 ) still retained at least 90% of their force capacity ( Fig 3D ) . The MUs that underwent the largest relative drop in force during this contraction were MUs 90–100 with only 40–45% of their force capacity remaining at the endurance limit . For comparison with the 20 and 50% contractions , Fig 4 shows simulation results for a sustained 80% force contraction . From the outset of the contraction , total muscle force capacity decreased steadily , which was counteracted by a progressive increase in excitation to maintain muscle force at the target level ( Fig 4A ) . However , after only about 15 s of activity , maximal excitation was reached and , as such , the declining total muscle force capacity could no longer be counteracted and the endurance limit was reached . This endurance time ( 14 . 8 s ) was only about 3% of that associated with the 20% force contraction and 15% of the 50% force contraction . All MUs were recruited at the start of the trial ( Fig 3B ) . Most motor units ( MUs 1–103 ) initiated their activities at their maximal firing rates . Consequently , their firing rates declined progressively over the trial because of the effects of firing rate adaptation , the degree of which varied as a function of MU threshold . In an experimental study , that used a similar target force as used in this simulation , MU firing rates also tended to progressively decrease with time [30] . Because the highest threshold MUs ( MUs 104–120 ) in the simulation were initially activated below their maximal firing rates , as excitation increased , their firing rates first increased toward their maximal rates followed by a gradual decline in firing rate due to adaptation . Due to the combined effects of firing rate adaptation and peripheral fatigue , MUs activated from the outset at their maximal rates showed a progressive decline in force ( Fig 4C ) with the greatest losses occurring in the strongest ( and most fatigable ) MUs . Motor units that were activated initially at rates below their maximum ( MUs 104–120 ) exhibited an initial increase in force as their firing rates increased , followed by a decline in force as firing rates adapted and the process of peripheral fatigue continued . At the endurance time , the degree of force capacity loss ( relative to the initial forces ) was relatively small for all MUs ( Fig 4D ) . For example , the MU exhibiting the greatest fatigue ( MU107 ) still retained ~86% of its force capacity at the endurance time . This contrasted with the 20% force contraction ( Fig 2D ) in which 28% of the MUs were completely exhausted . Nevertheless , for the 20% , 50% , and 80% force contractions , the simulations indicated a complex interplay of force contributions among the MU population ( Figs 2C , 3C and 4C ) with individual forces increasing and decreasing , at varying times and with different rates , but with the total muscle output maintaining the target force up until the endurance limit . Fig 5 shows simulations associated with a sustained maximum voluntary effort ( 100% MVE ) . The model predicted an immediate decrease in total muscle force capacity with an endurance time less than 1 s ( Fig 5A ) . To mimic what has been done experimentally for such contractions , we continued the simulation out to a time of 200 s . Total muscle force declined relatively steeply over the first ~ 40 s of the contraction , then somewhat less steeply up to about 120 s , and finally with a more gradual decline in force over the last ~ 80 s of the contraction . Because voluntary excitation was maximum throughout , muscle force and force capacity were the same ( Fig 5A ) . At the end of the 200 s contraction , force was down to ~15% of the initial force . Because excitation was maintained at 100% throughout the contraction , changes in firing rate ( Fig 5B ) were caused entirely by firing rate adaptation . Fig 5B also nicely illustrates the differential effects of adaptation across the MU population , with low threshold MUs showing little adaptation and high threshold units exhibiting marked adaptation . As expected , the inclusion of adaptation led to a greater decrease in muscle force ( black trace , Fig 5A ) as compared to simulations without adaptation ( gray trace , Fig 5A ) , especially in the first 160 s , after which time there was little difference . This was a consequence of a complex interaction between normalized firing rate , normalized force ( Fig 1C ) , instantaneous fatigability ( Eq 10 , Methods ) , and fatigue-related changes in contraction time ( Eq 11 , Methods ) . Since all MUs were recruited at the start of the trial , the effects of firing rate adaptation dominated in the first 35 s . However , because MU contraction times increased with MU fatigue , this tended to shift MUs higher on their force-frequency curves , thereby partially offsetting the force loss associated with declining firing rates due to adaptation . Consequently , the difference in the degree of force decline between simulations , that included and excluded adaptation , tended to dissipate in the latter third of the trial . Fig 5C shows the force contribution of the individual MUs over the course of the 100% force trial . It is important to note that , at the outset of the trial , before any fatigue had occurred , the forces produced by the highest threshold MUs were less than their theoretical maximum forces . For example , MU120 had a capacity to generate 100 times more force than MU1 , yet its initial force at 100% MVE was only 57 times greater than MU1 . This was due to: ( a ) the imposed ‘onion skin’ organization that limits the maximum firing rates of high threshold MUs to be less than that of low threshold MUs , and ( b ) the briefer contraction times of the high threshold MUs which decreased their normalized firing rates and led to lower forces . This implies that there is a reserve capacity of force ( mostly vested in the highest threshold motor units ) that normally is not accessed even during maximal voluntary efforts . There is substantial circumstantial evidence that lends support to this idea [31–34] . The initial steep drop in muscle force ( Fig 5A ) was primarily due to the rapid loss of force occurring in the highest threshold , strongest MUs ( Fig 5C ) . Those MUs lost force quickly because of a combination of greater firing rate adaptation and greater fatigability . After about 60 s , when firing rate adaptation was largely complete for all MUs , MU forces declined relatively steadily although with different slopes for different MUs related to their individual force capacities . An interesting exception existed with the highest threshold MUs . For example , after 60 s , the slope of force capacity decrease was less for MU120 than for MU100 ( Fig 5C ) . This was primarily due to the initial greater extent of adaptation for MU120 than MU100 , causing a larger decrease in its firing rate , which combined with its short contraction time to substantially shift it to the left on its force-frequency curve . This , in turn , led to a marked and early reduction in force output of MU120 such that it produced substantially less force than MU100 at 60 s . Because MU fatigability was partially dependent on normalized force ( Eq 10 , Methods ) , the rate of force decline was less for MU120 than MU100 for much of the contraction . The rate of total muscle fatigue decreased after ~120 s because many high threshold MUs became exhausted and could no longer further lose force capacity . Indeed , after 200 s of sustained maximum voluntary excitation , many high threshold motor units ( MU86-119 ) had lost virtually all their force generating capacity ( Fig 5D ) and MU120 generated only ~10% of its initial force . Because these high threshold MUs were also initially substantially stronger than the lower threshold units , such large losses in their force were associated with the large overall drop in total muscle force capacity during this trial . Endurance times were determined for a set of simulations ( like those shown in Figs 2 , 3 and 4 ) for target force levels at 15% of maximum , and from 20–100% maximum in 10% increments . The resulting relation between predicted endurance time and target force is shown in Fig 6 ( solid black line ) . The weighted average values of endurance times , determined experimentally for six different joints during submaximal contractions [35] , and from three joints during maximum contractions [36–39] are shown in Fig 6 . Overall , there was a good correspondence between predicted and actual endurance times across a wide range of forces , joints , and studies . Across the nine efforts from 15% to 90% MVC force , the average and RMS differences between the model-predicted and empirical endurance times were -33 . 7 s and 52 . 3 s , respectively . These amounted to -3 . 9% and 6 . 0% of the full range of empirical endurance times ( 884 s at 15% MVC ) . The largest absolute difference in endurance times , between simulated and empirical values , was -112 . 8 s for the 15% MVC effort , representing12 . 8% of the empirical mean of 884 s . For the higher forces of 70 , 80 and 90% MVC , there were larger relative differences between the model and empirical means ( see inset , Fig 6 ) . However , the absolute magnitude of these differences never exceeded 24 s . Fig 7 shows the simulated force and experimentally measured forces during sustained maximal contractions . In general , there was a reasonably good match between the simulated and experimental results . In the first 20 s , however , simulated force dropped somewhat more steeply ( 1 . 4% MVC/s ) than that recorded experimentally ( average of 1 . 0% MVC across the four experiments ) . The time at which force had dropped to 50% of maximum was ~70 s for the simulation . That time was quite similar to the ~61 s value averaged from the four experimental studies that had contractions of sufficient duration to cause at least a 50% decline [36–38 , 40] . The simulated loss of force beyond 20 s , and until 200 s , paralleled quite closely that of the one experimental study [38] that monitored sustained maximal contractions for 200 s . Nevertheless , the force at 200 s in that study was about 25% of the initial force , whereas the simulated force at that time was ~16% of the initial force . Given the reasonable correspondence between simulated force and experimental findings , we were encouraged to carry out further simulations involving somewhat unconventional tasks to highlight the potential of the model to predict fatigue under more complex circumstances . Fig 8 shows a simulation involving a ‘staircase’ task in which the force was maintained for 32 s at progressively increasing 20% MVC plateaus with a brief ramp increase in force between plateaus . The endurance time for this task was 101 . 5 s and occurred during the third plateau when the 60% MVC target could no longer be maintained ( Fig 8A ) . The first plateau ( 20% MVC ) was maintained with very little change in force capacities of the active MUs ( Fig 8C ) requiring only a subtle increase in excitation . As such , the firing rates of the MUs active during the first plateau ( MUs 1–90 ) changed little ( Fig 8B ) and only one new unit ( MU91 ) was recruited ( at ~15 s into the trial ) . The increase in excitation necessary to attain the second plateau ( 40% MVC ) was accompanied by ~13 imp/s increase in firing rate in MUs that were active during the first plateau ( Fig 8B ) , plus recruitment of an additional 15 MUs ( MUs 92–106 ) . However , this increase in firing rate had little effect on force generated by the lowest threshold MUs ( Fig 8C ) because their firing rates were already high enough to place them on the plateau of the force-frequency curve ( Fig 1C ) . For example , during the transition from the 20% to 40% plateau , MU40 started with a contraction time of ~63 ms and increased its firing rate from 23 . 7 to 33 . 6 imp/s . At 23 . 7 imp/s , the normalized firing rate is 23 . 7 imp/s x 0 . 063 s = 1 . 49 , which is associated with a force output of ~100% of maximum for that MU ( Fig 1C ) . As such , increasing the firing rate to 33 . 6 imp/s had a negligible effect on force for MU40 . However , the increased firing rate of higher threshold ( and faster contracting ) MUs ( e . g . MU80 ) , did translate into marked increases in force . Those increases , combined with the recruitment of higher threshold ( and stronger ) MUs , enabled the 40% target to be attained . Those units contributing the greatest amount of force during the 40% plateau were also relatively more fatigable ( e . g . MU 80 ) . As their force started to decline during the sustained 40% plateau ( Fig 8C ) , excitation progressively increased ( Fig 8A ) . The increased excitation caused firing rates to increase in those MUs ( 60–106 ) that had not yet reached their maximal firing rates ( Fig 8B ) . It is interesting to note that the slope of the firing rate increase varied systematically across these MUs during this time ( Fig 8B ) . The lower threshold MUs ( e . g . MU 80 ) had steeper slopes than higher threshold MUs ( e . g . MU100 ) . This was a consequence of firing rate adaptation being greater for higher threshold MUs , compared to lower threshold MUs , which more potently attenuated the increases in firing rate in these MUs during increased excitation . The increased excitation during the 40% plateau also led to the recruitment of three additional MUs ( Fig 8B ) . The increase in excitation needed to achieve the 60% MVC target resulted in increases in firing rates in those MUs that had not yet saturated ( MUs 72–109 ) and the recruitment of all the remaining MUs but MU120 ( Fig 8B ) . Force then dropped off relatively steeply in some high threshold MUs ( e . g . MU100 , Fig 8C ) because these MUs were assigned high fatigability values ( Eq 9 ) and they were no longer capable of increasing firing rate . This force loss was partially compensated for by increasing firing rates of the most recently recruited MUs and the recruitment of the last unit ( MU120 , Fig 8B ) . The increases in firing rate in these high threshold , strong and highly fatigable MUs led to an initial and brief escalation in their force ( Fig 8C ) followed by a steep decline , such that the muscle's maximum force capacity eventually fell below the target force ( at ~102 s ) . The MUs most impaired by this task , in terms of loss in force capacity at the end of the trial , were MUs 66–101 ( Fig 8D ) . There were no exhausted MUs and the MU exhibiting the greatest fatigue ( MU93 ) still retained ~65% of its force capacity at the endurance limit . A second set of somewhat unconventional tasks involved using target forces of 15% , 50% and 85% MVC but , in each case , the simulation was continued until muscle force had decreased below15% of maximum . As such , each case was associated with the same degree of total muscle fatigue ( ie . 85% , as conventionally defined ) but brought about by different ‘paths’ . We were interested to know whether these different routes to the same level of total muscle fatigue would have differential effects on the MU population . As shown in Fig 9A , force was maintained at the 85% MVC target ( blue trace ) for only about 10 s before declining and eventually dropping to 15% MVC at a time of 206 . 5 s . Likewise , force was maintained at the 50% target ( green trace ) for about 95 s before decaying to 15% MVC at a time of 234 . 5 s . For the 15% MVC target ( red trace ) , force was maintained at that level for 774 . 0 s . Fig 9B shows the remaining relative force capacity of each MU , at the time force dropped below the 15% MVC target , for each of the three cases . Despite the same 85% decrease in muscle force capacity , the profiles of fatigue across the MU population were strikingly different for the three cases . The 15% target trial ( red symbols ) resulted in substantially greater fatigue in the lower threshold MUs compared to the other two cases , but with less fatigue in the highest threshold MUs . On the other hand , for the 85% target trial ( blue symbols ) , the degree of fatigue was greater ( i . e . lower force capacities ) for the high threshold MUs as compared to the other two cases , but with less fatigue in the lower threshold MUs . In addition , the sets of MUs that became completely exhausted differed for the three different cases: MUs 56–102 , MUs 83–113 , and MUs 86–118 for the 15% , 50% , and 85% target force trials , respectively .
Many models have been published that predict the mechanical aspects of muscle fatigue [41–43] . For example , the three-compartment model of muscle fatigue developed by Liu et al . [43] has been used effectively by several investigators to predict fatigue for a variety of tasks ( e . g . [44–47] ) . However , that approach essentially simplifies muscle physiology to one type of MU and assumes MUs are fully rested , fully activated , or completely fatigued . Other models have been developed that predict the responses of groups of MUs ( e . g . [48 , 49] ) or individual MUs [50–52] ) . For example , the Dideriksen model [50] was a pioneering effort that used changes in metabolite concentrations within muscle as a key factor driving alterations in MU contractility and neural drive during fatigue . Such a mechanistic model was necessarily complex and , as such , the output of the model generally presented only the net effect of prolonged activity on total muscle capacity . Because our model was less complex , while still accounting for individualized responses for an entire MU population , it could make accurate predictions about total muscle fatigue ( Fig 6 ) and readily display the interplay of force contributions among the constituent MUs during a wide variety of fatiguing tasks ( Figs 2 , 3 , 4 , 5 and 8 ) . There were also some differences in the physiological representations of the Dideriksen et al . [50] model as compared to the present model . The Dideriksen model employed a ‘cross-over’ scheme to predict MU firing rates , wherein higher threshold MUs ultimately discharge at higher rates than low threshold MUs . While there are some data to support this type of organization [53 , 54] , many findings suggest a nested ‘onion-skin’ organization ( as used here ) in MU firing rate profiles [16–18 , 23 , 55 , 56] . Another difference between the two models is in the degree of fatigability of different MUs . In the Dideriksen model , none of the MUs would have been classified as ‘fatigable’ according to conventional criteria of Burke et al . [10] ( i . e . fatigue index values < 0 . 25 ) . In our model , the highest threshold and most fatigable MUs had fatigue-index values as low as 0 . 1 , consistent with the original data of Burke et al . [10] . However , questions remain as to how well data obtained from cat hindlimb MUs ( e . g . [10] ) might generally represent the properties of human MUs [9 , 57 , 58] . A third distinction between the two models relates to the implementation of adaptation of firing rates during sustained activity . The Dideriksen model did not account for intrinsic changes in excitability of motor neurons associated with spike frequency adaptation . Such adaptation , however , is a well-established property of motor neurons [59–69] . Furthermore , such intrinsic changes in excitability may partially explain the observed differences in firing responses across MUs ( i . e . some with decreasing firing rates while others are recruited and increasing their firing rates ) during fatiguing contractions that would be otherwise difficult to account for with broadly distributed sources of synaptic input [25 , 29 , 70] . The ability of the model to reveal the intricate interplay of MU contributions during fatigue provided interesting predictions and insights . For example , the model predicted that low-force contractions , sustained to their endurance limit , induce more fatigue across the MU population than high-force contractions ( compare Figs 2D and 4D ) . This prediction has implications for rehabilitation medicine , as it suggests that relatively weak contractions could provide a potent exercise stimulus for much of the MU population without the risks associated with intense , high-force contractions . In addition , the model predicted that loss in force capacity was always more pronounced among the upper-middle range of motor units ( from ~ MU60 –MU 110 ) while both the lowest threshold and highest threshold MUs subpopulations were less impaired across a wide range of tasks ( Figs 2D , 3D , 4D , 5D , 8D and 9B ) . Relative to low threshold MUs , the greater fatigue in the upper-middle range of MUs occurred simply because they were intrinsically more fatigable than the low threshold MUs , while being active for practically the same durations . On the other hand , greater fatigue in the upper-middle range of MUs , compared to the highest threshold MUs , was due to their more prolonged involvement in many of the tasks ( e . g . Figs 2B , 3B and 8B ) and because they tended to sustain higher levels of absolute force ( provoking greater fatigue ) than the highest threshold MUs ( e . g . Figs 4C and 5C ) . The highest threshold MUs , despite their intrinsic capability to generate the largest forces , never achieved their full force capacities because of the limits placed on their maximum firing rates . The physiological mechanisms underlying such firing rate saturation are not yet known despite several recent investigations into this phenomenon [71–75] . In addition , firing rates decreased more precipitously for the highest threshold MUs than other MUs ( e . g . Fig 4B ) due to greater firing rate adaptations . Such reductions in firing rate led to lower forces and , thereby , lessened their fatigue compared to the upper-middle threshold group of MUs . Another prediction involved simulations of sustained 15% , 50% and 85% MVC force . The model predicted marked differences in MU fatigue when the total force capacity decreased to the same level of 15% MVC ( i . e . 85% fatigue ) ( Fig 8 ) . Although all three cases led to large subsets of MUs that were exhausted , the specific MUs in each subset differed depending on the initial target force . Furthermore , the degree of fatigue , in those units still capable of producing force at the endurance limit , varied substantially across the three conditions . Therefore , despite an equivalence in the degree of muscle fatigue , based on the prevailing definition of fatigue ( a reduction in muscle force/power capacity ) , the physiological status of the motor unit population was quite different under the three conditions . Such differences could have important implications , for example , in determining the subpopulations of MUs receiving the greatest exercise stimulus in the context of strength or endurance training and for how the muscle responds to ensuing demands and recovery in the context of physical work occurring within an industrial setting . The recent development of high-density multi-electrode arrays [76–79] , combined with sophisticated decomposition algorithms [55 , 80 , 81] , enables tracking of a large numbers of MUs during a wide range of contractions ( e . g . [56 , 82 , 83] ) . Such technology should make it possible to evaluate some of the predictions made here , particularly with regards to patterns of MU activity . Unfortunately , however , few methods are presently available that can readily measure changes in force capacity of MUs during fatiguing contractions . One limitation of the present model is that it did not account for differences in the constellation of MU properties making up different muscles and/or occurring in different individuals . Instead , we opted here for a ‘one size fits all’ approach that , nevertheless , did make good predictions of endurance times for a wide range of muscles ( Fig 5 ) . However , because the model is flexible and all parameters are readily altered , one could easily carry out simulations of fatigue associated with different types of situations , such as might occur with muscles having different fiber type compositions , particular neuromuscular diseases , or with aging . Another limitation of the present model is that it simulates isometric force only . This is a critical limitation , as most behaviors involve dynamic muscle activity . This is an especially challenging limitation to overcome because there are so little experimental data involving lengthening and shortening contractions in individual MUs . In this regard , perhaps mechanistic models of fatigue ( e . g . [50 , 52] ) combined with Hill-type models of contractile dynamics MUs [84] could , from first principles , make good predictions about fatigue arising during tasks involving movement . A further limitation of the present model is that we used only one of a number of neural mechanisms that can contribute to central fatigue ( see [6] ) . For simplicity , diminished intrinsic excitability ( associated with firing-rate adaptation ) served as a representative mechanism underlying central fatigue . As such , fatigue-related alterations in descending drive ( e . g . impaired motor cortical output ) and sensory feedback ( e . g . increased inhibition associated with activation of metabolite-sensitive receptors in muscle ) were not explicitly simulated in the model . Nevertheless , there is a significant body of experimental work that has concluded , for example , that feedback from metabolite-sensitive receptors does not appear to significantly inhibit motor neurons during fatigue [85–88] . In addition , reduced excitability of motor neurons during fatigue does not appear to be due to diminished peripheral excitatory input [89] . Furthermore , some evidence indicates that the motor cortex is relatively unimpaired during voluntary fatiguing contractions [90] . On the other hand , there is compelling data indicating diminished intrinsic excitability of motor neurons during fatigue [91] . As such , it seemed reasonable to use reduced intrinsic excitability as a proxy for central fatigue in the present model . However , there are sure to be differences in the relative contributions of various central fatigue mechanisms that depend on the muscle group involved ( e . g . [92] ) or task [70 , 93] . And finally , an additional limitation with the present version of the model is that does not include recovery from fatigue . In real world situations , muscle fatigue usually does not occur in isolation—it is influenced by previous bouts of muscle activity and the degree of intervening rest . This issue is particularly relevant to physical ergonomics , a field that has produced many analysis tools to determine the acceptability of an isolated task , but almost no methods are available to estimate muscle fatigue and injury risk associated with the typical case of workers performing a combination of different subtasks , including brief periods of rest , as part of their whole job . In its next version , our fatigue model will be expanded to include both the fatigue and recovery of MUs , so that the effects of combined efforts can be assessed . This addition could assist in determining the acceptability of whole jobs and/or for optimizing task allocation and sequencing . The model could also then be used to design exercise and rehabilitation programs that set demand magnitudes and work/rest ratios to optimize the exercise stimulus , given the particular physiological state of the motor unit population .
The model presented here represents only one of many plausible schemes to simulate MU activity and force . Consequently , parameter selections were meant to be generally representative , but not definitive , characterizations of any specific skeletal muscle . The model enables users to readily specify parameters needed to simulate a variety of MU organizations . For the present study , the simulated muscle consisted of a pool of 120 MUs . MU twitches were modeled as the impulse response of a critically damped 2nd order system ( Fig 1A ) . Each MU , i , was assigned a unique twitch amplitude and twitch contraction time . The distribution of MUs based on twitch amplitude , P , was determined using the exponential function [19]: P ( i ) = e[ ln ( RP ) ( i − 1 ) / ( n−1 ) ] ( 1 ) where ln is the natural logarithm , RP is the desired range of twitch forces across the pool , and n is the number of MUs in the pool ( i . e . , 120 ) . For these simulations , RP was assigned a value of 100 . Such a representation yields a distribution with many weak MUs and relatively few strong MUs . Maximum MU forces were normalized to the force of MU ( 1 ) such that the force of MU ( 1 ) was 1 . 0 and MU ( 120 ) was 100 . 0 force units . Contraction times were assigned as an inverse function of twitch amplitude ( see [19] ) for specific formulation ) and ranged from 30 ms for the strongest unit , MU ( 120 ) , to 90 ms for the weakest , MU ( 1 ) ( Fig 1B ) . All MUs within the pool received the same level of excitatory drive ( E ) that could vary as a function of time ( t ) . The amount of excitatory drive needed to recruit each MU , referred to as 'recruitment threshold excitation' ( RTE ( i ) ) , was also determined as an exponential function that assigned many MUs to have low thresholds and few to have high thresholds , using: RTE ( i ) = e[ ln ( RR ) ( i − 1 ) / ( n−1 ) ] ( 2 ) where RR is the desired range of recruitment thresholds and was set to 50 for the present simulations . A MU , therefore , was recruited when the excitatory drive equaled or exceeded its assigned recruitment threshold excitation ( RTE ( i ) ) . Therefore , and in general accordance with the size principle , weaker MUs ( i . e . those with low twitch forces ) were recruited at lower levels of excitation than stronger MUs . At threshold excitation , MUs discharged at a minimum firing rate ( minR ) of 8 impulses ( imp ) /s . Firing rate ( R ) increased linearly with increased excitation up to an assigned maximum rate ( maxR ( i ) ) for each MU , beyond which no further increases in rate occurred ( i . e . firing rate saturated ) . The slope ( i . e . the gain , g ) of increased firing rate with excitation was set to be the same for all MUs ( 1 imp/s for each unit increase in excitation ) and firing rate was modeled as: R ( i , t ) = g[E ( t ) − RTE ( i ) ]+ minR ( 3 ) Based on considerable experimental findings , maximum firing rates ( maxR ( i ) ) were modeled as an inverse function of recruitment thresholds , yielding a nested or ‘onion skin’ arrangement of firing rates across the MU population ( e . g . , [18 , 56] ) . In the present model , maxR was assigned to be 35 imp/s for MU ( 1 ) and decreased uniformly to 25 imp/s for MU ( 120 ) . For excitation levels above that needed to bring a MU to its assigned maximum rate , firing rate was maintained at maxR ( i ) . Maximum excitation ( Emax ) to the MU pool was defined as the amount of excitation needed to bring the highest threshold MU to its assigned maxR ( i ) . Rearranging Eq 3 to solve for the excitation ( E ) associated with this situation yields: Emax = RTE ( 120 ) + ( maxR ( 120 ) —minR ) /g = 50 + ( 25–8 ) /1 . 0 = 67 excitation units , such that the last MU is recruited at 50/67 = 74 . 6% of Emax . The relation between MU ( or whole muscle ) force and firing frequency generally exhibits a sigmoid form [96 , 97] . The specific shape of the force-frequency relation depends on contractile speed , in that MUs with long duration contraction times attain tetanic fusion ( i . e . plateau on the sigmoid ) at lower rates than do MUs with brief contraction times [98–100] . If , however , the MU stimulus frequency or firing rate ( R ( i , t ) ) is normalized to the inverse of the twitch contraction time ( 1/CT ( i ) ) , force-frequency curves are similar for most MUs [19 , 99] . The normalized firing ( or stimulus ) rates ( NR ) can be represented as: NR ( i , t ) = R ( i , t ) / [1/CT ( i ) ] = R ( i , t ) x CT ( i ) ( 4 ) A composite linear and sigmoidal relationship ( see Fig 1C ) was used to estimate normalized force ( NF ) as a function of normalized firing rate ( NR ) , as originally derived by Fuglevand et al [19] but simplified to: for NR ( i , t ) ≤ 0 . 4 , NF ( i , t ) = 0 . 3 NR ( i , t ) ( 5 ) for NR ( i , t ) > 0 . 4 , NF ( i , t ) = 1−e−2 NR ( i , t ) 3 ( 6 ) The instantaneous force ( F ) of each MU was then scaled as a function of its assigned twitch force , P ( i ) : F ( i , t ) = NF ( i , t ) P ( i ) ( 7 ) The total muscle force was then calculated simply as the linear sum of all 120 individual MU forces at any given time . Central fatigue encompasses a host of mechanisms that can curtail the spiking output of motor neurons [4–7] . One category of such mechanisms is that related to diminution of net excitatory drive to motor neurons . This can occur due to decreases in excitatory input ( e . g . , from supraspinal centers ) and/or increase in inhibitory inputs ( e . g . via peripheral receptors and their spinal interneurons ) . In the model , such a reduction in net excitation could be implemented by decreasing the excitatory drive function , E ( t ) , or by reducing the maximum excitation , Emax ( Eq 4 ) . For simplicity in the present simulations , however , neither E ( t ) or Emax were explicitly reduced to simulate this category of central fatigue mechanisms . Another category of mechanisms underlying central fatigue are those intrinsic to motor neurons that contribute to time-dependent decreases in motor neuron firing in the presence of constant excitatory drive , referred to as firing-rate ( or spike-frequency ) adaptation . Here we simulated firing-rate adaptation using an approach similar to that described previously in detail by Revill and Fuglevand [117] . In brief , an exponentially rising outward ( i . e . inhibitory ) “current” was subtracted from the excitatory drive function to yield the net excitation acting at the spike initiation zone of a motor neuron . The extent of this intrinsic adaptation current , A , for any MU ( i ) , was a function of both the time since MU recruitment , TR ( i ) , and the excitation level , E ( t ) , namely , A ( t , E ) = q ( i ) [1−e− ( t−TR ( i ) ) /τ] ( 12 ) where τ is the time constant . We assigned the time constant a value of 22 s based on experimental observations of Sawczuk et al . [63] and Gorman et al . [60] . The parameter q ( i ) in Eq 12 designates the maximum value of the adaptation ( inhibitory ) current for each MU . Because the magnitude of adaptation tends to be larger with greater levels of depolarizing current and firing rate [59 , 60 , 63] , q was modeled to depend on a MU's firing rate in excess of its minimum firing rate [i . e . , R ( i , t ) —minR] . In addition , the magnitude of firing-rate adaptation appears to be more pronounced in high threshold compared to low threshold MUs [59 , 64] . Therefore , we also included recruitment threshold as an additional factor influencing the maximum extent of adaptation , q for each MU , using: q ( i ) =ϕ [R ( i , t ) − minR+ d] [RTE ( i ) − 1RTE ( n ) − 1] ( 13 ) where ( RTE ( i ) —1 ) / ( RTE ( n ) —1 ) indicates the recruitment threshold excitation of any MU ( i ) , relative to the largest threshold MU ( n ) , or RTE ( 120 ) in our model . The parameter ϕ was selected to match the magnitude of adaptation for different levels of excitation , as reported by Kernell and Monster [59] , and was assigned a value of 0 . 67 . The parameter d was included to account for the observation that the absolute minimum firing rate that a MU can sustain is lower at derecruitment than at recruitment [18 , 54 , 118] . Such a lower firing rate at derecruitment may be partially due to adaptation . Therefore , firing rate could decay with time below the initially specified minimum firing rate by a small amount determined by d . In the present simulations , d was assigned a value of 2 imp/s , similar to values reported experimentally [54 , 118] . As an example of how firing rate adaptation was implemented , consider one MU , say MU ( 60 ) , in the population of n = 120 MUs . From Eq 2 , the RTE for MU ( 60 ) = 6 . 96 excitation units . Under a constant excitatory drive of E = 20 excitation units , and in the absence of adaptation , Eq 3 would predict a steady firing rate of 21 . 04 imps/s . With adaptation , the adaptation current , A ( Eq 12 ) progressively undercuts the net excitatory drive acting on the MU and decreases the firing rates . At 20 units of excitation , the maximum extent of firing rate adaptation ( Eq 13 ) for MU60 would be q ( 60 ) = 0 . 67 [21 . 04–8 + 2] [ ( 6 . 96–1 ) / ( 50–1 ) ] = 1 . 23 . After 15 s of activity , firing rate adaptation ( Eq 12 ) for MU60 is = 1 . 23 x ( 1 –e-15/22 ) = 0 . 61 imp/s , and the adapted firing rate is 21 . 04–0 . 61 = 20 . 43 imp/s . When rested and at maximum voluntary excitation ( where E ( t ) = Emax = 67 excitation units ) , the modeled muscle generated a total maximum voluntary contraction force of 2 , 216 units and generated a minimum force of 1 force unit ( 0 . 045% MVC ) at E ( t ) = 1 . The model can be given a target at some percentage of the maximum force ( e . g . 40% MVC would be 886 . 4 units of force ) . For the initial time sample ( t = 0 ) , the muscle was assumed to be completely rested and the model incremented excitation in 0 . 01 steps beginning with E ( t ) = 1 . At each excitation step , the un-adapted firing rate ( Eq 3 ) , normalized firing rate ( Eq 4 ) , associated normalized force ( Eqs 5 & 6 ) , and the actual force developed ( Eq 7 ) was determined for each MU . The total muscle force was calculated as the sum of force values across all MUs . If the total muscle force was below the target force , excitation was increased by 0 . 01 . This process was repeated until the force target was met or slightly exceeded , at which time the model was advance 0 . 1 s ( sample rate = 10 Hz ) to the next time sample . During each subsequent interval , the existing force capacity of each motor unit ( PE ) was calculated as PE from the previous sample , minus the fatigue-related change during the 0 . 1 s interval ( using Eq 10 ) : PE ( i , t ) = PE ( i , t−0 . 1 ) − 0 . 10[ FAT ( i , t ) ] ( 14 ) Where PE ( i , 0 ) = P ( i ) , as the muscle has not yet had time to fatigue . At each subsequent iteration and for each MU: ( a ) fatigue affected force capacity and contraction time , ( b ) the duration of activity ( time since recruitment ) affected firing rates , and ( c ) these factors affected normalized firing rates , normalized forces , and actual exerted forces . Thus , with sustained isotonic contractions , more excitation would be needed to meet the force target over the course of the contraction , possibly necessitating the recruitment of higher threshold MUs not initially active under rested conditions . For target force levels at 15% MVC , and from 20–90% MVC in 10% increments , simulations were run for the duration necessary for the muscle force capacity to fall below the designated target force , and this duration was considered to be the endurance time . In addition to simulating total muscle force and endurance time , the model also enabled tracking of the instantaneous force and force capacity ( absolute and relative ) of each MU throughout the simulated contractions . A simulation was also run with a 100% MVC target for 200 s so it could be compared to the time-histories reported for this type of task in a number of experimental studies . Similarly , simulations were also performed using targets of 15% , 50% and 85% MVC , with each running until the total muscle force had decreased to 15% of maximum . This provided an interesting opportunity to compare the force capacities across MUs at the end of these trials for which the defined level of total muscle fatigue ( 85% decrease in total muscle force capacity ) would be the same in all cases . In addition , we simulated fatigue involving a ‘staircase’ task in which force was maintained for 32 s at progressively increasing 20% MVC increments with a 5-s linear ramp between steps .
|
Skeletal muscle fatigue reduces strength during work and play and profoundly impairs motor function in many neuromuscular disorders . Muscle is composed of groupings of fibres called motor units and these have an extensive range of characteristics from small , weak , and fatigue-resistant to large , strong , and highly fatigable . Our model tracks the fatigue of an entire population of motor units making up a muscle . The model predicted , with good fidelity , the endurance times for a wide range of tasks and provided new insights into the complex orchestration of motor unit contributions to muscle force during fatigue . The model should have wide application in the fields of ergonomics , rehabilitation and exercise to predict and better understand the nature of both motor unit and whole muscle fatigue .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Material",
"and",
"methods"
] |
[
"material",
"fatigue",
"medicine",
"and",
"health",
"sciences",
"legs",
"classical",
"mechanics",
"pathology",
"and",
"laboratory",
"medicine",
"fatigue",
"skeletal",
"muscles",
"limbs",
"(anatomy)",
"neuroscience",
"biomechanics",
"muscle",
"contraction",
"motor",
"neurons",
"simulation",
"and",
"modeling",
"signs",
"and",
"symptoms",
"materials",
"science",
"damage",
"mechanics",
"materials",
"physics",
"research",
"and",
"analysis",
"methods",
"muscle",
"physiology",
"musculoskeletal",
"mechanics",
"musculoskeletal",
"system",
"animal",
"cells",
"muscles",
"physics",
"cellular",
"neuroscience",
"diagnostic",
"medicine",
"cell",
"biology",
"anatomy",
"physiology",
"neurons",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"physical",
"sciences",
"ankles"
] |
2017
|
A motor unit-based model of muscle fatigue
|
DNA methylation is pervasive across all domains of life . In bacteria , the presence of N6-methyladenosine ( m6A ) has been detected among diverse species , yet the contribution of m6A to the regulation of gene expression is unclear in many organisms . Here we investigated the impact of DNA methylation on gene expression and virulence within the human pathogen Streptococcus pyogenes , or Group A Streptococcus . Single Molecule Real-Time sequencing and subsequent methylation analysis identified 412 putative m6A sites throughout the 1 . 8 Mb genome . Deletion of the Restriction , Specificity , and Methylation gene subunits ( ΔRSM strain ) of a putative Type I restriction modification system lost all detectable m6A at the recognition sites and failed to prevent transformation with foreign-methylated DNA . RNA-sequencing identified 20 genes out of 1 , 895 predicted coding regions with significantly different gene expression . All of the differentially expressed genes were down regulated in the ΔRSM strain relative to the parent strain . Importantly , we found that the presence of m6A DNA modifications affected expression of Mga , a master transcriptional regulator for multiple virulence genes , surface adhesins , and immune-evasion factors in S . pyogenes . Using a murine subcutaneous infection model , mice infected with the ΔRSM strain exhibited an enhanced host immune response with larger skin lesions and increased levels of pro-inflammatory cytokines compared to mice infected with the parent or complemented mutant strains , suggesting alterations in m6A methylation influence virulence . Further , we found that the ΔRSM strain showed poor survival within human neutrophils and reduced adherence to human epithelial cells . These results demonstrate that , in addition to restriction of foreign DNA , gram-positive bacteria also use restriction modification systems to regulate the expression of gene networks important for virulence .
DNA methylation has been shown to regulate diverse pathways across all domains of life [1] . In eukaryotes , cytosine methylation regulates developmental gene expression and aberrant DNA methylation patterns have been implicated in many disease states , including cancer [2 , 3] . Although studied in a limited number of prokaryotic organisms , DNA methylation has been implicated in a myriad of cellular processes , including protection from the invasion of foreign DNA , cell cycle regulation , DNA mismatch repair , and the regulation of gene expression [4] . It was recently shown that within the genomes of over 200 prokaryotes surveyed greater than 90% contained N6-methyladenosine ( m6A ) , N4-methylcytosine ( m4C ) , or 5-methylcytosine modifications ( m5C ) [5] . These results demonstrate that DNA methylation among prokaryotes is more pervasive than originally anticipated . What remains uncertain is if DNA methylation imparts any regulatory controls influencing virulence properties or other phenotypes amongst the array of diverse prokaryotic species . DNA methylation in bacteria has been well characterized in the context of restriction modification ( RM ) systems [4 , 5] . RM systems are a mechanism of bacterial host defense to prevent the invasion of foreign DNA . RM systems are generally comprised of a site-specific restriction endonuclease ( REase ) , methyltransferase ( MTase ) , and , in some cases , a specificity subunit that together form a protein complex that cleaves foreign DNA after it enters the cell . Methylation of the host DNA at the same recognition site serves to safeguard the host chromosome from cleavage . In addition to RM systems , DNA can also be methylated by orphan MTases . Orphan MTases methylate DNA in site-specific sequences and lack an active cognate endonuclease [5 , 6] . In bacteria , the two most well studied orphan MTases are Escherichia coli DNA adenosine methyltransferase ( Dam ) and Caulobacter crescentus cell cycle regulated methyltransferase ( CcrM ) [5 , 6] . Site-specific DNA methylation by Dam and CcrM has been shown to regulate DNA mismatch repair , cell cycle progression , origin sequestration , and gene expression , demonstrating that DNA methylation imparts critical regulatory functions [6] . Despite the importance of RM systems and orphan MTases , the lack of genome-wide detection tools has hindered the identification of DNA base modifications and characterization of the physiological consequences resulting from MTase inactivation in bacteria . The use of methylation-sensitive restriction endonucleases to identify sites of DNA base modifications is limited by the sequence specificity of the recognition site , potentially missing many base modifications that could occur outside of a particular sequence context ( [5] and references therein ) . While bisulfite sequencing allows for genome-wide detection of m5C in sequence specific-contexts , no such genome-wide detection tool has been available for the detection of m6A or m4C until the recent advent of Pacific Biosciences ( PacBio ) Single Molecule Real-Time ( SMRT ) sequencing platform [7–11] . SMRT sequencing relies on differences in DNA polymerase kinetics to detect base modifications in the template strand in a sequence-context specific manner without a priori knowledge of the modification . Our group previously used the PacBio SMRT sequencing platform to complete whole genome sequencing and reference genome assembly of two strains of the bacterial human pathogen Streptococcus pyogenes , or Group A Streptococcus ( GAS ) [12 , 13] . S . pyogenes causes a wide variety of human infections , ranging from the relatively common streptococcal pharyngitis and cellulitis to the relatively uncommon , but severe , streptococcal toxic shock syndrome and necrotizing fasciitis , which have high morbidity and mortality rates [14–16] . S . pyogenes is a model bacterial pathogen , not only for the infections it produces , but also for the great diversity of toxins and virulence factors expressed by the organism and the highly complex nature of regulatory mechanisms employed to control virulence factor expression [14 , 16–18] . Indeed , S . pyogenes utilizes over 30 recognized transcriptional regulatory proteins and 13 two-component regulatory systems to coordinate virulence factor expression in response to varying environmental signals ( e . g . , carbohydrate availability , temperature , pH , oxygen tension , salt concentrations , osmolality , etc . ) , growth phase , intracellular metabolite concentrations , and signaling pheromones involved in quorum sensing [17 , 18] . DNA methylation has not been previously investigated as a significant mechanism influencing virulence factor expression within S . pyogenes , and DNA methylation may represent an unrecognized target for therapeutic intervention to help prevent or treat severe streptococcal disease . In this study , we show that in S . pyogenes strain MEW123 , a representative derivative of a serotype M28 clinical pharyngitis isolate , the active Type I RM system SpyMEW123I is responsible for the bipartite m6A motif identified throughout the genome . We show that deletion of the RM system and subsequent loss of m6A from S . pyogenes results in the down regulation of a distinct set of operons involved in streptococcal virulence . Importantly , our study shows that methylation by a Type I RM system correlates with differential expression of Mga , a major transcriptional regulator of multiple virulence factors , surface adhesins , and immune evasion factors in S . pyogenes . The results presented here demonstrate that RM systems can integrate their methylation signal to influence the expression of gene networks important for bacterial virulence .
Previously we completed whole genome assembly using PacBio SMRT sequencing with S . pyogenes strain MEW123 , a representative serotype M28 isolate used by our group to investigate streptococcal mucosal colonization [12] ( for strain list refer to Table 1 ) . To begin our investigation , we performed methylation analysis of the SMRT sequencing data . We identified m6A DNA base modifications in the MEW123 genome at the consensus sequence 5' GCANNNNNTTYG and its corresponding partner motif 5' CRAANNNNNNTGC , consistent with m6A modification motifs previously reported by Blow et al . ( Table 2 ) [5] . Within the MEW123 genome , 412 occurrences of each m6A site within the bipartite recognition motif were identified; the majority occurred in predicted coding ( 92% ) and intergenic ( 6% ) regions of the MEW123 genome . The bipartite recognition motif is characteristic of Type I RM systems , which are typically comprised of three separate subunits , including a restriction endonuclease , a specificity subunit , and a methyltransferase subunit , that act together as a single protein complex and typically act at large distances from the methylation site . The RM system annotation pipeline used in Blow et al . identified the putative Type I restriction modification system , annotated as SpyMEW123I , consisting of a three-gene cluster with separate restriction endonuclease ( hsdR ) , specificity ( hsdS ) , and methyltransferase ( hsdM ) genes , as a predicted match for modification of the identified m6A motif in S . pyogenes [5 , 19] ( Fig 1A and Fig 1B ) . This three-gene cluster exhibits high amino acid sequence homology to the Type I RM system identified in S . pyogenes SF370 at Spy_1904 ( hsdR ) , Spy_1905 ( hsdS ) , and Spy_1906 ( hsdM ) , with 99% , 87% , and 99% identity , respectively [20] . This Type I RM system is present in virtually all sequenced S . pyogenes strains to date , with rare exception reported in some emm1 strains from Japan with spontaneous deletion of a two-component regulatory system and the adjacent Type I RM system [21] . Notably , we did not detect the 5mC modifications at CmCNGG reported by Euler et al . in our PacBio SMRT sequencing results , which is not surprising given the MTase , M . SpyI , is absent from the S . pyogenes M28 serotype [22] . The REase and MTase activities of SpyMEW123I are annotated as R . SpyMEW123I and M . SpyMEW123I , respectively . To determine if the SpyMEW123I RM system was responsible for the observed m6A modifications in strain MEW123 , an in-frame deletion mutation was constructed using a plasmid vector designed for allelic replacement ( pGCP213 ) as previously described [26] ( Table 1 and Fig 1A ) . Approximately 95% of the three-gene sequence encoding the hsdR , hsdS , and hsdM genes was deleted producing strain MEW513 ( referred to as ΔRSM ) ; the in-frame deletion was confirmed by PCR amplification and Sanger DNA sequencing ( Table 1 ) . Growth of the MEW123 parent strain , referred to as wild-type ( WT ) and the ΔRSM mutant were not significantly different in rate or final growth density when measured in either the nutrient rich Todd-Hewitt medium with 0 . 2% yeast extract ( THY broth ) or the low-carbohydrate C-medium ( S1 Fig ) . To confirm a reduction in m6A base modifications and to determine the sequence context lacking m6A base modifications in the ΔRSM strain , genomic DNA was isolated and sequenced via PacBio SMRT sequencing . Modification analysis showed loss of detectable m6A base modifications at 5' GCANNNNNTTYG and 5' CRAANNNNNNTGC sites , demonstrating that streptococci with a SpyMEW123I deletion no longer have m6A DNA base modifications at the consensus sequence identified in the WT strain ( Fig 1B , Table 3 ) . A number of additional methylation events were identified in MEW513; however , these occurred at far lower frequencies compared to the modifications at the consensus sequences in the parent strain and the quality of the read scores ( Mod QV ) were low compared to the RSM-dependent modifications . Based on these low quality read scores , we feel it is unlikely that these additional modifications reflect compensatory methylation events . Furthermore , SMRT sequencing of the MEW513 genome did not identify any unforeseen mutations outside of the in-frame deletion within hsdRSM that we anticipated . To further confirm that the MTase component of the RSM gene cluster , M . SpyMEW123I , was indeed responsible for producing m6A DNA modifications , genomic DNA was harvested from the WT and the ΔRSM strain and spotted onto a nitrocellulose membrane for immunodetection using an α-m6A antibody . We found that the α-m6A signal was substantially reduced in genomic DNA blots from the ΔRSM strain compared to the WT parent , suggesting a significant and near complete reduction in m6A base modifications in the ΔRSM strain ( Fig 1C ) . Complementation in trans of the ΔRSM mutant with a plasmid encoded copy of the three gene cluster ( hsdRSM ) produced strain MEW552 ( referred to as ΔRSM/pRSM ) and successfully restored detection of the α-m6A signal to levels comparable to the WT strain ( Fig 1C ) . These results demonstrate that the MTase activity of SpyMEW123I is responsible for base modifications at 5' GCANNNNNTTYG and 5' CRAANNNNNNTGC sites in vivo . Deletion of the three-gene cluster , hsdRSM , containing the predicted endonuclease , specificity , and methylation gene subunits abolished m6A base modifications in the ΔRSM mutant strain . In Type I RM systems , DNA cleavage is dependent on the MTase and specificity subunits , in addition to the REase subunits which are often independently regulated by a separate promoter [29] . Fully unmethylated recognition motifs induce REase activity that results in DNA cleavage typically between two fully unmethylated motifs at sites distant from the recognition sequence; this distance may range from 40 base pairs to several kilobases away from the RM site . Type I MTases can function to add m6A de novo on fully unmethylated DNA or act as maintenance MTases at hemi-methylated recognition sites [29–31] . Additional mechanisms also protect DNA from restriction , including proteolysis of the REase subunits or protection by DNA binding proteins that can protect unmethylated sites from cleavage in the host chromosome [32] . To establish the functionality of the REase component of SpyMEW123I , a transformation efficiency assay was performed using pJoy3 plasmid DNA methylated in an E . coli host ( Table 1 ) . This 6 . 3 kb plasmid contains eight predicted Dam MTase RM sites ( 5' GATC ) and is delivered in its native double-stranded circular form via electroporation into electrocompetent S . pyogenes where the plasmid is maintained and replicates extrachromosomally [27] . In addition to testing the effect of deleting the entire hsdRSM gene cluster in the ΔRSM mutant strain , we constructed an additional strain derivative of MEW123 with a spectinomycin-resistance cassette disrupting the hsdR REase gene subunit alone producing strain MEW489 ( referred to as ΩRE , Table 1 ) . If the SpyMEW123I RM system has true restriction enzyme activity to foreign-modified DNA , then we would expect that inactivating the hsdR gene subunit , either individually or within the entire RSM gene cluster , would enhance the transformation efficiency of the plasmid . Indeed , we found that the rates of transformation with foreign-methylated plasmid DNA increased significantly for both the ΔRSM mutant and the ΩRE mutant strains compared to the WT parent strain , providing evidence that the restriction endonuclease component of SpyMEW123I is active and functional ( Fig 2A ) . We were unable to compare our complementation strain ΔRSM/pRSM for transformation efficiency as this strain already carries the pJoy3 plasmid encoding the hsdRSM gene cluster . As a control , we undertook transformation of a MEW123 mutant in the gene encoding the C5a peptidase , scpA ( strain 489 or ΩscpA ) , as mutants in this gene would not be expected to show enhanced transformation efficiency; as expected , the transformation efficiency of ΩscpA was not significantly different than the WT ( Fig 2A ) . Interestingly , inactivation of the endonuclease subunit hsdR alone in the ΩRE mutant strain conferred significantly greater transformation efficiency than that observed in the ΔRSM mutant ( Fig 2A ) . In many Type I RM systems the restriction subunit is generally under control of a separate promoter than the specificity and methylation subunits in the RSM gene cluster [29] . We found that the α-m6A signal generated by dot blot of genomic DNA from the ΩRE strain was intermediate in intensity between the WT and ΔRSM strains ( Fig 1C ) . This result suggests that the methyltransferase subunit was still functional in the ΩRE strain , but that there may have been some degree of polar effect from the spectinomycin-resistance cassette used to inactivate hsdR that was reducing transcription of the hsdS and hsdM gene products compared to WT levels . We speculate that the residual functional activities of the specificity and methyltransferase subunits in the ΩRE mutant strain , even though less than WT levels , may have conferred additional stability to the incoming foreign-methylated plasmid DNA , possibly offering protection from other minor endonucleases , thereby enhancing overall transformation efficiency . As discussed above , the MTase activity of Type I RM systems may function to maintain the state of hemi-methylated or fully methylated DNA , whereas REase cleavage only occurs on fully unmethylated DNA . Thus , RM sites can exist in the genome in a hemi-methylated state while still conferring protection from digestion [29] . Having shown that SpyMEW123I functions as an active RM system , we sought to establish the fraction of reads that were called as methylated at each recognition site . Our analysis of sequencing data from WT S . pyogenes MEW123 found substantial variation in the fraction of sequencing reads modified at RM sites ( Fig 2B ) . The fraction of reads called as modified at a given RM site did not appear to be dependent on orientation or genome position . Of the m6A modifications called at RM sites , 4 . 9% of sites were called as m6A modified in less than 50% of sequencing reads , 23 . 7% of sites were called as modified in between 50–75% of sequencing reads , and finally 71 . 4% of sites were called as modified at greater than 75% of aligned reads . Previous studies have also reported heterogeneity in the frequency of SMRT sequencing reads with base modifications; it has been hypothesized that these differences are due to timing in DNA replication and subsequent methylation [33–35] . Whether there is a temporal component accounting for the heterogeneity in m6A DNA modifications , and whether this impacts other functions of m6A modifications , such as in influencing gene transcription in S . pyogenes , is unknown . Given the heterogeneity observed in the fraction of reads called as m6A methylated , we hypothesized that m6A modifications produced by the SpyMEW123I RM system might have additional functions outside of host protection from foreign DNA prompting the experiments below . In addition to functioning in RM systems , m6A base modifications from orphan MTases have been shown to function in cell cycle regulation , DNA mismatch repair , and the regulation of transcription [4] . In the pathogenic Escherichia coli serotype O104:H4 strain C227-11 associated with hemolytic uremic syndrome , deletion of the ϕStx104 RM system results in the differential expression of over 38% of the genes , including genes involved in motility , cell projection , and cation transport [33] . Mismatch repair is not coupled to methylation in S . pyogenes or most other gram-positive bacteria [36] . Therefore , we asked if m6A originating from the SpyMEW123I RM system in S . pyogenes might have additional functions outside of host defense from foreign DNA . We isolated RNA from streptococcal cells during mid-exponential growth phase in C media broth culture from WT and ΔRSM strains followed by RNA-sequencing . The results of the differential expression analysis showed that 20 genes were differentially expressed in the ΔRSM strain compared to WT ( adjusted p . val < 0 . 05 , log2 fold change >1 , data set available at NCBI repository ) . Interestingly , all 20 genes were down regulated in ΔRSM relative to WT suggesting a common regulatory mechanism ( Fig 3A and 3B , Table 4 ) . The three genes ( hsdRSM ) of the SpyMEW123I RM gene cluster showed the greatest log2 fold change in expression of -10 . 8 , -10 . 7 , and -11 . 7 , respectively , which was expected because these genes were deleted in the ΔRSM strain . The majority of the differentially expressed genes are located in approximately 6 separate operons or gene clusters as indicated in Table 4 . Interestingly , several of these gene groups are transcriptionally regulated , at least in large part , by activity of the multiple gene regulator protein , Mga [37–39] . During mid-exponential growth phase , Mga acts as a transcriptional activator to regulate a core set of virulence factors at the mga locus [37] . The mga locus consists of several components: a ) the M protein ( emm gene ) a major surface protein involved in resistance to phagocytosis and intracellular killing by neutrophils and used to distinguish S . pyogenes isolates , b ) a fibronectin-binding protein that binds host complement regulator factors , c ) an emm-like protein that binds IgG and fibrinogen , d ) the C5a peptidase ( ScpA ) which cleaves C5a chemotaxin , e ) the enn protein that binds IgA , and f ) the mga gene itself . All genes at the mga locus displayed log2 fold changes ranging from -1 . 2 to -8 . 7 in the ΔRSM strain relative to WT ( Fig 3A ) . To confirm this differential expression , we again isolated total RNA from strains during mid-exponential growth phase in C media broth culture and performed quantitative RT-PCR for detection of transcripts mga , emm28 , and scpA ( Fig 3B ) . Consistent with the RNA-seq results , the qRT-PCR results showed that these genes were significantly down regulated in the ΔRSM strain , with approximately 5-fold to over 300-fold decreased expression in the ΔRSM strain relative to WT ( Fig 3B ) . Complementation in trans in the ΔRSM/pRSM strain restored transcript expression patterns similar to WT values . Deletion of the mga gene produced qRT-PCR results in a similar trend to the ΔRSM strain for the examined transcripts , with significantly decreased detection of emm28 and scpA transcripts; mga transcript was not detected in the Δmga strain ( Fig 3B ) . Examination of these transcripts in the hsdR insertional inactivation mutant ΩRE showed transcript detection of mga and emm28 comparable to WT levels , with detection of scpA transcript approximately four to five-fold of WT levels . This transcript pattern was very different than those of the ΔRSM and Δmga mutant strains and more similar to the WT pattern . Even though the spectinomycin resistance cassette insertion into hsdR may have produced some polar effect with slightly decreased methyltransferase activity as noted on the α-m6A dot blot ( Fig 1C ) , it seems sufficient residual m6A base modifications persisted to not significantly disrupt gene expression ( Fig 3B ) . Taken together , these results from RNA sequencing and qRT-PCR provide evidence that m6A base modifications correlate with patterns of differential gene expression in S . pyogenes , including those of several recognized virulence factors and major regulators of virulence gene expression . Given that the genes in the Mga regulon were significantly down regulated in the ΔRSM strain relative to WT , we were interested in determining the impact of disrupting m6A DNA modifications on S . pyogenes virulence using a murine subcutaneous infection model [40 , 41] . C57BL/6J mice were inoculated at the shaved flank with 1 x 107 CFUs of either MEW123 ( WT ) or the ΔRSM mutant strain and resulting skin ulcers were photographed daily for sizing the skin ulcer area . As shown in Fig 4A , there was no significant difference in skin lesion size at day two post-infection in comparison of the mice infected with either the WT or the ΔRSM strains . However , by three to four days post-infection , and for the remainder of the experiment , the skin lesions of mice infected with the ΔRSM strain were significantly larger than those of mice infected by the WT strain ( Fig 4A ) . No strain caused a lethal infection among any of the mice with the 1 x 107 CFU inoculum . Representative images of skin lesions for mice infected with the WT , the ΔRSM strain , and the complemented ΔRSM/pRSM strain over time are shown in Fig 4B , with skin lesions of mice infected with the ΔRSM strain notably larger on average at 4 and 6 days compared to those of mice infected with the WT or complemented strain . Complementation of the ΔRSM mutation in trans by strain MEW552 ( ΔRSM/pRSM ) produced murine skin lesions smaller than the ΔRSM mutant but not significantly different than the WT strain throughout the duration of the experiment ( Fig 4C ) . Skin lesion sizes reached a mean peak size at four to six days post infection . To determine if the difference in skin lesion size correlated with the concentration of viable streptococci at the site of infection , the skin lesions of mice were dissected and homogenized at day four post-infection to obtain viable CFU counts . Upon dissection , we made the observation that skin lesions from mice infected with the ΔRSM strain were grossly more purulent than lesions of mice infected by the WT and complemented ΔRSM/pRSM strains . The skin lesions contained on average CFU counts of approximately 1 x 106 to 1 x 107 CFUs; while there was a slight trend to higher CFU counts on day four post-infection for the ΔRSM streptococci compared to the WT and complemented strain CFUs , there were no statistically significant differences in CFU counts between these groups ( S2 Fig ) . We noted that skin lesions of mice infected with the WT and complemented strain ΔRSM/pRSM strains seemed to heal more quickly than those of mice infected with the ΔRSM strain ( Fig 4C ) . With the subcutaneous ulcer model , skin lesion size tends to correlate closely with the degree of the host immune response , with particular regards to the neutrophil influx [40 , 41] . To compare the inflammatory response in skin lesions of mice infected with the WT and the ΔRSM strain , we performed skin biopsies for cytokine analysis and histologic examination at six-days post-infection; this time point was chosen as it was the time point with the greatest difference in skin lesion size between the experimental groups . Measurements of interleukin-1 beta ( IL-1β ) , interleukin-6 ( IL-6 ) , interleukin-17A ( IL-17A ) , and tumor necrosis factor alpha ( TNFα ) , were obtained as evidence of pro-inflammatory cytokine activity . Cytokine concentrations for all four cytokines measured were significantly greater from mice infected with WT streptococci than mice mock-infected with sterile phosphate-buffered saline ( PBS ) ( Fig 5A ) . Cytokine concentrations from mice infected with the ΔRSM strain were significantly greater than mock-infected or mice infected with the WT strain ( Fig 5A ) . Furthermore , histologic analysis of skin lesions shows predominantly increased neutrophil influx , but also a modest increase in the number of macrophages in the subcutaneous tissue of mice infected with the ΔRSM strain compared with WT ( Fig 5B ) . Infiltration of T lymphocytes was not appreciably different between skin lesions of mice infected with WT or the ΔRSM strain ( Fig 5B ) . Cytokines IL-6 and IL-17A , in particular , are important for coordinating neutrophil trafficking to areas of infection [42–44] . Our results in mice infected with the ΔRSM strain showing enhanced pro-inflammatory cytokine detection , increased neutrophil infiltration , and larger skin lesions , suggests an effect of altered gene transcription patterns in the ΔRSM strain and a more robust host inflammatory response compared to mice infected with the WT parent strain . Given the known association of several of the streptococcal gene transcripts down regulated in the ΔRSM strain , including mga , emm28 , and scpA , with immune evasion properties , we hypothesized that m6A DNA modifications and proper regulation of gene expression are important contributors to immune evasion strategies and/or disruption of host immune responses by S . pyogenes . To determine if the loss of specific virulence factors recapitulates the phenotype of the ΔRSM strain in the murine subcutaneous ulcer model , we infected mice with derivatives of strain MEW123 with in-frame deletions of mga ( strain MEW480 , ΔMga ) , and spectinomycin-resistance cassette disruption mutations of emm28 ( strain 409 , Ωemm28 ) and scpA ( strain 380 , ΩscpA ) . As shown in Fig 4C , infection of mice by the ΔMga strain produced skin lesions significantly larger than the WT strain and comparable to the ΔRSM strain in size throughout the experiment . Infection by the Ωemm28 strain was not statistically different than the WT strain at day 2 and day 4 post-infection; however , by day 6 and day 8 post-infection , the Ωemm28 strain produced lesions that were statistically significantly larger than the WT ( Fig 4C ) . Infection of mice by the ΩscpA strain produced the widest range of murine skin lesion sizes , with some mice having very large lesions following infection ( Fig 4C ) ; however , at no time point was the average size of the lesions produced by the ΩscpA strain statistically different than WT . Overall , these results suggest that the presence of m6A DNA base modifications produced by M . SpyMEW123 activity correlate with differential transcriptional expression of several S . pyogenes virulence factors , especially those within the Mga operon , and that these seem to influence host-pathogen interactions at the site of infection . A major function of the S . pyogenes M protein is to promote streptococcal survival , resisting killing by human leukocytes by interfering with bactericidal activity within neutrophils following phagocytosis [45 , 46] . Staali et al . found that S . pyogenes strains with or without M protein underwent phagocytosis by neutrophils to similar levels , but only strains expressing M protein survived intracellularly whereas strains lacking M protein expression were rapidly killed [45] . Given our findings that elimination of m6A DNA modifications was associated with decreased transcript expression for mga and emm28 , we wished to compare survival within human neutrophils . Purified human neutrophils were incubated with WT or ΔRSM S . pyogenes strains using a neutrophil bactericidal assay similar to a previous report [45] . Briefly , streptococci and neutrophils were mixed together allowing the neutrophils to internalize S . pyogenes strains followed by elimination of extracellular bacteria with penicillin and gentamicin . It was previously determined that there was no significant difference in susceptibility to penicillin and gentamicin at the high concentrations used in these experiments between the WT or ΔRSM strains ( S1 Fig ) . Streptococcus surviving within neutrophils were liberated by treatment with the detergent saponin and plated for viable CFUs . As shown in Fig 6 , we utilized serotype M14 HSC5 and a derivative strain with disruption in the M14 emm gene ( Ωemm14 ) as positive and negative controls , respectively . As expected , the Ωemm14 mutant was significantly attenuated for intracellular survival within neutrophils compared to the M14 parent strain ( Fig 6 ) . Similarly , we compared survival of the MEW123 parent strain ( M28 ) and its cognate strain with disruption of the M28 emm gene ( Ωemm28 ) or the ΔRSM mutant . We found that both the Ωemm28 and the ΔRSM mutant were significantly attenuated for intracellular survival compared to the M28 parent strain , further confirming the role of M protein in promoting intracellular neutrophil survival by the serotype M28 MEW123 strain , in addition to demonstrating correlation of m6A DNA base modifications with differential expression of M protein ( Fig 6 ) . These results provide further support for m6A DNA base modifications in S . pyogenes as important for promoting streptococcal virulence , possibly by influencing virulence factor expression . From the RNA-seq results we found that the ΔRSM strain had significantly decreased transcript expression of several recognized and known adhesin proteins , including M28 , M-like protein , collagen-binding protein , and fibronectin-binding proteins , as well as several hypothetical surface proteins [38 , 47–49] . As a group , serotype M28 S . pyogenes are overrepresented in cases of human infection within the female urogenital tract , including vulvovaginitis and puerperal sepsis ( a . k . a . “childbed fever” ) [50–53] . Serotype M28 S . pyogenes have a particular predilection for cervical and vaginal epithelium due to surface proteins , including protein R28 among others , which may explain the overrepresentation of this serotype with infections in this niche [15 , 54] . Therefore , we asked if m6A DNA modifications influenced adherence of the serotype M28 MEW123 strain to human vaginal epithelial cells . As shown in Fig 7A , disruption of m6A DNA modifications in the ΔRSM strain was indeed associated with significantly decreased adherence to human vaginal epithelial cells in vitro compared to the WT parent strain . The attenuation in vaginal epithelial cell adherence by the ΔRSM strain was comparable to a strain lacking expression of the M protein ( Ωemm28 ) , suggesting that decreased expression of M protein , among other adhesins , by the ΔRSM strain was at least partly responsible for decreased adherence ( Fig 7A ) . To determine if impaired adherence to human vaginal cells in vitro translated to impaired vaginal mucosal colonization in vivo , we utilized a murine vaginal model and compared streptococcal carriage burdens over time [40 , 55] . In contrast to the results of the in vitro adherence assay , using the murine vaginal carriage model we found no significant difference in vaginal streptococcal burdens in comparison of mice inoculated with either the WT or the ΔRSM strains over the course of the 28-day experiment ( Fig 7B ) . Given that human cells are the natural hosts of S . pyogenes , this may be an example of the human-restricted nature of S . pyogenes in which a murine model cannot adequately replicate the natural human environment in which this pathogen evolved to survive . Nevertheless , our overall results showed several key differences in virulence phenotypes correlating with alterations in gene transcription associated with streptococcal m6A DNA methylation .
In this report , we provide evidence that m6A DNA base modifications influence gene transcription patterns and overall virulence properties in a major gram-positive bacterial pathogen of humans , S . pyogenes . The S . pyogenes RM system , SpyMEW123I , is a Type I RM system and is responsible for the majority of m6A base modifications distributed throughout the S . pyogenes genome . The target consensus sequences identified by our study , 5' GCANNNNNTTYG and its corresponding partner motif 5' CRAANNNNNNTGC , were consistent with m6A motifs identified in S . pyogenes previously reported by Blow et al [5] . We found approximately 412 occurrences of each m6A site with the majority found within coding regions . Interestingly , we found that not all m6A sequence motifs were consistently modified to the same extent; only about 70% of consensus sites were modified in at least 75% of sequencing reads , suggesting that m6A modifications may be intermittently present with additional functions beyond simple protection from restriction , including influencing gene expression patterns based on timing of hemi- or full-methylation status . It is not known at this time whether all of the m6A sites , or only the sites within the intergenic regions , would participate in influencing transcriptional expression , but methylation events modifying access of transcriptional regulators to intergenic promoter regions would be a potential mechanism . With the introduction of SMRT sequencing , groups have now identified m6A DNA modifications within a diversity of prokaryotes , including E . coli , Campylobacter jejuni , Salmonella enterica serovar Typhimurium , Vibrio breoganii , Geobacter metallireducens , Chromhalobacter salexigens , Bacillus cereus , and Borrelia burgdorferi [33–35 , 56 , 57] . Additional evidence of 5-methylcytosine ( m5C ) DNA modifications influencing transcriptional expression of multiple genes with an impact on several phenotypic traits has recently been described in Helicobacter pylori , further expanding the recognized influence of prokaryotic methylation modifications [58] . Some of the DNA modifications described have been linked to orphan MTases without an associated endonuclease , such as DNA Adenine Methyltransferase ( Dam ) of S . enterica , E . coli , and Haemophilus influenzae [56 , 59 , 60] . Uncoupling DNA methylation from restriction endonuclease protection is conceptually easier to envision with an orphan MTase , freeing the orphan MTase to have roles in DNA mismatch repair and influencing gene expression of potential virulence factors [6] . Indeed , Dam-dependent DNA modifications in S . enterica have been linked to alterations of gene expression and virulence [56] . However , two examples have recently been reported in C . jejuni and B . burgdorferi of intact RM systems also influencing gene expression patterns [34 , 35] . Both of the RM systems in these organisms are representatives of Type IIG RM systems , which differ significantly from the Type I RM system described here for S . pyogenes in that they consist of a single polypeptide with both REase and MTase activity [34 , 35 , 61] . The effects on gene expression conferred by these systems in C . jejuni and B . burgforferi were noted by Casselli et al . to be more modest in terms of numbers of genes influenced by m6A base modifications when compared to the larger number of transcriptional changes found from the standalone activity of Dam MTase in Salmonella [34 , 35 , 56] . It would seem that with an intact RM system the conditions involved in determining gene expression is more stringent and regulates a fewer number of genes than orphan MTases . DNA methylation from Type I RM systems has also been well-established in phase variation in a number of gram-positive pathogens , including Streptococcus pneumoniae , Streptococcus suis , Listeria monocytogenes , and Mycoplasma pulmonis , which can have downstream effects on gene expression [62–65] . In phase variation , switching of specificity subunits of Type I RM systems results in cells with different sites of methylation within the population , which can create heterogeneity in gene expression . The role of methylation in phase variation differs from our findings here as we show that loss of methylation at a single site ( i . e . not switching of specificity subunits to create methylation at diverse sites ) results in the down regulation of a very defined subset of genes . The M . SpyMEW123I MTase activity we describe here modifies 412 sites in the MEW123 genome , whereas Dam-modified recognition sites approximate 19 , 000 per chromosome [56] . Perhaps the context of the m6A recognition motif in a particular intergenic promoter region , combined with specific transcription factors sensitive to the presence or absence of m6A modifications , determines the specificity of which genes an intact RM system will influence . Our results reported here demonstrate that the S . pyogenes Type I RM system is functional as a protective mechanism with restricting uptake of foreign DNA ( Fig 2A ) . Similar results were found by Okada et al . in a series of emm1 S . pyogenes isolates from Japan with spontaneous deletions in their Type I RM systems; isolates lacking the Type I RM system had significantly increased rates of transformation with foreign plasmid DNA [21] . While their study did not specifically address virulence properties of isolates lacking the RM system , the authors speculated that enhanced rates of DNA uptake and transformation exhibited by strains lacking REase activity may be beneficial by allowing uptake of potentially advantageous genes from the environment contributing to overall fitness . Inactivation of the SpyMEW123I RM system was associated with significant dysregulation of gene transcript expression in broth culture , with 20 genes from at least six separate gene clusters/operons significantly down regulated ( Table 4 ) . Notable among the down regulated genes were the trans-acting regulator Mga , the M-like protein , M28 protein , C5a peptidase ( ScpA ) , a cell surface protein , a collagen-like surface protein ( SclA ) , the Serum Opacity Factor ( SOF ) , and a fibronectin-binding protein ( SfbX ) . Most of these genes are regulated by the Mga transcriptional regulator in serotypes that have been investigated . Mga is a ubiquitous stand-alone regulator primarily active during exponential growth phase and is responsible for influencing expression of over 10% of the S . pyogenes genome , primarily genes involved in metabolism , but also many virulence factors including adhesins and surface proteins involved in immune evasion [37 , 38] . Mga binds to upstream promoter regions to activate high-level transcription of genes in the Mga core regulon [66] . The majority of Mga-regulated promoters , including most of the genes in the core Mga regulon , contain a single Mga binding site centered around position -54 and overlapping the -35 region of the gene promoter , likely interacting with the α-subunit of RNA polymerase [67] . In theory , m6A base modifications at or around this site could potentially influence Mga and RNA polymerase binding to the promoter region , perhaps by stabilizing or localizing Mga to the proper site , promoting activation of gene transcription . Consistent with this hypothesis , examination of the genome sequences upstream of the Mga open reading frame for S . pyogenes strains MEW123 , MEW427 , and SF370 , all reveal the existence of the m6A consensus motifs approximately 800 bp upstream of the mga start codon [12 , 13 , 20] . It is unclear if , or how , this m6A motif site located upstream of the predicted Mga promoter region activates Mga expression . The mechanism of m6A-dependent regulation of the mga locus is the subject of active investigation by our group . Regulation of virulence factor expression in response to different environmental cues and stresses is critical to the success of S . pyogenes survival and pathogenesis . Over 30 recognized transcriptional regulatory proteins and 13 two-component regulatory systems must function to coordinate virulence factor expression properly [17 , 18] . We found that loss of m6A DNA modifications in our ΔRSM mutant correlated with significant changes in virulence properties of S . pyogenes . In a murine model of subcutaneous ulcer formation , we noted that mice infected with the ΔRSM mutant displayed enhanced inflammatory responses compared to mice infected with the WT strain , with comparatively larger skin lesions , increased detection of pro-inflammatory cytokine levels , and enhanced neutrophil infiltrates on histologic examination ( Figs 4 and 5 ) . Disruption of m6A DNA modifications and an associated dysregulation of gene transcript expression may result in failed activation of multiple important adhesins and streptococcal proteins involved in evading host immunity ( Fig 3 and Table 4 ) . For example , neutrophilic infiltration in response to bacterial infections is enhanced by activity of host chemotaxins , chiefly complement protein C5a . A major virulence determinant of S . pyogenes aiding immune evasion is to degrade complement C5a through activity of ScpA , a surface-expressed , serine-protease specifically degrading host C5a and interfering with neutrophil recruitment [68] . We found that the ΔRSM mutant exhibited significantly decreased transcript expression for ScpA which may partly explain a more exaggerated neutrophil response to infection with the ΔRSM mutant strain , resulting in more inflammation and larger skin lesions ( Figs 4C and 5 ) . Previous investigation into the contribution of ScpA to host immune responses was performed using a murine air sac model of subcutaneous infection performed by Ji et al . [69]; air sacs infected with S . pyogenes lacking ScpA expression exhibited a significantly enhanced host inflammatory response compared to the WT parent , with a neutrophil predominance analogous to our results . Another report found similar to slightly larger skin lesions in mice infected subcutaneously with S . pyogenes lacking ScpA compared to WT [70] . The effect of S . pyogenes virulence factors in murine models is not always similar to activity in the human environment; it is known that ScpA does cleave murine C5a , but at slower rates compared to human C5a , and these differences may impact our ability to detect phenotypes in these non-human systems [71] . Similar to our own results with the ΩscpA strain infections , the results reported by Li et al . were not statistically significant suggesting that the individual contribution of ScpA in this murine model may be modest , but when the expression of multiple virulence factors is disrupted the effects may be more apparent . Indeed , our experiments in the skin lesion model with the ΔRSM and the Δmga strains showed significant differences in lesion size and inflammatory response compared to the WT and complemented mutant strains . Both of these mutant strains would be expected to have similar patterns of differential gene expression and as result they phenocopy each other in this model . Decreased expression of several adhesins and other factors may have contributed to enhanced spread of the infection together with an exaggerated host inflammatory response resulting in larger areas of inflammation and larger skin lesion formation . Decreased M protein expression , among other adhesins , also explains the decreased in vitro adherence of the ΔRSM mutant to human vaginal epithelial cells . Interestingly , the decreased adherence to human vaginal epithelial cells in vitro did not correlate with disrupted carriage in the murine vaginal mucosa colonization model . This suggests that there are additional adhesins not influenced by m6A DNA modifications that are important for promoting and maintaining carriage in vivo . One example would be the R28 adhesin of serotype M28 S . pyogenes strains , which is a major streptococcal adhesin to human cervical epithelial cells [54] . Our RNA-sequencing experiments did not find significant differences in the transcription of the MEW123 R28 gene ( AWM59_02815 ) between WT and the ΔRSM mutant ( full data set available in NCBI repository ) . With only 20 genes significantly downregulated in the ΔRSM mutant clearly not all major S . pyogenes adhesins and virulence factors are impacted by m6A DNA modifications . Our data show that only a few gene operons , or regulons as in the case of Mga , are differentially expressed in the absence of m6A base modifications in S . pyogenes and that down regulation of these genes impacts virulence . In this study , we have demonstrated that the SpyMEW123I RM system and m6A DNA modifications in S . pyogenes significantly influence DNA restriction activity , in addition to correlating with differential gene transcription and virulence properties of this important human pathogen . Disruption of the SpyMEW123I Type I RM in S . pyogenes altered the transcriptional profile of the mutant strain resulting in attenuated virulence and impaired evasion of the host immune response in both in vitro and in vivo models . Similar to our results , disruption of Type IIG RM systems in C . jejuni and B . burgdorferi also interfered with genetic regulation of virulence factors of those pathogens [34 , 35] . Together , these findings demonstrate that intact RM systems in these bacterial pathogens , and likely many other prokaryotes , can exert multiple functions , including restriction-mediated protection from foreign DNA in addition to influencing gene expression . Understanding how m6A DNA modifications influence virulence properties in these organisms could potentially identify targets for therapeutic intervention , potentially changing patterns of virulence factor expression resulting in strain attenuation helping to prevent human disease . Further investigation is necessary to fully comprehend the many functions of DNA methylation and the complex nature of bacterial physiology and pathogenesis .
Experimental protocols involving the use of mice were reviewed and approved by the Institutional Animal Care and Use Committee ( IACUC ) of the University of Michigan Medical School ( Ann Arbor , MI , USA ) . The University of Michigan IACUC complies with the policies and standards as outlined in the Animal Welfare Act and the “Guide for the Care and Use of Laboratory Animals , ” [72] . The protocol numbers approved by the University of Michigan IACUC are as follows: Skin and Soft Tissue Infection Model of Streptococcus pyogenes Virulence ( PRO00007495 ) , and Murine Vaginal Colonization Model for Streptococcus pyogenes ( PRO00007218 ) . For consistency , all experiments utilized female C57BL/6J mice at approximately 6 weeks of age at the time of use . Mice were purchased from The Jackson Laboratories ( catalog #000664 ) ( Bar Harbor , ME , USA ) , and maintained in a University of Michigan animal facility with biohazard containment properties . Following arrival , mice were allowed to acclimatize in the facility for one week prior to beginning experiments . When manipulated , mice were briefly sedated by inhalation of isoflurane via drop jar dosing . Animals were inspected at least once daily for evidence of suffering , manifested by significantly diminished or no activity , decreased appetite , poor grooming , increased respiratory rate , or weight loss greater than 15% of body weight; if evidence of suffering was identified , then the mouse was euthanized . Euthanasia was primarily through carbon dioxide asphyxiation with a subsequent secondary method of euthanasia , including induction of bilateral pneumothorax , decapitation , and/or removal of a vital organ . The principal strain used in this study was S . pyogenes MEW123 , a streptomycin-resistant ( rpsLK56T ) , serotype M28 pharyngeal isolate [55] . Other strains used are listed in Table 1 . Growth rates and yields of MEW123 and associated mutant strains were measured using a Synergy HTX plate reader ( BioTek , Winooski , VT , USA ) in 96 well plates ( Greiner Bio-One , Monroe , NC , USA ) . Briefly , 4μl of overnight culture grown in THY broth was inoculated into 200μl of the described fresh media , with identical strains and conditions measured in at least triplicate . Growth was at 37°C , room air , in static conditions for 12 hours and OD620nm was measured every 3 seconds . Unless otherwise noted , all S . pyogenes strains had equivalent growth rates and yields under all in vitro conditions tested ( S1 Fig ) . Routine culture of S . pyogenes was performed in Todd-Hewitt medium ( Becton Dickinson , Franklin Lakes , NJ , USA ) supplemented with 0 . 2% yeast extract ( Difco Laboratories , Detroit , MI , USA ) ( THY media ) . Where required , Bacto agar ( Difco ) was added to a final concentration of 1 . 4% ( w/v ) to produce solid media . Gene expression experiments used C-Medium , a lower-glucose , higher-protein media that more closely resembles in vivo conditions [73] . Incubation was performed at 37°C under anaerobic conditions ( GasPack™ , Becton Dickinson ) for solid media , or in sealed tubes without agitation for broth media . Aerobic culture was conducted as described [74] . For inoculation of mice , S . pyogenes was harvested from culture in THY broth at early logarithmic-phase ( OD600 0 . 2 ) , washed once in PBS , briefly sonicated on ice to break up long streptococcal chains , and resuspended in PBS to 108 CFU/mL . Molecular cloning used Escherichia coli strain DH5a ( Invitrogen , Grand Island , NY , USA ) cultured in LB broth . When appropriate , antibiotics were added at the following concentrations: erythromycin , 500 μg/mL for E . coli and 1 μg/mL for S . pyogenes; chloramphenicol , 20 μg/mL for E . coli and 3 μg/mL for S . pyogenes; spectinomycin , 100 μg/mL for both E . coli and S . pyogenes; and streptomycin , 1000 μg/mL for S . pyogenes . In some experiments , growth was monitored in THY broth supplemented with either penicillin , gentamicin , or erythromycin at concentrations ranging from 0 . 05 μg/mL to 100 μg/mL . All antibiotics were obtained from Sigma Chemical Co . , St . Louis , MO , USA . Streptococcus pyogenes MEW123 was used as a source strain for DNA , Genbank CP014139 . 1 [12] . Bacterial strains and plasmid vectors are listed in Table 1 . The primers used for PCR amplification and cloning are listed in Table 5 . For cloning and routine DNA Sanger sequencing , the Phusion High-Fidelity DNA Polymerase ( New England Biolabs , Inc . , Ipswich , MA , USA ) was used . For routine endpoint PCR amplification standard Taq DNA Polymerase was used ( New England Biolabs , Inc . ) . Polymerase chain reaction products were digested with indicated restriction enzymes and ligated to pJRS233 or pGCP213 for in-frame deletions , pSPC18 for insertional mutations , or pJoy3 as a plasmid vector for transformation efficiency assays . In-frame deletion mutants and insertional mutants were constructed essentially as described [25 , 26] , [28] , and [27] , respectively . The ΔRSM in-frame deletion allele was cloned by splice overlap extension ( SOE ) PCR [75] . Corresponding GenBank accession numbers for the MEW123 restriction endonuclease gene hsdR , specificity subunit hsdS , and the methyltransferase subunit hsdM , are AWM59_07895 , AWM59_07900 , and AWM59_07905 , respectively . The upstream region of the gene cluster was PCR amplified using primers MEW123 Del-RSM F1 and MEW123 Del-RSM R2 , producing a 1 . 02 kb amplicon . The downstream region of the gene cluster was PCR amplified using primers MEW123 Del-RSM F3 and MEW123 Del-RSM R4 , producing a 1 . 02 kb amplicon . These two amplicons contain complementary ends that anneal together and essentially will produce an in-frame deletion of the three-gene restriction endonuclease , specificity subunit , and DNA methyltransferase open reading frames . The two amplicons were mixed together as template and further amplified using primers MEW123 Del-RSM F1 and MEW123 Del-RSM R4 , the resulting amplicon was approximately 2 . 04 kb and contained a unique EcoRI site at the 5’ end and a unique HindIII site at the 3' end . The resulting amplicon was digested with EcoRI and HindIII , and inserted within same restriction sites of the E . coli to S . pyogenes temperature-sensitive vector for allelic replacements , pGCP213 [26] , producing plasmid pKJ24 . The pKJ24 plasmid was confirmed by Sanger DNA sequencing using primers MEW M13 F and MEW M13 R , which bind just outside and flank the multiple cloning site region within pGCP213 . Electrocompetent cells of MEW123 were prepared and transformation was performed essentially as previously described [76] . The pKJ24 plasmid carrying the RSM in-frame deletion was transformed into electrocompetent S . pyogenes MEW123 through electroporation with conditions as described above . Erythromycin-resistant transformants were handled according to the temperature-sensitive selection protocol as previously described [26] . Final clones of S . pyogenes that had successfully replaced the full-length genomic RSM gene cluster with the in-frame deletion allele were screened by endpoint PCR and confirmed by Sanger DNA sequencing . The resulting strain containing the in-frame deletion allele ( ΔRSM ) was identified as MEW513 . GenBank accession numbers for the MEW123 restriction endonuclease gene hsdR , specificity subunit gene hsdS , and the methyltransferase subunit gene hsdM , are AWM59_07895 , AWM59_07900 , and AWM59_07905 , respectively . The operon was cloned by PCR using primers pJoy3_123_RSM_F and pJoy3 123 RSM R , producing an amplicon of approximately 6 kb . This fragment was inserted into plasmid pJoy3 linearized by digestion with EcoRI and SphI using the NEBuilder® HiFi DNA Assembly kit ( New England Biolabs , Inc . ) , producing plasmid pEH01 . This plasmid was transformed into electrocompetent S . pyogenes MEW513 through electroporation with conditions as described above . Chloramphenicol-resistant clones were selected and screened by endpoint PCR , with restoration of m6A methylation activity confirmed by dot blot . The resulting strain containing the plasmid encoded RSM operon for complementation ( ΔRSM/pRSM ) was identified as strain MEW552 . The GenBank accession number for the restriction-endonuclease subunit gene , hsdR , is AWM59_07895 . A fragment of the endonuclease open reading frame was cloned by PCR using primers 123_7895 F and 123_7895 R , producing an amplicon of approximately 950 bp . This fragment was inserted into plasmid pSpc18 linearized by digestion with HindIII and BamHI using the NEBuilder® HiFi DNA Assembly kit ( New England Biolabs , Inc . ) , producing plasmid pKJ19 . This plasmid was transformed into electrocompetent S . pyogenes MEW123 through electroporation with conditions as described above . Spectinomycin-resistant clones were selected and screened by endpoint PCR , with final confirmation by Sanger DNA sequencing . The resulting strain containing the spectinomycin-resistance cassette insertion disrupting the restriction endonuclease gene hsdR ( ΩRE ) was identified as strain MEW489 . The GenBank accession number for the MEW123 Mga protein , gene mga , is AWM59_08335 . An in-frame deletion allele of mga was cloned by splice-overlap extension ( SOE ) PCR [75] . The upstream region of the mga gene was cloned using primers M28 Mga 5’ SalI and M28 Mga 5’ SOE R , producing an amplicon of approximately 420 bp . The downstream region of the mga gene was cloned using primers M28 Mga 3’ BamHI and M28 Mga 3’ SOE F , producing an amplicon of approximately 410 bp . The two amplicons are mixed together as template and amplified using the outside primers M28 MGA 5’ SalI and M28 Mga 3’ BamHI , producing an amplicon of approximately 830 bp . This amplicon was subsequently digested with BamHI and SalI and ligated into the E . coli to S . pyogenes temperature-sensitive vector for allelic replacements , plasmid pJRS233 [25] , cut similarly with BamHI and SalI . The resulting plasmid of was named pIL01 , with confirmation by Sanger DNA sequencing and PCR verification . Electrocompetent cells of MEW123 were prepared and transformation with plasmid pIL01 was performed essentially as previously described [76] . Erythromycin-resistant transformants were handled according to the temperature-sensitive selection protocol as previously described [26] . Final clones of S . pyogenes that had successfully replaced the full-length genomic mga allele with the in-frame deletion allele were screened by endpoint PCR and confirmed by Sanger DNA sequencing . The resulting strain containing the in-frame deletion allele ( Δmga ) was identified as MEW480 . The GenBank accession number for the MEW123 scpA gene is AWM59_08315 . A fragment of the scpA open reading frame was cloned by PCR using primers M28 ScpA SalI F and M28 ScpA SacI R , producing an amplicon of approximately 1 . 1 kb . The amplicon was digested with SalI and SacI and ligated into plasmid pSpc18 linearized with SalI and SacI , producing plasmid pIL09 . This plasmid was transformed into electrocompetent S . pyogenes MEW123 through electroporation with conditions as described above . Spectinomycin-resistant clones were selected and screened by endpoint PCR , with final confirmation by Sanger DNA sequencing . The resulting strain containing the spectinomycin-resistance cassette insertion disrupting the scpA gene ( ΩscpA ) was identified as strain MEW380 . The GenBank accession number for the M28 protein , gene emm28 , is AWM59_08325 . A fragment of the emm28 open reading frame was cloned by PCR using primers M28 Emm Hindlll F and M28 Emm BamHl R , producing an amplicon of approximately 1 . 1 kb . This amplicon incorporated unique sites for HindIII and BamHI , and the amplicon was accordingly restriction digested and ligated into plasmid pSpc18 opened with HindIII and BamHI , producing plasmid pIL03 . This plasmid was transformed into electrocompetent S . pyogenes MEW123 through electroporation with conditions as described above . Spectinomycin-resistant clones were selected and screened by endpoint PCR , with final confirmation by Sanger DNA sequencing . The resulting strain containing the spectinomycin-resistance cassette insertion disrupting the emm28 gene ( Ωemm28 ) was identified as strain MEW409 . Transformation efficiency was assessed by electroporation of electrocompetent S . pyogenes strains with 0 . 5 μg plasmid pJoy3 conferring chloramphenicol resistance isolated from E . coli DH5α . Electroporation was performed using a Gene Pulser II system ( BioRad , Hercules , CA , USA ) under the following settings; Volts at 1 . 75 kV , capacitance at 400Ω , and resistance at 25 μf . Transformants were plated onto THY agar supplemented with chloramphenicol . In addition , a separate aliquot of the sample was plated onto THY agar with no antibiotics to determine the total viable cell count . Transformation efficiency was determined as the number of chloramphenicol resistant cells per total viable cell count . Genomic DNA was purified from S . pyogenes strains MEW123 ( WT ) and MEW513 ( ΔRSM ) using the Wizard Genomic DNA Purification Kit ( Promega , Madison , WI ) . Genomic DNA preparation , library preparation , and sequencing of MEW123 was performed as previously described [12] . For MEW513 , one Single Molecule Real-Time ( SMRT ) cell was used to sequence the library prepared with 5 kb mean insert size on the Pacific Biosciences RSII sequencer by the University of Michigan Sequencing Core ( https://brcf . medicine . umich . edu/cores/dna-sequencing ) . Modification and motif analysis was performed using RS_Modification_and_Motif_Analysis . 1 version 2 . 3 . 0 using the published MEW123 reference genome with an average reference coverage of 501 and 539 for MEW123 and MEW513 , respectively . Data generated in this analysis have been deposited in NCBI’s Gene Expression Omnibus [77] and are accessible through GEO Series accession number GSE130428 ( https://www . ncbi . nlm . nih . gov/geo/query/acc . cgi ? acc=GSE130428 ) . Genomic DNA was isolated from S . pyogenes strains MEW123 ( WT ) and MEW513 ( ΔRSM ) , as described above . DNA was treated with RNAse during purification to remove any contaminating mRNA or rRNA potentially containing m6A base modifications . DNA was denatured by heating at 98°C for 10 min and then placed immediately on ice for 5 minutes . Denatured DNA or unmodified oligonucleotides as a negative control was then spotted at 500 ng per spot onto nitrocellulose membranes and allowed to air dry . Membranes were then placed onto Whatman paper soaked with PBS containing 0 . 5% Tween 20 ( PBST ) , and DNA was cross-linked to the membranes using a Bio-Rad GS Genelinker using two 125 mJ delivery cycles . Membranes were blocked in 5% milk protein in PBS for 1 h at room temperature and then incubated with a dilution of anti-m6A primary rabbit antibody ( 2 μg/mL ) ( EMD Millipore ABE572 Anti-N6-methyladenosine ( m6A ) Antibody ) in 5% milk PBS overnight at 4°C . Primary antibody was removed and the membrane was washed three times with PBST for 5 minutes each wash . The membrane was then incubated with a 1:5 , 000 dilution of horseradish peroxidase-conjugated anti-rabbit secondary antibody in 5% milk PBS at room temperature for 1 hour . The secondary antibody was removed , and the membrane washed with PBST three times for 5 minutes each wash . Chemiluminescent substrate ( Pierce SuperSignal West Femto HRP Substrate , ThermoFisher Scientific , Waltham , MA ) was applied and the membrane was visualized . Streptococcus from fresh overnight growth on THY agar plates was inoculated into 40 mL of C-media broth and grown statically to mid-log phase OD600nm of 0 . 6 . RNA was then purified using the RiboPure RNA Purification Kit ( Life Technologies ) , for bacteria according to the manufacturer’s recommendations . The University of Michigan Sequencing Core performed ribosomal rRNA depletion using the Ribo-Zero Magnetic Kit , bacteria and subsequent library preparation . Fifty-base single end reads were sequenced on the Illumina HiSeq 4000 . Sequence alignment was performed using the Burrows-Wheeler Aligner ( BWA ) version 0 . 7 . 8-r455 to the MEW123 reference genome [12] . Subsequent differential expression analysis was performed using the limma package in R [78] . Differentially expressed genes were called as those that had a Benjamini-Hochberg adjusted p-value less than 0 . 05 and a log2 fold change greater than 1 . Log2 CPM values were computed using edgeR and were subsequently used to construct the heatmap using the aheatmap function as part of the NMF package in R [79 , 80] . Data generated in this analysis have been deposited in NCBI’s Gene Expression Omnibus [77] and are accessible through GEO Series accession number GSE130427 ( https://www . ncbi . nlm . nih . gov/geo/query/acc . cgi ? acc=GSE130427 ) . Based on results of the most significantly differentially expressed genes between WT ( MEW123 ) and ΔRSM ( MEW513 ) , three genes were selected for independent reverse-transcription cDNA preparation and real-time PCR amplification for relative comparison of transcript expression; mga , emm28 , and scpA . RNA was isolated as described above from strains grown in C-media broth to mid-log phase OD600nm of 0 . 6 . Synthesis of cDNA was performed using the iScript™ cDNA Synthesis Kit ( BioRad ) . Real time amplification of select genes was performed using an iCycler Thermocycler ( BioRad ) and iQ SYBR Green Supermix ( BioRad ) . Sequences for RT-PCR primers are as shown in Table 5 . Relative transcript levels were determined using the recA transcript as reference by the 2 ( -ΔΔCt ) method [81] . All RNA was stored at -80ºC . All cDNA was stored at -20ºC or utilized directly for comparative RT-PCR analysis . For each experiment , three biological replicates were analyzed in duplicate . Statistical significance was examined using the paired t-test in Prism 6 ( GraphPad ) . Inflammatory infection of murine subcutaneous tissue was conducted as described in detail [41] . On the day of infection , mice sedated by inhalation of isoflurane received a subcutaneous injection of 100 μl PBS containing 1 x 107 S . pyogenes into the shaved flank . Following infection , the resulting ulcers were photographed over several days and the areas of the irregular lesions were calculated using ImageJ software as described in detail elsewhere [82 , 83] . Skin biopsies were obtained from euthanized mice and homogenized in 1 mL ice cold PBS using a FastPrep-24 homogenizer ( MP Biomedicals , LLC . , Santa Ana , CA ) ; tissue was homogenized in 2 mL conical screw top vials with 3 . 2 mm stainless steel beads ( Fisher Scientific , Pittsburg , PA ) with two FastPrep cycles of speed 6 . 0 for 45 sec , with a 5 min ice incubation between pulses to prevent overheating . Samples of mouse skin and subcutaneous tissue homogenates were harvested at six-days post-infection . Cytokine protein concentrations were determined by a multiplex murine ELISA assay ( EMD Millipore , Billerica , MA , USA ) according to the manufacturer's protocol . Murine skin biopsies were obtained at six-days post-infection and were fixed in 4% formalin and dehydrated up to 70% ethanol prior to paraffin embedding through the University of Michigan Pathology Core for Animal Research ( PCAR ) . H&E staining and immunohistochemistry services were performed by the PCAR using commercially available anti-CD3 ( T lymphocytes ) , and anti-F4/80 ( macrophages ) antibodies ( Abcam , Cambridge , MA , USA ) . Digital images were obtained with an EC3 digital imaging system ( Leica Microsystems , Buffalo Grove , IL , USA ) using Leica Acquisition Software ( Leica Microsystems ) . Adjustments to contrast in digital images were applied equally to all experimental and control images . Adherence of S . pyogenes strains was assessed to an established human vaginal epithelial cell line , VK2/E6E7 , using methods similar to those previously described [84–86] . The human vaginal epithelial cell line VK2/E6E7 was purchased from the American Type Culture Collection ( ATCC , Manassas , Virginia ) , and cells were grown and maintained in media and conditions as recommended by ATCC . Human cells were grown to confluence in 12-well tissue culture-treated plates and washed with sterile PBS prior to inoculation with bacteria . S . pyogenes strains were grown in THY broth to early stationary phase ( OD600nm 0 . 6 ) , washed twice in sterile PBS , and adjusted to give an inoculum of ~5 x 106 CFU in 1 mL per well , for a multiplicity of infection ( MOI ) of ~5 . Bacteria and human cells were incubated at 37°C in 5% carbon dioxide for 60 min , after which time the supernatants were removed and cells were washed four times with 2 mL sterile PBS to remove non-adherent organisms . To recover S . pyogenes from the epithelial cells , each well was treated with 0 . 2 mL 0 . 25% Trypsin-EDTA ( Invitrogen ) and incubated at 37°C for 5 min , and then lysed by addition of 0 . 8 mL sterile water at pH 11 . Lysis in water at pH 11 was shown to result in a more complete eukaryotic cellular breakdown with maximal recovery of bacteria from the surface in addition to intracellular reservoirs [87] . This method recovers all cell-associated streptococci , predominantly extracellular adherent cells with a relatively smaller amount of intracellular cells . The cell suspension was serially diluted in PBS and plated onto THY agar for determination of viable CFU count . The total cell-associated CFU percentage was calculated as ( total CFU recovered from the well/CFU of the original input inoculum ) x 100% . Experiments were performed as previously described [55] . To synchronize estral cycles , sedated mice were estrogen supplemented by intra-peritoneal injection with 0 . 5 mg β-estradiol 17-valerate ( Sigma ) dissolved in 0 . 1 mL sterile sesame oil ( Sigma ) 2 days prior to streptococcal inoculation and again on the day of inoculation ( considered day #0 ) . On day #0 , sedated mice were inoculated with ~1 x 106 colony forming units ( CFUs ) instilled into the vaginal vault using a P20 micropipetter ( Gilson , Inc . , Middleton , WI ) in a total volume of 20 μL PBS . At successive intervals over a 1-month period post-inoculation , the vaginal vaults of sedated mice were gently washed with 50 μL sterile PBS and serial dilutions in sterile PBS were plated onto THY agar plates supplemented with 1000 μg/mL streptomycin to determine viable CFUs . This concentration of streptomycin suppressed growth of normal mouse vaginal flora but had no effect on the plating efficiency of the streptomycin-resistant S . pyogenes strains . For colonization experiments , between 5 to 20 mice were tested per S . pyogenes strain , as indicated in the relevant figure legends . Human neutrophils were purchased from a commercial supplier ( Astarte Biologics , Bothell , WA , USA ) and prepared according to supplier recommendations to a concentration of 5 x 106 cells/mL in room temperature Hank’s Balanced Salt Solution ( HBSS ) . Neutrophil bactericidal assay was performed similar to that reported by Staali et al . [45] . Briefly , streptococcal strains were grown in fresh C-media to mid-log phase ( OD600nm of 0 . 6 ) and were washed twice in HBSS with calcium and magnesium , but without Phenol Red ( Sigma , St . Louis , MO , USA ) . Streptococci were counted using a hemocytometer and adjusted to a concentration of 5 x 107 CFU/mL in room temperature HBSS . Neutrophils and streptococci were mixed in a 1:10 ratio of neutrophils to bacteria , and were incubated together for 10 minutes at 37°C . Next , extracellular streptococci were eliminated by addition of gentamicin ( 100 μg/mL ) and penicillin ( 5 μg/mL ) in HBSS for 20 minutes at 37°C . Next , cells were diluted in 1 mL of HBSS , centrifuged at 400g x 5 min , and washed with 1 mL fresh HBSS . The wash was repeated a second time in HBSS and the final cell pellet was resuspended in 50 μL of 2% saponin in distilled water at pH 11 and allowed to remain at room temperature for 20 minutes to lyse neutrophils and release viable intracellular streptococci . The cells were diluted in distilled pH 11 water and aliquots plated onto fresh THY agar media for CFU counts . Three biological replicate experiments for each strain were performed . Comparison of nonparametric data sets was performed using the Mann-Whitney U-test to determine significant differences . Differences between groups for recovery of CFU in vaginal washes were tested using a repeated measures analysis of variance . Differences in relative transcript levels were tested for significance with a two-tailed paired t-test . Differences in VK cell adherence and in neutrophil bactericidal survival assays were compared using a non-paired t-test . For all tests , the null hypothesis was rejected for P < 0 . 05 . Computation utilized the resources available in GraphPad Prism™ ( GraphPad Software , Inc . , San Diego , CA ) .
|
DNA methylation is common among many bacterial species , yet the contribution of DNA methylation to the regulation of gene expression is unclear outside of a limited number of gram-negative species . We characterized sites of DNA methylation throughout the genome of the gram-positive pathogen Streptococcus pyogenes or Group A Streptococcus . We determined that the gene products of a functional restriction modification system are responsible for genome-wide m6A . The mutant strain lacking DNA methylation showed altered gene expression compared to the parent strain , with several genes important for causing human disease down regulated . Furthermore , we showed that the mutant strain lacking DNA methylation exhibited altered virulence properties compared to the parent strain using various models of pathogenesis . The mutant strain was attenuated for both survival within human neutrophils and adherence to human epithelial cells , and was unable to suppress the host immune response in a murine subcutaneous infection model . Together , these results show that bacterial m6A contributes to differential gene expression and influences the ability of Group A Streptococcus to cause disease . DNA methylation is a conserved feature among bacteria and may represent a potential target for intervention in effort to interfere with the ability of bacteria to cause human disease .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"blood",
"cells",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"pathology",
"and",
"laboratory",
"medicine",
"gene",
"regulation",
"pathogens",
"immunology",
"microbiology",
"streptococcus",
"pyogenes",
"signs",
"and",
"symptoms",
"sequence",
"motif",
"analysis",
"epigenetics",
"dna",
"bacteria",
"dna",
"methylation",
"chromatin",
"bacterial",
"pathogens",
"neutrophils",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"white",
"blood",
"cells",
"chromosome",
"biology",
"animal",
"cells",
"medical",
"microbiology",
"bioinformatics",
"streptococcus",
"microbial",
"pathogens",
"gene",
"expression",
"lesions",
"chromatin",
"modification",
"dna",
"modification",
"biochemistry",
"diagnostic",
"medicine",
"group",
"a",
"streptococci",
"cell",
"biology",
"nucleic",
"acids",
"database",
"and",
"informatics",
"methods",
"genetics",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"organisms"
] |
2019
|
DNA methylation from a Type I restriction modification system influences gene expression and virulence in Streptococcus pyogenes
|
Memory-like CD8+ T cells expressing eomesodermin are a subset of innate T cells initially identified in a number of genetically modified mice , and also exist in wild mice and human . The acquisition of memory phenotype and function by these T cells is dependent on IL–4 produced by PLZF+ innate T cells; however , their physiologic function is still not known . Here we found that these IL-4-induced innate CD8+ T cells are critical for accelerating the control of chronic virus infection . In CIITA-transgenic mice , which have a substantial population of IL-4-induced innate CD8+ T cells , this population facilitated rapid control of viremia and induction of functional anti-viral T-cell responses during infection with chronic form of lymphocytic choriomeningitis virus . Characteristically , anti-viral innate CD8+ T cells accumulated sufficiently during early phase of infection . They produced a robust amount of IFN-γ and TNF-α with enhanced expression of a degranulation marker . Furthermore , this finding was confirmed in wild-type mice . Taken together , the results from our study show that innate CD8+ T cells works as an early defense mechanism against chronic viral infection .
Conventional T cells take on naive phenotypes when they emigrate out from the thymus , whereas innate T cells from the thymus are phenotypically of the effector/memory form [1] . Compared with conventional T cells , these innate T cells , such as natural killer T ( NKT ) cells , mucosal-associated invariant T ( MAIT ) cells and H2-M3-specific T cells , are selected by interaction with hematopoietic cells rather than thymic epithelial cells , and their development is dependent on IL–15 and the SAP ( SLAM-associated protein ) signaling pathway [1] . Moreover , most innate T cells express T cell receptors ( TCRs ) specific for MHC class Ib molecules [1 , 2] . Memory-like CD8+ T cells expressing eomesodermin ( Eomes ) are another subset of innate T cells [3] . Although this type of cells is not abundant in wild type C57BL/6 mice , they initially described in Tec-kinase-deficient mice [4 , 5] and subsequently found in the thymus of a variety of mice in which T-cell-associated genes are deficient [6–13] or CIITA-transgenic ( CIITATg ) mice in which MHC class II molecules are expressed in thymocytes [14] . Recently , a substantial number of these innate CD8+ T cells was also identified in wild-type BALB/c mice [6] and in human [14] . Eomes+ CD8+ T cells from both mice and human thymus exhibit immediate effector function upon TCR stimulation [6 , 14]; however , this type of CD8+ T cells has unique characteristics that make them different from MHC class Ib-restricted innate T cells . Firstly , common gamma chain cytokines , particularly IL–4 in this case , drive the expression of Eomes during the intrathymic developmental process [6 , 14] . Promyelocytic leukemia zinc finger protein ( PLZF ) + NKT cells are the major source of IL–4 in wild-type BALB/c and Klf2-deficient mice [6] , whereas in CIITATg mice PLZF+ T-T CD4+ T cells are responsible for the production of IL–4 [14 , 15] . In humans , IL–4 would be produced by both PLZF+ T-T CD4+ T and NKT cells [14] . Secondly , MHC class Ib-restricted innate T cells have a highly restricted TCR repertoire [16] , whereas IL-4-induced Eomes+ innate CD8+ T cells from CIITATg mice have a diverse TCR repertoire very much like conventional T cells [14] . This difference in TCR repertoire suggests that they are selected by diverse self-peptides presented by classical MHC class I molecules and raises the possibility that IL-4-induced innate CD8+ T cells perform some functions distinct from those of MHC class Ib-restricted innate T cells during a variety of immune responses . However , the biological relevance of IL-4-induced innate CD8+ T cells has not been elucidated . Although CD8+ T cells are crucial for the control or elimination of various viral infections , many viruses are able to establish a chronic infection by escaping virus-specific CD8+ T cell responses . The functional inactivation of antigen-specific CD8+ T cells through the triggering of co-inhibitory receptors such as programmed death–1 ( PD–1 ) and cytotoxic T-lymphocyte antigen–4 ( CTLA–4 ) is currently considered to be a conserved mechanism for not only maintaining viral persistence , but also for limiting immunopathology [17 , 18] . Moreover , an increase in frequency of virus-specific naïve CD8+ T-cell precursors was reported to help control initial viremia but cause differing outcomes with either clearance of wild-type chronic virus or emergence of a T-cell epitope escape mutant virus [19] . However , little is known regarding the impact that the type of CD8+ T cells present during infection has on protection against viral persistence . In the present study , we investigated the in vivo role of IL-4-induced innate CD8+ T cells in controlling initial viremia using the lymphocytic choriomeningitis virus ( LCMV ) clone 13 ( CL–13 ) chronic virus infection model . One of the most notable findings from this experiment is that IL-4-induced innate CD8+ T cells produce a robust amount of cytokines such as IFN-γ and TNF-α upon LCMV infection , resulting in the efficient control of viruses from the body and providing an effective barrier to the establishment of viral persistence .
To explore the in vivo function of IL-4-induced Eomes+ CD8+ T cells , we used CIITATg mice in which thymocytes express MHC class II molecules . As reported previously [14] , thymus of CIITATg mice contain high numbers of Eomes+ CD8+ T cells , whereas wild-type C57BL/6 mice have only a small number of these cells ( Fig 1A ) . These Eomes+ CD8+ T cells exhibited a phenotype similar to that of Eomes+ memory-like CD8+ T cells identified in other types of gene-manipulated mice [3 , 6] in that they highly express CXCR3 , CD124 ( IL-4Rα ) , CD122 ( IL-2Rβ ) and CD44 , and exhibit low expression of CD24 ( Fig 1B ) . We initially infected both CIITATg and wild-type mice with a conventional dose ( 2 x 106 PFU/mouse ) of LCMV CL–13 and found that CIITATg mice succumbed to early death , whereas wild-type mice did not ( Fig 1C ) . Histopathological analysis of LCMV CL-13-infected CIITATg mice showed edematous lungs where most of the alveolar spaces were filled with transudate ( Fig 1D ) , suggesting that the mice died due to immunopathologic tissue damage . We next infected mice with a diverse range of viral doses to determine the minimal infectious dose capable of causing chronic infection in wild-type C57BL/6 mice . We found that inoculation of LCMV CL–13 into wild-type mice established chronic virus infection with sustained expression of PD–1 molecules on CD8+ T cells ( S1A Fig ) and virus persistency in serum ( S1B Fig ) . Based on this , we challenged CIITATg and wild-type mice with this dose of LCMV CL–13 for subsequent studies and monitored the frequency of CD8+ T cells and viral titer in the blood for 28 DPI . Notably , the frequency and number of endogenous CD8+ T cells specific for the LCMV GP33-41 epitope ( GP33 ) were greatly enhanced in CIITATg mice when compared with those in wild-type mice ( Fig 2A and 2B ) . In addition , sustained expression of PD–1 on GP33-specific CD8+ T cells was not induced in CIITATg mice upon LCMV CL–13 infection , although wild-type mice exhibited strongly induced PD–1 expression patterns ( Fig 2C and 2D ) . In the CIITATg mice , the expression of CD127 , which are known to be downregulated on exhausted virus-specific CD8+ T cells [19 , 20] was significantly increased ( Fig 2E ) . This data coincides with a rapid drop in virus titer in blood of CIITATg mice , but not in that of wild-type mice ( Fig 2F ) . We next asked whether the enhanced CD8+ T-cell response was also present in peripheral tissues , particularly during the late phase of viral infection . To this end , we dissected the CD8+ T-cell response with respect to their number and cytokine responses in the spleen and lungs at 31 DPI . In this experiment , endogenous CD8+ T cells specific for the LCMV GP276-286 epitope ( GP276 ) as well as for LCMV GP33 were analyzed phenotypically and functionally . As expected , the CIITATg mice contained higher numbers of GP33- or GP276-specific CD8+ T cells in the spleen ( Fig 3A ) and these cells exhibited much lower levels of PD–1 expression on their surface ( Fig 3B ) compared with wild-type mice . As was the case in peripheral blood , a substantial population of CD127hi virus-specific memory CD8+ T cells was detected in the spleen and lung of CIITATg mice ( Fig 3C ) . A particularly important point in this experiment is the fact that CD8+ T cells from CIITATg mice showed very strong cytokine responses compared with T cells from wild-type mice , in this case IFN-γ and TNF-α release upon ex vivo restimulation with GP33 , GP276 , or pooled peptides ( Fig 3D and 3E ) . Like that of peripheral blood , this enhanced function of virus-specific CD8+ T cells was associated with decreased viral titers in the spleen and especially in the kidney , which is well known to be a life-long reservoir of chronic LCMV , of CIITATg mice compared with wild-type mice ( Fig 3F ) . Taken together with data from peripheral blood , this strongly suggests that enhanced virus-specific CD8+ T-cell responses both quantitatively and qualitatively contribute to the accelerated control of viremia in CIITATg mice . The development of IL-4-induced innate CD8+ T cells is dependent on PLZF+ T cells , such as PLZF+ innate CD4+ T cells [14] and NKT cells , as a source of IL–4 [6] . Consistent with the previous report [14] , the frequency of innate CD8+ T cells co-expressing CD44 and CXCR3 in the thymus was significantly lower in CIITATgIL-4KO mice than in CIITATg mice ( Fig 4A ) . Thus , to determine whether the enhanced anti-viral CD8+ T-cell response in CIITATg mice is dependent on IL-4-induced innate CD8+ T cells , we used CIITATgIL-4KO mice . When these mice were infected with CL–13 and virus-specific CD8+ T-cell numbers in peripheral blood were monitored until 31 DPI , the overall magnitude of these cells in CIITATgIL-4KO mice was found to be as low as that in control IL-4KO mice infected with CL–13 ( Fig 4B ) . Functionally , virus-specific CD8+ T cells in both CIITATgIL-4KO and IL-4KO mice exhibited an exhausted phenotype in terms of sustained PD–1 expression ( Fig 4C and 4D ) . In addition , the robust cytokine production exhibited in cells from CL-13-infected CIITATg mice upon ex vivo restimulation with epitope peptides was not present in CIITATgIL-4KO mice ( Fig 4E ) . This defect in CD8+ T-cell function suggests a failure of virus control in CIITATgIL-4KO mice . Indeed , virus was not cleared in these mice ( Fig 4F ) . Thus , these results imply that the enhanced CD8+ T-cell response in CIITATg mice infected with CL–13 is dependent on the existence of IL-4-induced innate CD8+ T cells . It is possible that IL–4 deficiency probably has effects not only on the loss of the IL-4-induced innate CD8+ T cells but also on the development of antibody responses , an important component in determining the outcome of virus infection . To exclude the possibility , we examined cytotoxic T lymphocyte ( CTL ) and antibody response between wild-type and IL-4KO mice during the course of CL–13 infection . The number of virus-specific CD8+ T cells and their PD–1 expression were not different between wild-type and IL-4KO mice ( S2A and S2B Fig ) . When serum level of LCMV-specific IgG was measured , the level was also similar between these mice ( S2C Fig ) . Accordingly , absolute numbers of follicular helper T ( TFH ) cells and germinal center ( GC ) B cells was not different between IL-4KO and wild-type mice ( S2D and S2E Fig ) , which both could not control viruses for a certain period ( S2F and S2G Fig ) . These data suggest that IL–4 itself does not affect the control of chronic viruses . Next , to confirm the anti-viral function of IL-4-induced innate CD8+ T cells , we generated these cells in P14 TCR transgenic mice , which express a TCR recognizing LCMV GP33 peptide presented by H-2Db molecules . To produce P14 Eomes+ CD8+ T cells , we injected a mixture of BM cells isolated from CIITATg and Thy1 . 1 P14 mice into irradiated CIITATgPIVKO mice ( Fig 5A , T-T P14 ) . In CIITA promoter type IV null ( PIVKO ) mice , MHC class II molecules are not expressed only on cortical thymic epithelial cells [21] , therefore most of CD4+ T cells are selected by MHC class II+ thymocyte-thymocyte interaction pathway in CIITATgPIVKO mice [15] . As a result , CIITATgPIVKO mice generate much higher number of PLZF+ CD4+ T cells as compared to CIITATg mice , so that these mice are able to get much higher amount of IL–4 in thymic environment . As expected , this mixed chimerism allowed most of the P14 CD8+ T cells to express Eomes ( Fig 5B ) . Consistent with a previous report [14] , Eomes+ P14 cells developed with a CD127hiCD44hi memory phenotype ( Fig 5C ) . In contrast , wild-type C57BL/6 mice that received Thy1 . 1 P14 BM cells contained only Eomes- naïve P14 cells ( Fig 5A–5C , T-E P14 ) . Eomes+ P14 cells also exhibited enhanced ability for effector cytokine production and cytotoxicity compared with the Eomes- counterpart: a higher fraction of Eomes+ cells produced IFN-γ and expressed a marker of degranulation , CD107a , upon in vitro stimulation with GP33 peptide ( Fig 5D ) . After generation of virus-specific IL-4-induced innate CD8+ T cells via mixed BM chimerism , we evaluated the anti-viral function of these innate CD8+ T cells in vivo . For this , we transferred Eomes+ or Eomes- P14 cells into congenic hosts , which were then infected with CL–13 ( Fig 6A ) . This strategy allowed us to track the response of IL-4-induced and Eomes- conventional CD8+ T cells after CL–13 infection and to judge their individual contributions to the protection against viral persistence . To examine the proliferation capability of Eomes+ or Eomes- P14 cells after CL–13 infection , each of the cells were labelled and transferred into naïve congenic mice . After 2 . 5 DPI , Eomes+ P14 cells showed a more division and a higher expression of CD44 compared to Eomes- P14 cells ( Fig 6B ) . In the spleen , a 10-fold higher number of P14 cells accumulated at an early time point ( 5 DPI ) in the mice that received Eomes+ P14 cells than in those that received Eomes- cells , although these cells were detected with similar abundance at a later time point ( 18 DPI ) ( Fig 6C ) . In the mice that received Eomes+ P14 cells , PD–1 was only transiently expressed on P14 cells at 5 DPI and was then later downregulated ( Fig 6C ) . In contrast , the high level of PD–1 expression was sustained on Eomes- P14 cells . Functionally , P14 cells from mice received Eomes+ P14 cells were superior to those from recipients of Eomes- cell , with a higher fraction of these cells capable of producing both IFN-γ and TNF-α ( Fig 6D ) . To test whether the increased CD8+ T-cell activity corresponded to better viral control we measured viral titer in the serum . Indeed , Eomes+ P14 cells also had an effect on reducing viral load during CL–13 infection ( Fig 6E ) . These results indicate that Eomes+ P14 cells provide superior support for anti-viral activity compared to Eomes- P14 cells . Interestingly , most of the responding CD8+ T cells of recipients were PD-1-negative in the mice that received Eomes+ P14 cells ( Fig 6C ) compared to those that received Eomes- P14 cells at 18 DPI . Thus , we examined the PD–1 expression on endogenous virus-specific CD8+ T cells , which are initially Eomes-negative before infection , and their cytokine production . PD–1 expression on endogenous GP276 tetramer+ CD8+ T cells was comparable to the donor Eomes+ P14 cells ( Fig 6F ) . In addition , endogenous virus-specific CD8+ T cells in the mice that received Eomes+ P14 cells were not exhausted ( Fig 6G ) . These data suggest that PD–1 expression in both donor Eomes+ P14 cells and endogenous virus-specific CD8+ T cells during the course of CL–13 infection presumably depend on antigen levels rather than intrinsic property of the Eomes+ or Eomes- responding cells . Our observation that IL-4-induced Eomes+ innate CD8+ T cells are more effective at controlling the CL–13 infection than Eomes- conventional CD8+ T cells ( Fig 6 ) raises a question regarding the underlying mechanism . As shown in Fig 5D , Eomes+ P14 cells were able to produce more effector cytokine per cell level upon antigen stimulation than Eomes- P14 cells . In addition to an elevated effector function , Eomes+ P14 cells also showed better proliferative capability ( Fig 6B ) , resulting in the higher frequency of these cells in the mice that received Eomes+ P14 than Eomes- P14 cells ( Fig 6C ) . Therefore , an enhanced quantity and quality of Eomes+ P14 cells compared to Eomes- P14 cells could contribute to the accelerated control of CL–13 infection . However , it is possible that functional and numerical differences in the two populations depend on viral titers . To address this issue , we co-transferred both Eomes+ and Eomes- P14 cells into same mice ( Fig 7A ) . At 5 DPI , we found that the frequency of Eomes+ P14 cells was significantly higher than that of Eomes- P14 cells ( Fig 7B ) . In addition , the ability to produce effector cytokines , IFN-γ and TNF-α , was better in Eomes+ P14 cells than in Eomes- P14 cells ( Fig 7C ) . These data indicate that functionality and proliferative capability of Eomes+ and Eomes- P14 cells upon antigen stimulation is intrinsically different independently of viral titer . Next , we wanted to determine whether the anti-viral effect of IL-4-induced innate CD8+ T cells also occur in a wild-type host . Unlike C57BL/6 mice , BALB/c mice have an abundant population of IL-4-induced innate CD8+ T cells , whose generation is supported by IL–4 produced by intrathymic PLZF+ NKT cells [6] . Prior to investigating anti-viral CTL responses in BALB/c strain , we compared IL–4 induced innate CD8+ T-cell phenotypes on CD8+ T cells in uninfected wild-type BALB/c and CD1dKO BALB/c , in which IL-4-induced innate CD8+ T-cell generation is defective due to the absence of NKT cells . The number of virus-specific CD8+ T cells , NP118-126 ( NP118 ) tetramer+ CD8+ T cells , was not different between wild-type and CD1dKO BALB/c mice ( Fig 8A ) . However , the NP118 tetramer+ CD8+ T cells displayed a slightly higher Eomes expression level and the population co-expressing CD44 and CXCR3 among the NP118 tetramer+ CD8+ T cells was significantly increased in wild-type BALB/c mice than in CD1dKO BALB/c mice ( Fig 8B ) . Similarly , the other CD8+ T cells than NP118 tetramer+ CD8+ T cells also contained the higher population of CD44hiCXCR3+ cells in wild-type BALB/c mice than in CD1dKO BALB/c mice ( Fig 8C ) , indicating that the frequency of pre-existing innate CD8+ T cells is higher in wild-type BALB/c than in CD1dKO BALB/c prior to infection , irrespectively of antigen-specificity of CD8+ T cells . Next , to compare anti-viral CD8+ T-cell responses in wild-type BALB/c mice with those of CD1dKO BALB/c mice , we challenged wild-type BALB/c mice and CD1dKO BALB/c mice with CL–13 ( 2 x 105 PFU/mouse ) . The viral load in peripheral blood of wild-type BALB/c mice was significantly decreased compared with that of CD1dKO hosts ( Fig 8D ) . Moreover , wild-type BALB/c mice exhibited decreased viral titer in spleen and kidney and more abundant virus-specific CD8+ T cells in spleen and lungs than CD1dKO mice ( Fig 8E and 8F ) . As was the case in CIITATg mice , PD–1 expression on virus-specific CD8+ T cells in wild-type BALB/c mice was not sustained at a later time point after viral infection in wild-type BALB/c mice ( Fig 8G ) . Furthermore , we observed that CD8+ T cells from wild-type BALB/c mice produced higher amounts of effector cytokines such as IFN-γ and TNF-α than did those from CD1dKO mice ( Fig 8H and 8I ) . These data suggest that IL-4-induced innate CD8+ T cells in a wild-type BALB/c host also contributed to reduce CL–13 viral load compared with CD1dKO mice .
Over the course of viral infection there may be a limited time period during which the host system can eliminate the virus [22] . When viruses are not eliminated within this period of time , virus-specific CD8+ T cells are exhausted via PD–1 ( programmed death 1 ) and its ligand , PD-L1 interaction , resulting in a chronic infection [23] . In this study we demonstrated that IL-4-induced innate CD8+ T cells are able to effectively control the chronic viral infection . For this , we first compared T-cell responses to chronic viral infection induced by LCMV CL–13 in CIITATg and wild-type C57BL/6 mice . The immune system of the CIITATg mouse resembles that of humans with respect to MHC class II expression in both thymic epithelial cells and thymocytes making it a suitable model [24–27] . Thus , PLZF+ T-T CD4+ T cells are generated in response to TCR signals from the MHC class II/peptide complex expressed on thymocytes [15 , 27 , 28] and provide IL–4 for the development of IL-4-induced innate CD8+ T cells [14] . When mice were infected with CL–13 , CIITATg mice were able to control viral titers below detection levels in selected tissues such as the spleen and serum within a month , whereas wild-type mice succumbed to persistent infection . Furthermore , the ability of CIITATg mice to control the virus was dependent on IL-4-induced innate CD8+ T cells as CIITATg and control mice did not show a difference in serum viral titers on the IL–4 deficient background , which causes a lack of intrathymic generation of IL-4-induced innate CD8+ T cells . Moreover , adoptive transfer of LCMV-specific IL-4-induced innate CD8+ T cells into wild-type hosts further confirmed the crucial role of these innate T cells as the primary effector mechanism for viral control . We also demonstrated that wild-type BALB/c mice , which have abundant IL-4-induced innate CD8+ T cells , exhibit notably enhanced anti-viral CTL responses compared with CD1dKO BALB/c mice , which only possess a very small fraction of these cells . As expected , expression level of Eomes and innate CD8+ T cell marker ( CD44 and CXCR3 ) was higher in virus-specific CD8+ T cells wild-type mice , as compared to those of CD1dKO BALB/c mouse ( Fig 8B ) , while we could not found any difference in LCMV NP118-specific CD8+ T cell numbers in these mice ( Fig 8A ) . Considering the results from P14 cell adoptive transfer and CL–13 infection ( Fig 6 ) , these data favor the idea that IL-4-induced innate CD8+ T cells in a wild-type BALB/c host contributed to reduce CL–13 viral load compared with CD1dKO mice , although the genetic perturbation in CD1dKO mice is not specific for the innate CD8+ T cell population . Eomes is a key transcription factor in the cytotoxic T-cell lineage [29] . During the activation and differentiation of mature CD8+ T cells Eomes induces effector function and cooperates with T-bet to sustain memory CD8+ T-cell homeostasis [30] . In particular , the central memory population is diminished in CD8+ T cells lacking Eomes [29 , 30] . Moreover , in the thymus Eomes also seems to confer effector function and memory phenotypes to innate CD8+ T cells as IL-4-induced innate CD8+ T cells acquire a CD44hiCD62Lhi central memory cell-like phenotype [14] . In addition to the previous reports , our comparison data of phenotype and function in between IL-4-induced innate CD8+ T cells and memory CD8+ T cells showed that the expression levels of Eomes and CXCR3 were similar but those of CD44 , CD124 , CD24 , and NKG2D were different ( S3A Fig ) . IFN-γ production and degranulation ability upon antigen stimulation were also different ( S3B Fig ) . Although these IL-4-induced innate CD8+ T cells are less functional than memory CD8+ T cells , their function is evidently better than those of naïve CD8+ T cells . These data demonstrate that IL-4-induced innate CD8+ T cells are phenotypically and functionally different from conventional memory CD8+ T cells as well as naïve CD8+ T cells . On the other hand , the expression of Eomes mRNA and protein are markedly elevated in exhausted CD8+ T cells during chronic virus infection compared to that in effector or memory CD8+ T cells [20 , 31] . These finding suggest that Eomes alone is not sufficient to stimulate the effector function of exhausted CD8+ T cells under the conditions of established chronic virus infection . This is despite the fact that upregulation of Eomes initially triggers the effector function of CD8+ T cells upon TCR stimulation and contributes to preserve the functionality of memory CD8+ T cells . In the present study , when CD8+ T cells already express a significantly high level of Eomes , these cells acquire obviously enhanced ability to produce effector cytokines such as IFN-γ and TNF-α upon viral antigen challenge . Taken together , these data suggest that a high level of Eomes expression allows IL-4-induced innate CD8+ T cells to exhibit their prompt and significant effector function , thereby controlling viremia during the early phase of virus infection . Many researchers have attempted to control chronic virus infection using immunotherapeutic interventions such as blockade of the inhibitory receptors PD–1 and CTLA–4 , administration of type I IFN , and regulation of microRNAs [32] [33 , 34] . Additionally , for chronic hepatitis B virus infection treatment , adoptive T-cell therapy using either in vitro expanded hepatitis B virus antigen-specific T cells or grafting T cells with recombinant TCR has been investigated as an approach [35] . Based on our data , virus-specific IL-4-induced innate CD8+ T cells have the potential to be used in adoptive T-cell therapy . Interestingly , when we adoptively transferred Eomes+ LCMV-specific innate CD8+ T cells into CL-13-infected mice , the CTL response was slightly increased even though the transferred cells were seemed still exhausted ( S4 Fig ) . From this experiment , we hypothesize that combination therapy with adoptive transfer of IL-4-induced innate CD8+ T cells for prompt control of virus and PD–1 blockade for rejuvenating CTL function could be effective . Further experiments will be required to test this theory . X-linked lymph proliferative disease ( XLP ) is a human immunodeficiency caused by germ-line mutations in SH2D1A gene and characterized by an inability to respond appropriately to infections such as Epstein-Barr virus [36] . The SH2D1A gene encodes the SAP molecule that is a component of the SLAM ( signaling lymphocytic activation molecule ) signaling pathway . Signaling through the SLAM family of receptors is crucial for the development of NKT cells , and thus the absence of the SAP causes an arrest in NKT cell development [37] [38 , 39] ) . In this context , the deficiency of NKT cells has been considered as one of the cellular bases of XLP [38] . However , considering the crucial role of SAP for T-T CD4+ T-cell development [40] , the development and function of T-T CD4+ T cells would also be expected to be defective in XLP patients [27] . Both NKT and T-T CD4+ T cells are engaged in the generation of innate CD8+ T cells via IL–4 production and thus , the development of IL-4-induced innate CD8+ T cells is also dependent on the adaptor SAP [41] . In the present study , we demonstrated that IL-4-induced innate CD8+ T cells are able to rapidly proliferate , secrete cytokines , and decrease viral load after LCMV CL–13 infection . Taken together , it is necessary to consider the possibility that defective development of IL-4-induced innate CD8+ T cells causes the heighten susceptibility to Epstein-Barr virus infection in XLP patients .
Animals were maintained and procedures were performed with approval of the IACUCs of Seoul National University ( permit number: SNUIBC-R100524-1 ) and Yonsei University ( permit number: 2013–0115 ) in accordance to LABORATORY ANIMAL ACT of Korean Ministry of Food and Drug Safety for enhancing the ethics and reliability on animal testing through appropriate administration of laboratory animals and animal testing . C57BL/6 , IL-4KO , BALB/c and BALB/c-CD1dKO mice were purchased from the Jackson Laboratory . The CIITATg mice were previously generated at Seoul National University [26] and CIITATg mice were bred to IL-4KO and PIVKO mice in our laboratory to generate CIITATgIL-4KO and CIITATgPIVKO . LCMV epitope-specific TCR transgenic P14 Thy1 . 1+ Ly5 . 2+ mice were obtained from the Emory Vaccine Center , USA and P14 Thy1 . 1+ Ly5 . 1+ mice were obtained from POSTECH , Korea . All mice were maintained in the specific pathogen-free facility of the Yonsei Laboratory Animal Research Center at Yonsei University and the Center for Animal Resource Development at Seoul National University College of Medicine ( Seoul , Korea ) . LCMV CL–13 , a variant derived from an LCMV ARM CA1371 carrier mouse [42] , was obtained from Rafi Ahmed ( Emory Vaccine Center , Atlanta ) . Six- to ten-week old mice were infected with 1 x 105 to 2 x 106 PFU of LCMV CL–13 diluted in serum-free RPMI medium per 20 g of mouse body weight by intravenous infection or with 2 x 105 PFU of LCMV Armstrong diluted in serum-free RPMI medium per 20 g of mouse body weight by intraperitoneal infection . For serum virus titration , three to four drops of blood were individually collected by microcapillary tube at the indicated time points post infection , and the serum was directly stored at -70°C . For tissue titration , small pieces of the spleen and kidney were put in DMEM containing 1% FBS ( HyClone ) and stored at -70°C . The tissues were later homogenized completely using a homogenizer ( Kinematica ) before titration . Viral titers from sera or homogenized samples were determined by plaque assay on Vero cells as previously described [43] . Undetectable samples were given a half of each detection limit . Peripheral blood mononuclear cells ( PBMCs ) were isolated from peripheral blood using density gradient centrifugation underlaid with Histopaque–1077 ( Sigma-Aldrich ) . Lymphocytes from the thymus , spleen and lung were isolated as previously described [23] . For phenotypic analysis of lymphocytes , single-cell suspensions were stained with the following antibodies; fluorochrome-conjugated antibodies against CD4 ( RM4-5 ) , CD127 ( A7R34 ) , and PD–1 ( RMP1-13 ) were from BioLegend , antibodies against CD8 ( 53–6 . 7 ) , CD24 ( 30-F1 ) , CD44 ( IM7 ) , CXCR3 ( CXCR3-173 ) , and CD19 ( eBio1D3 ) were from eBioscience , and antibodies against CD44 ( IM7 ) , CD90 . 1 ( Thy1 . 1; OX–1 ) , CD122 ( IL-2Rβ; TM-b1 ) , CD124 ( IL-4Rα; mIL4R-M1 ) , CXCR5 ( 2G8 ) , CD45R/B220 ( RA3-6B2 ) , CD95 ( Fas; Jo2 ) , T- and B-cell activation antigen ( GL7 ) , and NKG2D ( CX5 ) were from BD Biosciences . H-2Db tetramers bound to GP33-41 or GP276-286 peptide and H-2Ld tetramer bound to NP118-126 peptide were generated and used as previously described [44] . To detect cytokine production by virus-specific CD8+ T cells , splenocytes from C57BL/6 mice were restimulated in vitro with 0 . 2 μg/ml of LCMV GP33-41 , GP276-286 , or peptide pool including GP33-41 , GP276-286 , GP70-77 , GP92-101 , NP166-175 , NP205-212 , NP235-249 , and NP396-404 in the presence of Golgi plug/Golgi stop ( BD Biosciences ) and anti-CD107a ( 1D4B , BD Biosciences ) Ab for 5 hours followed by intracellular cytokine staining using anti-IFN-γ ( XMG1 . 2 , BD Biosciences ) and anti-TNF-α ( MP6-XT22 , BioLegend ) antibodies . In case of in vitro restimulation of splenocytes from BALB/c mice , LCMV NP118-126 , GP283-291 , or NP313-322 peptide was used . For IL-4-induced innate CD8+ T-cell staining , lymphocytes were fixed and permeabilized with Fixation/Permeabilization solution ( eBioscience ) and stained with anti-Eomes ( Dan11mag , eBioscience ) Ab . The Live/Dead fixable Stain Kit ( Invitrogen ) was used to remove the dead cell population in most staining procedures . All stained samples were read by FACS CANTO II or LSR II ( BD Biosciences ) , and analyzed by FlowJo software ( Tree Star ) . Lungs were fixed in 10% neutral buffered formalin ( Sigma-Aldrich ) , and paraffin-embedded tissues were sectioned to a thickness of 4 μm and stained with hematoxylin and eosin . Microscopic observations were performed with an ECLIPSE 80i Bright-Field Microscope Set ( Nikon ) equipped with CFI 10×/22 eyepiece , Plan Fluor objectives ( with 4× , 10× , 20× , and 100× objectives ) and DS-Fi1 camera . We used NIS-Elements BR 3 . 1 software ( Nikon ) for image acquisition . LCMV-specific antibodies were measured by ELISA . LCMV CL-13-infected BHK–21 cell lysate was used as capture antigen . Ninety-six-well Polysorp plates ( Nunc ) were coated with sonicated lysate for 3 days before performing the ELISA . Three fold serial dilutions of serum samples were incubated and detected with IgG-specific horseradish peroxidase ( HRP ) -conjugated goat anti-mouse immunoglobulin ( Southern Biotech ) . TMB ( Sigma-Aldrich ) was used as a substrate , and the reaction was stopped by sulfuric acid and read at 450nm . BM cells were isolated from femurs and tibias of donor mice and T cells were depleted through magnetic sorting using a mixture of CD4 and CD8 depleting microbeads ( Miltenyi Biotech , Auburn , CA ) . Then , 3 x 106 T cell-depleted BM cells were intravenous injected into each recipient mouse preconditioned with 13 , 000 rad irradiation ( two doses of 650 rad applied 4 h apart ) from a 137Cs source 1 day prior . For CD8+ T cell enrichment , CD8+ T cells were isolated from the spleen of wild-type BALB/c and CD1dKO BALB/c mice after negative selection using a CD8+ T-cell isolation kit ( Miltenyi Biotech ) . For adoptive transfer of P14 Thy1 . 1+ Ly5 . 2+ CD8+ T cells and P14 Thy1 . 1+ Ly5 . 1+ CD8+ T cells , the cells were isolated from the spleen of BMT chimera mice after negative selection using a CD8+ T-cell isolation kit and subsequent positive selection using Thy1 . 1+ cell microbeads ( Miltenyi Biotech ) . After isolation , 5 x 103 purified P14 Thy1 . 1+ CD8+ T cells were transferred alone or with P14 Ly5 . 1+ CD8+ T cells into wild-type C57BL/6 mice via tail vein . Mice were infected with LCMV CL–13 at 1d after the adoptive transfer . LCMV GP33-41-specific P14 CD8+ T cells were isolated from P14 transgenic mice using CD8+ isolation kit ( Myltenyi Biotec ) . Purified P14 CD8+ T cells were labelled with CellTrace Violet ( CTV ) proliferation kit at concentration of 10uM ( Invitrogen ) . Labelled P14 CD8+ T cells ( 1 x 106 cells ) were adoptively transferred into naïve mice . Mice were infected with LCMV CL–13 at 1d after the adoptive transfer . Thy1 . 2+ congenic C57BL/6 mice were infected with 2 x 106 PFU of LCMV CL–13 . At 3 weeks post infection , P14 Thy1 . 1+ CD8+ T cells were isolated from the spleen of BMT chimera mice for adoptive transfer after negative selection using a CD8+ T-cell isolation kit and subsequent positive selection using Thy1 . 1+ cell microbeads ( Miltenyi Biotech ) . After isolation , 2 x 106 purified P14 Thy1 . 1+ CD8+ T cells were transferred into CL-13-infected C57BL/6 mice via tail vein . Survival curve was evaluated by log-rank ( Mantel-Cox ) test and other data were analyzed by two-tailed unpaired Student’s t-test using GraphPad Prism software . A p value less than 0 . 05 was considered statistically significant .
|
Over the course of viral infection there may be a limited time period during which the host system can eliminate the virus . When viruses are not eliminated within this period of time , virus can establish persistent infection . Here , we show that IL-4-induced innate CD8+ T cells are able to effectively control chronic virus infection . Innate T cells are heterogeneous population of T cells that acquire effector/memory phenotype as a result of their maturation process in thymus , unlike conventional T cells that differentiate into memory cells after antigen encounter in periphery . Previous data suggest that innate T cells might serve as a first-line of defense against certain bacterial pathogens . IL-4-induced innate CD8+ T cells are a unique subset of innate T cells that were recently identified in both mouse and human . We found that IL-4-induced innate CD8+ T cells immediately accumulated after viral infection and produced a robust amount of effector cytokines . Thereby , IL-4-induced innate CD8+ T cells provide an effective barrier to the establishment of persistent infection via effective virus control during the early phase of viral infection . Collectively our data show that IL-4-induced innate CD8+ T cells works as an early defense mechanism against chronic viral infection .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
IL-4 Induced Innate CD8+ T Cells Control Persistent Viral Infection
|
Viruses within a family often vary in their cellular tropism and pathogenicity . In many cases , these variations are due to viruses switching their specificity from one cell surface receptor to another . The structural requirements that underlie such receptor switching are not well understood especially for carbohydrate-binding viruses , as methods capable of structure-specificity studies are only relatively recently being developed for carbohydrates . We have characterized the receptor specificity , structure and infectivity of the human polyomavirus BKPyV , the causative agent of polyomavirus-associated nephropathy , and uncover a molecular switch for binding different carbohydrate receptors . We show that the b-series gangliosides GD3 , GD2 , GD1b and GT1b all can serve as receptors for BKPyV . The crystal structure of the BKPyV capsid protein VP1 in complex with GD3 reveals contacts with two sialic acid moieties in the receptor , providing a basis for the observed specificity . Comparison with the structure of simian virus 40 ( SV40 ) VP1 bound to ganglioside GM1 identifies the amino acid at position 68 as a determinant of specificity . Mutation of this residue from lysine in BKPyV to serine in SV40 switches the receptor specificity of BKPyV from GD3 to GM1 both in vitro and in cell culture . Our findings highlight the plasticity of viral receptor binding sites and form a template to retarget viruses to different receptors and cell types .
Interactions of a virus with receptors on host cells are crucial for viral entry and infection , and determine host range and tissue tropism of the virus . As a result , receptor specificity and affinity are tightly regulated , with changes in either producing a virus with different spread and infectivity . Many zoonotic transmissions are based on a virus acquiring binding capability for a new receptor . Despite their importance , most viral specificity switches are poorly characterized , especially for carbohydrate-binding viruses , as methods capable of structure-specificity studies are only being developed for carbohydrates . Here we characterize the receptor specificity of the human BK Polyomavirus in depth , and retarget it to use the Simian Virus 40 ( SV40 ) receptor GM1 . The human polyomavirus BK Virus ( BKPyV ) is a non-enveloped , double-stranded DNA ( dsDNA ) virus that belongs to the family Polyomaviridae . Other members of the family include Simian Virus 40 ( SV40 ) , the human JC Virus ( JCPyV ) , Merkel Cell Polyomavirus ( MCPyV ) and at least eight other recently discovered human polyomaviruses [1] . BKPyV was first isolated from a kidney transplant recipient in 1971 [2] . It establishes a persistent asymptomatic infection in the genitourinary tract of approximately 70% of the adult population [3] , [4] , [5] . A key modulator of BKPyV reactivation is immunosuppression of the host that leads to an increase in viral replication [3] . Complications of BKPyV reactivation include the development of polyomavirus-induced nephropathy ( PVN ) in kidney transplant recipients , and hemorrhagic cystitis in bone marrow transplant recipients [3] , [6] , [7] . BKPyV attachment is mediated by cell-surface sialic acid [8] . The most common sialic acid type in humans is 5-N-acetyl neuraminic acid ( NeuNAc ) [9] . Simians and most other mammals , however , possess an enzyme that can attach an additional hydroxyl group to NeuNAc , yielding 5-N-glycolyl neuraminic acid ( NeuNGc ) . In contrast to humans , these animals therefore carry both NeuNAc and NeuNGc . Gangliosides were found to mediate cell attachment of BKPyV [10] , and gangliosides GD1b and GT1b were later shown to function as specific receptors for BKPyV [11] ( Fig . 1 ) . Gangliosides are ceramide-based glycolipids , which are used as receptors for most of the well-characterized polyomaviruses , for example GD1a and GT1b for Polyoma , or GM1 for SV40 [12] . A polyomavirus capsid consists of 72 pentamers of the major capsid protein VP1 [13] , [14] . Crystal structures of Polyoma , SV40 , MCPyV and JCPyV VP1 show that receptors are bound in shallow grooves formed by VP1 loop structures on the outer surface of the virus . These loops contribute to different receptor specificities and are the only parts of VP1 that are not well conserved [14] , [15] , [16] , [17] , [18] . In this study , we use viral infection assays to demonstrate that the b-series gangliosides GD3 , GD2 , GD1b and GT1b enhance BKPyV infection . We then define the common α2 , 8-disialic acid motif on these gangliosides as the primary binding epitope for BKPyV by NMR spectroscopy . In order to understand how the disialic acid motif is recognized by the virus , we solved the crystal structure of a BKPyV VP1 pentamer in complex with GD3 and generated a model of the pentamer with the larger GD1b oligosaccharide . Analysis of these complexes reveals extensive interaction with the terminal sialic acid , specificity-defining contacts with the internal sialic acid , thus explaining the requirement for a disialic acid motif , and additional contacts with the branched GD1b . Mutagenesis of residues contacting the disialic acid motif abolishes infectivity and a comparison with the SV40 VP1-GM1 complex attributes the different viral receptor specificities to one point mutation . Introduction of this mutation into BKPyV switches specificity , enabling BKPyV to bind GM1 and abolishes binding to GD3 , as shown by saturation transfer difference ( STD ) NMR spectroscopy and carbohydrate microarray analyses . The microarray analyses moreover reveal that the mutant is specific for the ‘human’ sialic acid NeuNAc . This contrasts with SV40 VP1 , which has a preference for the more prevalent NeuNGc found in simian species and many other nonhuman mammals . The specificity of the mutant thus sheds light on the influence of sialic acid on species tropism .
To date , only four gangliosides had been tested to support BKPyV infection , two of which , the b-series gangliosides GD1b and GT1b , were confirmed as receptors [11] . The carbohydrate moieties of gangliosides typically consist of two “arms” ( Fig . 1A ) . In this manuscript , we number the carbohydrates sequentially ( starting from the lipid anchor ) and use “L” and “R” designations to indicate whether a carbohydrate is part of the left or the right arm , respectively . For example , NeuNAc 3R is the third carbohydrate and located in the right arm ( Fig . 1A , schematic on the left hand side ) . The right arm consists of only sialic acids and is used to classify gangliosides . The b-series gangliosides , e . g . GD1b and GT1b , carry both NeuNAc 3R and NeuNAc 4R , while a-series gangliosides such as GM1 carry only NeuNAc 3R ( Fig . 1A ) . We tested the effects of all common b-series gangliosides on BKPyV infection by supplementing permissive Vero cells with GD3 , GD2 , GD1b or GT1b ( Fig . 1B ) . Consistent with previous reports [11] , gangliosides GD1b and GT1b enhanced infectivity of Vero cells . However , gangliosides GD2 and GD3 , which had not been tested previously , also enhanced infection of the cells ( Fig . 1B ) . Incorporation of all b-series gangliosides into the plasma membrane of Vero cells increased the binding of labeled VP1 pentamers to cells ( data not shown ) . In control experiments , supplementing cells with the a-series ganglioside GM1 had no effect on infection or binding ( Fig . 1B , and data not shown ) . The ability of b-series gangliosides to enhance BKPyV infection was greater in the presence of the left arm , with GD1b and GT1b supporting infection best ( Fig . 1B ) . Taken together , these data show that the α2 , 8-disialic acid motif in b-series gangliosides is the minimal requirement for binding , with the left arm of GD1b and GT1b contributing some additional interactions . We analyzed pentamer binding to GD3 and GD1b oligosaccharides by STD NMR spectroscopy [19] , which identifies ligand atoms that contact a protein in solution . The strongest saturation transfer from BKPyV VP1 to GD3 was observed for the methyl group of the terminal NeuNAc 4R , followed by the methyl group of the internal NeuNAc 3R ( Fig . 1C ) . No significant transfer was observed to any of the anomeric protons , or to the NeuNAc H3 protons . Interestingly , no transfer was observed for the Glc and Gal residues of GD3 ( Fig . 1C ) , suggesting that they may not participate in binding . We repeated the same experiment for GD1b oligosaccharide , observing transfer to essentially the same set of protons from the disialyl moiety plus additional transfer to Gal 4L and GalNAc 3L in the GD1b left arm as well as the branching Gal 2 residue ( Fig . 1D ) . Resonances H4 and H1 from Gal 2 and H1 from GalNAc 3L can be unambiguously assigned , but H3 from Gal 2 and H4 from GalNAc 3L overlap and cannot be distinguished . From the Gal 4L ring , only the anomeric proton can be assigned in the STD difference spectrum . The STD spectra thus show that while the right arms of both GD3 and GD1b interact with BKPyV VP1 in a similar way , additional contacts are provided by the left arm of GD1b . To define the structural features underlying the receptor-binding specificity of BKPyV , we solved the structure of the BKPyV VP1 pentamer at 2 . 0 Å resolution ( Table 1 ) . The VP1 pentamer is a doughnut-shaped ring , with the five monomers arranged around a central pore that aligns with the five-fold symmetry axis ( Fig . 2A ) . The monomers adopt a β-sandwich fold with jelly-roll topology that is present in many viral capsid proteins . The β-strands B , I , D , G and C , H , E , F ( designated alphabetically from the N-terminus of the full-length protein ) are linked by extensive loops that decorate the surface of the protein . For clarity , the long BC-loop is subdivided into loops BC1 and BC2 , which face in different directions . We incubated BKPyV VP1 crystals in 20 mM GD3 oligosaccharide solution and solved the structure of the resulting complex at 1 . 7 Å resolution ( Table 1 ) . Attempts to form a complex with GD1b oligosaccharide by soaking or co-crystallization failed in several crystal forms , likely because of the size of the GD1b hexasaccharide . However , the GD3 structure encompasses the minimal motif required for binding and provides a template for understanding interactions with the larger GD1b oligosaccharide . The structures of unliganded and liganded BKPyV VP1 are virtually identical ( r . m . s . d . of 0 . 4 Å for all atoms of one monomer ) , indicating that GD3 binding does not induce a conformational change in the protein . GD3 engages BKPyV VP1 at the top of the pentamer , which corresponds to the outer surface of the virion ( Fig . 2A , B ) . Contacts involve residues in the BC1- , HI- and DE-loops of one monomer , as well as the BC2-loop of the clockwise neighboring VP1 monomer ( BC2cw ) and the DE-loop of the counterclockwise neighbor ( DEccw ) . Four of the five binding sites within one pentamer are occupied by ligand , while one is inaccessible due to crystal packing . Two of the four occupied binding sites , however , participate in crystal contacts . As their conformations are influenced by these non-physiologic interactions , they will not be considered further . The two remaining GD3 oligosaccharides do not participate in crystal contacts and assume an identical conformation , which therefore should represent a physiologically relevant complex . In both cases , only the terminal NeuNAc-α2 , 8-NeuNAc motif interacts with BKPyV VP1 , while the Gal-Glc moiety projects into solution . This is in agreement with the STD NMR data that showed little saturation transfer between BKPyV VP1 and the Gal-Glc portion of GD3 . The terminal NeuNAc 4R is the main contact of GD3 with BKPyV VP1 . In all four occupied binding sites on the BKPyV VP1 pentamer , NeuNAc 4R adopts the same conformation and makes identical interactions with the protein . The sialic acid carboxylate group is recognized by two hydrogen bonds to the side chains of S274 and T276 ( Fig . 2C ) . Additional , water-mediated hydrogen bonds are formed to the side chain of S273 and the backbone of S274 . The O4 hydroxyl group of NeuNAc 4R interacts via water-mediated hydrogen bonds with the side chain of N272 , the backbone of G131 , and the backbone nitrogen of F75cw . The N-acetyl group makes a hydrogen bond to N272 and a water-mediated hydrogen bond to R169cw . Its methyl group inserts into a tight-fitting , hydrophobic pocket on BKPyV VP1 that is formed by four non-polar residues ( L62 , F65 , F270 and F75cw ) . Finally , the glycerol chain of NeuNAc lies in a shallow groove and makes van der Waals contacts to the side chains of P58 , L62 , L67 , K68 and Q278 . The glycerol chain also forms a single hydrogen bond to the K68 backbone nitrogen . The conformation of the NeuNAc 4R ring is stabilized by an intramolecular hydrogen bond from the carboxylate group to the O8 hydroxyl group . The second sialic acid of GD3 , NeuNAc 3R , has weaker electron density and makes fewer contacts with the protein ( Fig . 2C ) . Its carboxylate group forms a salt bridge with the K68 side chain . The methyl group of its N-acetyl chain stacks against a hydrophobic surface created by parts of the side chains of H138ccw , S274 and T276 . The prominent role of van der Waals interactions with the methyl group mirrors the NMR results , which feature prominent saturation transfer to the methyl groups of both NeuNAcs ( Fig . 1C ) . The BKV-GD3 complex structure enabled us to model the interaction of BKV VP1 with the longer GD1b oligosaccharide , using the tightly bound terminal NeuNAc 4R as an anchor . A large number of possible GD1b oligosaccharide conformations was calculated and superposed on the terminal NeuNAc 4R in the BKV-GD3 complex structure . The oligosaccharide conformations were filtered for presence of the specificity-defining contacts between the protein and the internal NeuNAc 3R . The remaining conformations could be classified into two groups: one in which the left arm of GD1b pointed away from the protein , not engaging in interactions , and one in which this arm made additional contacts with BKV VP1 . Some of the latter conformations were then subjected to a final round of molecular dynamics ( MD ) simulations in explicit water . We found that the ‘left’ arm is involved in several weaker interactions with amino acids P58 , D59 , L67 , K68 , E81 , and K83 . A snapshot of the complex is shown in ( Fig . 2D , E ) . According to this model , the Gal 4L residue of GD1b can adopt a position enabling hydrogen bonds between its 2- and 3-hydroxyl groups and the side chain of E81 , as well as hydrogen bonds between its 6-hydroxyl group and the side chain of K83 and the backbone carbonyl of P58 . Moreover , the left arm of GD1b is supported by van der Waals interactions with the side chain of L67 , and there is an intramolecular hydrogen bond between the carboxylate group of the internal NeuNAc 3R and the N-acetyl group of GalNAc 3L . The model is in accord with the observed increase in BKV infection with increasing length of the left arm of b-series gangliosides , and also with the STD NMR data that show signals for some protons in the left arm . To test the biological relevance of these interactions mutations were introduced into an infectious clone of BKPyV . We first probed the interaction with the tightly bound terminal NeuNAc 4R ( Fig . 3A ) with mutations designed to abolish carbohydrate binding either by reducing the number of hydrogen bonds ( S274A , T276A , and S274A/T276A ) , eliminating van der Waals contacts ( F75V ) , or by introducing steric hindrance ( L62W , F75W ) . Vero cells were transfected with mutant or wild-type ( WT ) BKPyV plasmid DNA . Viral gene expression was scored every 3 days over a 22 day growth period . The first data point after transfection indicated no difference between wild-type and the mutants in terms of protein expression and localization ( data not shown ) . While WT BKPyV resulted in viral production that continued to spread with time , all mutants that targeted the binding site for terminal sialic acid did not propagate , highlighting the importance of these interactions ( Fig . 3B ) . We then targeted the binding site of the internal NeuNAc 3R . Mutant H138A , in which a van der Waals contact is removed from the second sialic acid , propagated at a significantly reduced level compared to WT ( Fig . 3B ) . We also introduced the mutations in our recombinant pentamer construct . Flow cytometry binding assays using WT and mutant pentamers show that the mutants have greatly reduced cell binding ( Fig . 3C ) , suggesting that the loss of viral propagation is due to an attachment defect . The structural integrity of mutant pentamers was verified with circular dichroism spectroscopy , and their ability to assemble into pentamers was confirmed with gel filtration ( data not shown ) . All the mutations were in the receptor binding site , which is distant from the sites important for capsid assembly . Thus , the mutations are very unlikely to cause defects in capsid assembly . Finally , we mutated residues in the putative binding site for the left arm of b-series gangliosides ( Fig . 3D ) . Again , we designed mutations to either introduce steric hindrance ( D59Y and L67W ) or to remove contacts ( D59A , L67A and K83A ) . The mutations that created steric hindrance significantly reduced BKPyV spread in culture . The D59A mutation had no significant effect , but L67A and K83A both reduced BKPyV growth . In addition , the E81A mutant , which also removes a contact from the second arm , was described in an earlier paper to have slightly reduced growth [20] . The mutations likely did not abolish growth altogether because they still permit interactions with the primary contact , terminal sialic acid . While the D59Y and L67 mutations might in theory also interfere with primary sialic acid binding due to the branched nature of GD1b , the K83A and E81A mutations certainly only target the second arm . Thus , mutagenesis confirms our structural model and highlights the importance of specific contacts with the second arm of GD1b . Taken together , our biological data confirm that the binding site for terminal sialic acid is indispensable for viral infection , while peripheral interactions further enhance binding and infection . BKPyV is most closely related to SV40 and JCPyV , with amino acid identity among their VP1 proteins as high as 74% . Nevertheless , the three viruses recognize different sialic acid containing receptors . BKPyV interacts with α2 , 8-linked b-series gangliosides , while SV40 binds the branched α2 , 3-linked GM1 ganglioside [12] and JCPyV attaches to the linear α2 , 6-linked sequence in LSTc [17] . In receptor complexes of all three viruses , the terminal sialic acid engages in critical and highly conserved interactions that anchor the ligand to VP1 ( Fig . 4 ) [16] , [17] . Specificity for the three different oligosaccharide receptor sequences arises in each case from a small number of unique contacts outside the sialic acid binding site . JCPyV recognizes an L-shaped conformation of the LSTc oligosaccharide . The key residue that makes contacts to both legs of this L-shaped glycan is N123 [17] ( Fig . 4C ) . BKPyV ( Fig . 4A ) and SV40 ( Fig . 4B ) both have a glycine at the equivalent position and thus cannot form similar contacts . BKPyV VP1 specificity for the α2 , 8-disialyl motif can be attributed to residues K68 and H138 , which form contacts with the internal NeuNAc 3R . These two residues are not conserved in SV40 or JCPyV , and thus neither virus is able to specifically interact with α2 , 8-disialic acid carrying glycans in the same manner [16] , [17] . Moreover , none of the BKPyV residues contacting the left arm of gangliosides are conserved in either SV40 or JCPyV . Although the SV40 receptor , GM1 , resembles GD1b with an identical left arm , BKPyV and SV40 VP1 bind the left arm at different sites on the proteins ( Fig . 4A , B ) . The two pockets recognize the left arm in different ways . The GalNAc 3L methyl group of GM1 in the SV40 complex is bound at a similar position on VP1 as the NeuNAc 3R methyl group in the BKPyV complex . BKPyV cannot bind the left arm of GM1 in the orientation seen in the SV40 complex because the binding site is blocked by the large side chain of K68 , which would lead to clashes with GalNAc ( Fig . 4A , B ) . Apart from this difference , BKPyV and SV40 VP1 share similar surface features at the SV40 left arm binding site and display the same main chain conformation in their surface loops . Thus , the inability of BKPyV to bind to GM1 appears to be determined by the amino acid at position 68 . To validate the conclusions derived from the structural comparisons , we introduced a K68S mutation into the BKPyV VP1 pentamer expression construct . Purified K68S pentamers were analyzed by STD NMR for binding to GD3 and GM1 . Unlike the WT BKPyV-GD3 pair , almost no saturation transfer was observed for BKPyV K68S and GD3 , indicating that the mutation virtually abolished binding to the disialic acid motif of GD3 ( Fig . 5A ) . However , saturation transfer from BKPyV K68S VP1 to GM1 was as efficient as for the SV40 VP1-GM1 pair , which was included for comparison ( Fig . 5B–D ) . This indicates that the K68S mutation switches the binding preference of BKPyV VP1 from GD3 to GM1 . The STD NMR spectra of SV40 VP1 and BKPyV K68S VP1 with GM1 are almost indistinguishable , suggesting that GM1 engages in the same contacts with both proteins . Saturation transfer is primarily observed to protons of the NeuNAc 3R and Gal 4L rings . In addition , both the GalNAc and the NeuNAc methyl groups in GM1 received considerable saturation in the complexes , with the NeuNAc methyl group being more affected . Our observations are in good agreement with the crystal structure of the SV40 VP1-GM1 complex [16] and demonstrate that a single amino acid mutation suffices for BKPyV to adapt to the SV40 receptor . BKPyV , JCPyV and SV40 differ in one aspect of their sialic acid binding site . The cavity engaging the methyl group is tight-fitting and lined with hydrophobic residues in BKPyV and JCPyV , but significantly enlarged and partially hydrophilic in SV40 ( Fig . 4A–C ) . This difference may reflect the different hosts of these viruses , humans and simians , and the different types of sialic acids characteristic for each host . While the most prominent sialic acid in humans is NeuNAc , simians carry in addition to NeuNAc larger amounts of N-glycolyl neuraminic acid ( NeuNGc ) , in which the methyl group is replaced by the bigger and more hydrophilic glycolyl ( CH2-OH ) group [9] , [21] . SV40 preferentially binds to NeuNGc-GM1 , and the glycolyl group likely engages polar residues in the cavity [16] , [22] . By contrast , the smaller and more hydrophobic cavity of BKPyV and JCPyV cannot accommodate the glycolyl group in a similar manner , thus making BKPyV and JCPyV specific for NeuNAc . To assess their specificity for human-type and simian-type sialic acids , the VP1 proteins of WT BKPyV , mutant BKPyV K68S , and SV40 were analyzed using a focused ganglioside microarray comprised of 21 ganglioside-related saccharide probes , which included the b-series gangliosides as well as GM1 variants NeuNAc-GM1 and NeuNGc-GM1 ( Supplemental Table S1 ) . Microarray analyses revealed differing binding specificities of the three VP1 proteins ( Fig . 5E–G ) . With the WT BKPyV , the only detectable binding was to the b-series ganglioside GD1b ( position 16 ) and the signal intensity was relatively low . No binding to any other b-series gangliosides GD3 , GD2 , GT1b or GQ1b was detected , possibly due the lower binding avidity to these probes compared to GD1b in the array ( Fig . 5E ) . The BKPyV K68S mutant showed barely detectable binding to GD1b , but highly specific and strong binding to the two NeuNAc-GM1 probes ( positions 10 and 11 ) , which differed only in the composition of their lipid moieties ( Fig . 5F ) . Interestingly , there was no binding to the simian-type NeuNGc-GM1 probes , in contrast to SV40 VP1 , which showed preferential binding to the two NeuNGc-GM1 probes ( positions 12 and 13 ) ( Fig . 5G ) . This finding is in accord with earlier observations and consistent with our structural analysis ( Fig . 4 ) . We next tested whether BKPyV K68S was able to use human-type NeuNAc-GM1 to attach to cells . Purified K68S or WT BKPyV VP1 pentamers were incubated with simian ( Vero ) and human ( HEK ) cells , and binding was detected by flow cytometry . In both Vero and HEK cells , K68S mutant pentamers had reduced binding compared to WT pentamers , reflecting either lower affinity or a lower number of receptor molecules ( Fig . 6A , C ) . There was no significant change in the binding of K68S or WT VP1 pentamers to Vero cells that were supplemented with 3 µM NeuNAc-GM1 prior to incubation with the pentamers ( Fig . 6A ) . However , binding of K68S VP1 to HEK cells was increased upon supplementation with GM1 , whereas WT binding levels were unchanged ( Fig . 6C ) . This finding might be linked to the enzyme CMP-sialic acid hydroxylase , which converts NeuNAc to NeuNGc and is present on simian but not human cells [9] . We performed a competitive binding assay with GM1 treated HEK cells in the presence and absence of Cholera toxin subunit B ( CTX ) , which uses GM1 as a receptor [23] . CTX abolished binding of K68S VP1 , confirming that the K68S mutant is in fact retargeted to GM1 ( Fig . 6C ) . CTX had no effect on the binding of WT BKPyV VP1 ( Fig . 6C ) . The K68S mutation was also assayed in long-term viral growth assays . We found that while transfection of WT BKPyV plasmid into Vero cells resulted in viral propagation and spread , transfection with the K68S plasmid failed to propagate ( Fig . 6B ) . In human cells however the K68S mutant spread as efficiently as WT regardless of supplementation with GM1 ( Fig . 6D ) .
In this structure-function study , we investigated the interaction of BKPyV with its glycan receptors and identified key determinants of specificity . We show that the conserved α2 , 8-disialic acid motif on the right arm of b-series gangliosides is the minimal binding epitope for BKPyV , with the variable left arm contributing some additional contacts . Point mutations in the receptor binding site abolish viral spread and infectivity , demonstrating the physiological relevance of the observed interactions . Our data demonstrate that all of the b-series gangliosides tested support BKPyV infection . As attachment likely requires multiple interactions , the virus is predicted to engage a mixture of gangliosides on the cell surface , depending on lipid composition . While gangliosides are likely entry receptors for BKPyV , the main binding epitope of BKPyV , α2 , 8-disialic acid , is not only present on gangliosides , but also on glycoproteins . It has been shown for another ganglioside-binding polyomavirus that such sialylated glycoproteins act as decoy receptors [24] . The additional contacts with the left arm of b-series gangliosides therefore may increase BKPyV binding affinity for ganglioside ligands and distinguish those from glycoproteins , which likely would lead the virus along non-infectious entry pathways . The importance of b-series gangliosides for BKPyV infection may have implications for BKPyV tropism and pathogenesis . Biochemical analyses indicate that the kidney , where BKPyV persists , is rich in diverse sphingolipids and particularly gangliosides . The most abundant gangliosides in adult human kidney are GM3 and GD3 , but small amounts of more complex gangliosides were also detected [25] , [26] . The relative abundance of simple gangliosides differs between the kidney and the brain , where complex gangliosides are most abundant in adults [27] . Therefore , the differences in affinity toward b-series gangliosides are only one determinant of their usage as receptors in vivo , as a lower affinity can be balanced by a greater abundance in the host tissue . Moreover , gangliosides are differentially expressed in cortical tubular , medullary and glomerular tissues of adult human kidney and developmental changes in ganglioside expression have been observed in bovine kidney [26] . Thus , our observation of differing affinities toward b-series gangliosides raises the question whether developmental or drug-induced changes in ganglioside distribution may play a role in BKPyV latency and reactivation . Structural comparison of BKPyV-GD3 with the closely related SV40-GM1 complex suggested that a protruding lysine at position 68 may prevent BKPyV from binding GM1 . To test this hypothesis , we introduced the smaller SV40 residue S68 into BKPyV VP1 . This single mutation switched the oligosaccharide specificity of BKPyV from b-series gangliosides to GM1 , altering a key attachment property of BKPyV . Known BKPyV VP1 sequences contain lysine , arginine or histidine at position 68 . All of these amino acids would block engagement of GM1 but promote or at least tolerate binding of b-series gangliosides . BKPyV strains do not carry a serine at position 68 , and this may indicate that recognition of GM1 instead of b-series gangliosides may not be advantageous to the virus in the context of the host organism . Possible explanations could be that GM1 is not very abundant or is differentially localized in human kidneys [25] , [26] . There likely exists an evolutionary constraint on BKPyV to bind b-series gangliosides , not GM1 , especially as the remainder of the SV40 GM1 binding site is mostly conserved in BKPyV . We have shown that unlike SV40 , the BKPyV K68S mutant is specific for GM1 containing the human sialic acid NeuNAc , and cannot engage its simian counterpart NeuNGc due to steric hindrance . As the WT BKPyV and K68S binding sites for terminal sialic acid are identical , WT BKPyV shares the same sialic acid specificity . The inability of K68S to propagate in simian Vero cells and its ability to attach to and propagate in human HEK cells highlights the importance of sialic acid specificity for viral species tropism . Collectively , our data on BKPyV , JCPyV and SV40 suggest that each virus has adapted to the most prominent sialic acid in its host . SV40 , JCPyV and BKPyV all feature a conserved platform of core residues that allows them to efficiently engage terminal sialic acid in a similar manner . These core residues mediate the vast majority of interactions . However , each virus achieves its distinct receptor specificity with a small number of strategically positioned satellite residues , such as K68 in the case of BKPyV , that form distinct contacts with additional carbohydrate moieties . Thus , these satellite residues define the context in which a terminal sialic acid can be bound , and as demonstrated here they present attractive opportunities for switching receptor specificities . It is tempting to speculate that at least some members of the polyomavirus family have evolved from an initial sialic-acid binding template through subtle modification of their satellite residues , thereby expanding their host range and tropism . The switching of specificity can occur naturally in viruses , and often triggers altered pathogenicity and species tropism . In many cases , switching is due to exceedingly small changes in the virus capsid structure . Prominent examples include different serotypes of adenoviruses , the canine and feline parvoviruses , as well as avian , swine and human influenza viruses [28] , [29] , [30] . In many of these cases , however , the molecular functions of these switches , such as how specific mutations alter the interaction with receptors , are not well understood at the atomic level . Switching polyomavirus receptor specificities , as demonstrated here in a first example , may therefore be a useful tool to study parameters that define host receptor recognition , viral uptake , and entry pathways .
Cells ( ATCC , Manassas , VA ) were maintained at 37°C in Cellgro Minimum Essential Medium Eagle ( MEM ) supplemented with 5% heat inactivated fetal bovine serum ( Atlanta Biologicals ) and penicillin ( 10 , 000 U/ml ) and streptomycin ( 10 , 000 µg/ml ) ( Gibco ) . Cells seeded in 24-well dishes were pre-incubated with media , dimethyl sulfoxide ( DMSO ) or gangliosides GM1 , GD2 , GD3 , GD1b , and GT1b ( Matreya ) at 0 . 3–30 µM for 17 h at 37°C . Prior to infection cells were chilled for 20 min at 4°C and washed with 2% MEM . Cells were infected with 8×105 Fluorescent Forming Units ( FFU ) per ml of BKPyV for 1 h at 37°C . The infectious media was then removed and replaced with fresh growth media . Infection was scored 72 h post infection by staining for VP1 and analyzed by indirect immunofluorescence . Construction of the BKPyV pUC-19 expression plasmid was previously described [20] . BKPyV VP1 mutants were generated by site directed mutagenesis using QuickChange XL ( Stratagene , La Jolla , CA ) . Mutant and WT plasmids were digested with BamHI ( Promega ) . Vero or HEK cells were transfected with ( 0 . 5 µg ) of linearized mutant or WT BKPyV plasmid DNA using Fugene 6 ( Roche ) . To detect expression of viral antigens cells were fixed with 2% paraformaldehyde in phosphate buffered saline ( PBS ) for 20 min at 25°C and permeabilized with 1% Triton-X 100 in PBS for 15 min at 37°C . Cells were incubated with the primary mouse monoclonal antibody PAb597 ( 1∶10 [20] , or PAb416 ( Ab-2 ) ( 0 . 2 mg/ml ) ( Calbiochem ) , [31] , used at 8 ng/µl to stain for BKPyV T-Ag . After incubation cells were washed with PBS and incubated with Alexa Fluor 488-labeled goat anti-mouse antibody in PBS ( Invitrogen ) . Vero or HEK cells were incubated with media or 3 µM GM1 for 17–18 h . Cells were washed and suspended in 100 µl ( 10 µg/mL ) ( Sigma ) of CTX or PBS for 30 min on ice with 10 min agitation . Cells were washed and then incubated with 100 µl of purified wild type or mutant BKPyV VP1 pentamers ( 100 µg/mL ) in PBS on ice for 2 h with 30 min agitations or with PBS alone . Cells were washed and suspended in 100 µl of Penta-His-AlexaFlour 488 conjugated antibody ( 10 µg/mL ) ( Qiagen ) in PBS on ice for 1 h with 15 min agitations . Cells were washed and fixed in 1% paraformaldehyde and binding analyzed using a BD FACSCanto II flow cytometer ( Benton , Dickinson , and Company ) . Data were analyzed using Flow Jo ( Tree Star Inc . ) software . We expressed and purified a truncated form of BKPyV VP1 that assembles into pentamers but does not form capsids . DNA coding for amino acids 30–300 of BKPyV VP1 was amplified by PCR and cloned into the pET15b expression vector ( Novagen ) in frame with an N-terminal hexahistidine tag ( His-tag ) and a thrombin cleavage site . The protein was overexpressed in E . coli BL21 ( DE3 ) and purified by nickel affinity chromatography and gel filtration on Superdex-200 . For crystallization , the tag was cleaved with thrombin before gel filtration , leaving non-native amino acids GSHM at the N-terminus . All NMR spectra were recorded using 3 mm tubes on a Bruker DRX 500 MHz spectrometer fitted with a 5 mm cryogenic probe at 283 K and processed with TOPSPIN 2 . 0 ( Bruker ) . For all proteins used for STD NMR , two NMR samples were prepared , containing either 1 mM GM1 oligosaccharide ( Alexis ) or 1 mM GD3 oligosaccharide ( Sigma ) . Protein concentrations were between 19 µM and 22 µM . An additional sample contained 20 µM WT BKPyV VP1 and 1 mM GD1b oligosaccharide ( Elicityl , F ) . Additional protein-free samples were prepared that only contained 1 mM GM1 , GD3 or GD1b oligosaccharide . These samples were used to verify that no direct excitation of ligand resonances occurred during STD NMR measurements , and they served as samples for the spectral assignment . 0 . 1 mM trimethylsilyl propionate was then added to the GD3 sample to allow 1H referencing . The buffer used for all NMR measurements contained 20 mM deutero-Tris pH 7 . 5 , 150 mM NaCl , and 20 mM deutero-DTT . Samples were prepared in D2O and no additional water suppression was used to avoid affecting the anomeric proton signals . The off- and on-resonance frequencies were set to 80 ppm and 7 ppm , respectively . The total relaxation delay was 4 s . A cascade of 40 Gaussian-shaped pulses with 50 ms duration each , corresponding to a strength of 65 Hz , and a saturation time of 2 s was used for selective excitation . A 10 ms continuous-wave spin lock filter with a strength of 3 . 7 kHz was employed in order to suppress residual protein signals . 32 k points were collected and zero filling to 64 k data points was employed . Spectra were multiplied with an exponential line broadening factor of 1 Hz prior to Fourier transformation . To assign oligosaccharide proton resonances , series of 1D 1H-TOCSY and COSY spectra as well as 1H , 13C-HSQC spectra were acquired . Literature values for related oligosaccharides served as additional assignment controls [32] , [33] , [34] , [35] . Assignment of the acetate methyl groups was taken from [34] for GD3 and [32] for GM1 . For crystallization , BKPyV VP1 was supplemented with 20 mM DTT and concentrated to 6 . 6–7 . 0 mg/ml . The protein was crystallized at 20°C by sitting drop vapor diffusion against a reservoir of 16–18% PEG 3 , 350 , 0 . 1 M HEPES pH 7 . 5 and 0 . 25 M LiCl ( drop size 300 nl protein+300 nl reservoir ) . Crystals were harvested into reservoir solution containing 14–16% PEG 3 , 350 and cryoprotected by soaking in harvesting solution supplemented with 30% ( v/v ) glycerol for 10 s before flash-freezing them in liquid nitrogen . For oligosaccharide complex formation , crystals were soaked in harvesting solution containing 20 mM GD3 oligosaccharide for 15 min before cryoprotection . Diffraction data were collected at the SLS ( Villigen , CH ) and processed with xds [36] , and the structure was solved by molecular replacement with Phaser [37] using the core of the SV40 VP1 pentamer ( 3BWQ ) as the search model . After rigid body and simulated annealing refinement in Phenix [38] , missing parts of the model were built in Coot [39] . Refinement proceeded by alternating rounds of refinement in Refmac5 [40] and model building in Coot . Fivefold non-crystallographic symmetry restraints were used throughout refinement . Oligosaccharide residues were located in weighted 2mFo-DFc and mFo-DFc electron density maps and refined with restraints from the CCP4 monomers library; only the α2 , 3- and α2 , 8- glycosidic linkages had to be user-defined . Data collection and refinement statistics are given in Table 1 . Coordinates and structure factor amplitudes were deposited with the RCSB data bank ( www . rcsb . org ) with entry codes 4MJ0 ( BKPyV VP1 bound to GD3 ) and 4MJ1 ( unliganded BKPyV VP1 ) . Structure figures were prepared with PyMol ( Schrödinger Inc . ) . To generate a model for the complex , we first explored the conformational space of GD1b alone using high-temperature molecular dynamics ( MD ) [41] and subsequently positioned the individual sampled snapshots into the binding site using the α2 , 8-disialic acid motif of crystal structure as an anchor point . Molecular dynamics simulation of GD1b was performed at 700 K for 100 ns using the MM3 force field as implemented in the TINKER software ( http://dasher . wustl . edu/tinker/ ) . Torsion restraints were applied on the ring torsions to avoid inversion of the carbohydrate rings during MD . Snapshots were recorded every 0 . 5 ps resulting in a conformational ensemble consisting of 200000 frames . Further processing of the data was performed using Conformational Analysis Tools ( CAT ) ( http://www . md-simulations . de/CAT/ ) . Conformational maps were calculated as described [41] in order to check that the accessible conformational space of the glycosidic linkages was sufficiently explored ( Supplemental Fig . S1 ) . All snapshots were positioned into the crystal structure using three atoms of residue NeuNAc 4R as an anchor and conformations that result in atom-overlaps with the protein were removed . Additionally two filters were applied on the remaining snapshots that control the position and orientation of residue NeuNAc 3R: Only snapshots were allowed to pass that have the center of the carboxylate and the N-acetyl group within 3 Å of the corresponding residue of the crystal structure . Several conformations were manually selected and refined based on 5 ns MD simulations at 300K in explicit water using YASARA [42] . Model coordinates are available from the authors upon request . Microarrays comprised lipid-linked oligosaccharide probes , neoglycolipids ( NGLs ) and glycolipids , robotically printed on nitrocellulose-coated glass slides using a non-contact instrument [43] , [44] . For the analyses , an array set of 21 ganglioside-related probes ( 18 sialylated and 3 non-sialylated , in house designation Ganglioside Dose Response Array set 1 ) was used , in which each probe was arrayed at four levels: 0 . 3 , 0 . 8 , 1 . 7 and 5 . 0 fmol/spot . The microarray analyses were performed essentially as described [17] . In brief , microarrays were blocked in 5 mM HEPES ( pH 7 . 4 ) , 150 mM NaCl , 0 . 3% ( v/v ) Blocker Casein ( Pierce ) , 0 . 3% ( w/v ) bovine serum albumin ( Sigma ) and 5 mM CaCl2 ( referred to as HBS-Casein/BSA ) . WT and mutant BKPyV VP1 were diluted in HBS-Casein/BSA and overlaid at 300 µg/ml and 150 µg/ml , respectively , followed by incubation with mouse monoclonal anti-poly-histidine and biotinylated antimouse IgG antibodies ( both from Sigma ) . SV40 VP1 was tested as a protein-antibody complex that was prepared by preincubating with mouse monoclonal anti-poly-histidine and biotinylated anti-mouse IgG antibodies at a ratio of 4∶2∶1 ( by weight ) and diluted in HBS-Casein/BSA to provide a final SV40 VP1 concentration of 150 µg/ml . The SV40 and K68S protein samples had been supplemented with DTT to prevent dimerization of pentamers , while WT BKPyV VP1 was analysed without DTT . Binding was detected with Alexa Fluor-647-labelled streptavidin ( Molecular Probes ) . Microarray data analyses were as described [45] .
|
Viruses need to bind to receptors on host cells for viral entry and infection , and the type of receptor bound determines the range of hosts and tissues the virus can infect . Viruses within a family often vary in their tissue distribution and pathogenicity because changes in receptor specificity can produce a virus with different spread and infectivity . In fact , many transmissions between species are based on a virus acquiring binding capability for a new receptor . The structural changes that underlie such receptor switching are not well understood . We have analyzed the structural requirements for receptor binding and switching of the human BK polyomavirus ( BKPyV ) , the causative agent of polyomavirus-associated nephropathy . We show that BKPyV uses specific gangliosides that all contain a common α2 , 8-disialic acid motif to infect cells , and have characterized the interaction in atomic detail . Our data explains the requirement for this disialic acid motif and in particular highlights a single amino acid that is central to determining specificity . Mutation of this residue switches the receptor specificity , enabling BKPyV to infect cells bearing a different class of gangliosides . Our findings highlight the plasticity of viral receptor binding sites and form a template to retarget viruses to different receptors and cell types .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2013
|
A Structure-Guided Mutation in the Major Capsid Protein Retargets BK Polyomavirus
|
Evolutionary theory predicts that sexually antagonistic mutations accumulate differentially on the X chromosome and autosomes in species with an XY sex-determination system , with effects ( masculinization or feminization of the X ) depending on the dominance of mutations . Organisms with alternative modes of inheritance of sex chromosomes offer interesting opportunities for studying sexual conflicts and their resolution , because expectations for the preferred genomic location of sexually antagonistic alleles may differ from standard systems . Aphids display an XX/X0 system and combine an unusual inheritance of the X chromosome with the alternation of sexual and asexual reproduction . In this study , we first investigated theoretically the accumulation of sexually antagonistic mutations on the aphid X chromosome . Our results show that i ) the X is always more favourable to the spread of male-beneficial alleles than autosomes , and should thus be enriched in sexually antagonistic alleles beneficial for males , ii ) sexually antagonistic mutations beneficial for asexual females accumulate preferentially on autosomes , iii ) in contrast to predictions for standard systems , these qualitative results are not affected by the dominance of mutations . Under the assumption that sex-biased gene expression evolves to solve conflicts raised by the spread of sexually antagonistic alleles , one expects that male-biased genes should be enriched on the X while asexual female-biased genes should be enriched on autosomes . Using gene expression data ( RNA-Seq ) in males , sexual females and asexual females of the pea aphid , we confirm these theoretical predictions . Although other mechanisms than the resolution of sexual antagonism may lead to sex-biased gene expression , we argue that they could hardly explain the observed difference between X and autosomes . On top of reporting a strong masculinization of the aphid X chromosome , our study highlights the relevance of organisms displaying an alternative mode of sex chromosome inheritance to understanding the forces shaping chromosome evolution .
As males and females differ in their optimal values for most phenotypic traits , selection often runs in opposite directions in the two sexes , a situation called sexual antagonism [1] . Since males and females share most of their genome , intra-locus conflicts appear when the same gene is selected for different optima in each sex . Because sex chromosomes have a sex-biased transmission pattern they are expected to accumulate different types of sexually antagonistic mutations than autosomes , as originally shown by Rice [2] and elaborated by further models [3]–[6] . The Y chromosome is expected to accumulate alleles that are good for males ( even if detrimental for females ) because the beneficial effects are always achieved but not the costs . It has been shown that the Y is indeed enriched in genes with male-specific functions ( [7]–[9] , see [10] for W in ZW systems ) . The situation of the X chromosome for XX/XY or XX/X0 systems ( or Z in ZZ/ZW systems ) is however more complex . X-linked recessive mutations are always exposed to selection in the heterogametic sex ( i . e . XY males or ZW females ) , and their spread in the population is thus facilitated if beneficial or impeded if deleterious for the heterogametic sex . Conversely , for a dominant or partly dominant sex-linked mutation , the homogametic sex ( i . e . XX females or ZZ males ) has the highest influence on the evolutionary fate of mutations , because such a mutation experiences 2/3 of the time selective pressures acting on the homogametic sex . Overall , Rice's model underlines the crucial effect of the recessive or dominant character of mutations , with the X accumulating either male- or female-beneficial mutations depending on their dominance status . Although Rice's model assumes that the dominance coefficient of each allele is the same in both sexes , which may not necessarily be the case ( e . g . [3] , [11] ) , more general models show that its predictions regarding the spread of sexually antagonistic alleles on the X versus autosomes still hold when dominance differs among the sexes , the outcome depending on dominance in the heterogametic sex [3] . For example in a XY species , male-beneficial alleles invade more easily the X chromosome when they are recessive in males , and the autosomes when they are dominant in males , while female-beneficial alleles invade more easily the X when they are dominant in males ( and the autosomes when they are recessive in males ) . Interestingly , some of Rice's predictions have been experimentally validated by engineering a novel sexually antagonistic allele in Drosophila melanogaster using genetic constructs [12] . The evolution of sex-biased gene expression has been proposed as a possible way of resolving conflicts raised by the spread of sexually antagonistic alleles over protein-coding sequence [2] , see also Box 5 in [13] ( throughout , male-biased [or female-biased] genes refer to genes overexpressed in males [or females] ) . Indeed , once a sexually antagonistic allele is present ( either segregating or fixed ) in a population , any modifier of expression ( not necessarily physically linked to the target ) that reduces the expression of the gene in the harmed sex will be selected for and fixed , allowing the allele that favors the other sex to reach fixation too ( if it is not already fixed ) [2] . Although this hypothesis is frequently presented as plausible ( e . g . [13]–[15] ) , we lack empirical demonstration that sex-biased gene expression might have been selected because it allowed to solve past sexual antagonism , presumably because of the difficulties to demonstrate that a given substitution in a genome corresponds to the fixation of a sexually antagonistic allele . Accordingly , empirical support for this hypothesis is at best correlative: if intra-locus sexual conflicts are frequently resolved by the evolution of a sex-biased gene expression , we expect an accumulation of either male- or female-biased genes on the X ( or Z ) , depending on the average level of dominance of sexually antagonistic mutations . This could account for the non-random distribution of genes with sex-biased expression between the X ( or Z ) and the autosomes observed in different groups of animals . Male-biased genes are overrepresented on the X chromosome in mammals [16]–[20] , but under-represented in nematodes [21] , flies [22]–[24] but see [25] , mosquito [26] , [27] but see [28] and flour beetle [29] . Female-biased genes are also overrepresented on the X in some species [16] , [22] , [29] but are under-represented in nematodes [21] . In systems where female is the heterogametic sex ( i . e . ZW systems ) , the Z is enriched with testis genes in the silkworm [30] and birds [19] , [31]–[33] , but depleted from female-biased genes in birds [31] , [33] , [34] but see [35] . However , several other factors could explain the opposite trends observed in different species . In particular , the inactivation of the X during late spermatogenesis ( Meiotic Sex Chromosome Inactivation , MSCI , [36] , [37] ) drives spermatogenesis genes out of the X in Drosophila and mammals [16] , [20] , [36] , [38] , [39] . Recent evidence for the absence of X dosage compensation in Drosophila testis [40] also explains the apparent paucity of genes expressed in the male germline on the Drosophila X [41] . Furthermore , the X in the whole body of male Drosophila is naturally hyper transcribed as a whole to equalize X∶A expression rate for dosage compensation [42] , so that evolving further overexpression of X-linked genes in males may be difficult [43] , [44] . Finally , the absence of a global mechanism of dosage compensation in birds ( a ZW system ) might also account for the overrepresentation of male-biased genes on the Z [45] , [46] . As a result , the non-random distribution of sex-biased genes between autosomes and sex chromosomes does not in itself demonstrates that sex-biased gene expression evolved to solve past intra-locus sexual conflicts over protein coding sequences . The most convincing ( through indirect ) evidence for the resolution of intra-locus sexual conflicts via the evolution of sex-biased gene expression comes perhaps from the non-recombining old homomorphic ZW sex chromosomes of the emu , a ratite bird [14] . While it is widely accepted that the cessation of recombination between proto-sex chromosomes has been favored because of the accumulation of sexually antagonistic alleles in the vicinity of the sex-determining region [47]–[49] , in the emu , the evolution of sex-biased gene expression may have provided an alternative solution to alleviate the segregation load due to sexually antagonistic alleles [14] . This could explain the occurrence of old and homomorphic sex chromosomes in ratite birds . Alternative sex-determining systems are of high interest , because they allow studying the selective forces driving the different patterns from another perspective , e . g . [33] , [50] . In this article , we investigate the evolutionary forces driving the chromosomal location of sexually antagonistic mutations in aphids . Aphids have an XX/X0 sex determination system whereby females carry two X chromosomes and males only one X ( while both sexes are diploid for the autosomes ) . Yet , aphids are peculiar because in addition to males and sexual females , apomictic parthenogenetic females ( diploid at the X and autosomes ) represent a major component of their life cycle . This could set the stage for a three-way genetic conflict since mutations may be beneficial to either males , sexual females or asexual females . Furthermore , the alternation of asexual and sexual reproduction results in an unusual ( autosome-like ) inheritance of the X ( see Figure 1 and [51] ) . During the first part of the cycle ( spring and summer ) , asexual females ( XX/AA ) reproduce through parthenogenesis . In autumn , asexual females generate males and sexual females in response to photoperiodic cues: sexual females are therefore strict clones of asexual females ( hence also XX/AA ) , while one of the X is lost to generate X0/AA males [52] . The fusion of an ovule ( haploid for the X and for the autosomes ) and a sperm ( always haploid for the X and for the autosomes because males produce only X-bearing gametes ) restores diploidy at both Xs and autosomes to generate an egg from which an asexual female will hatch in spring ( Figure 1 ) . Hence , the X is transmitted equally through males and sexual females in aphids: one half of the Xs found in the sexual progeny comes from the mother and the other half from the father ( i . e . the X have an “autosome-like” inheritance , Figure 1 , see also [51] ) . This contrasts with standard XY systems , where the X is transmitted twice more often through females than through males . These differences between aphids and standard systems have been shown to influence the neutral diversity and gene evolutionary rates of the X chromosome in aphids [51] , and could also affect the evolutionary forces that promote the accumulation of sexually antagonistic mutations on sex chromosomes in standard XX/XY or ZZ/ZW systems [2] , . Here we investigate how the particular inheritance of the X , and the alternation of sexual and asexual reproduction involving specialized reproductive morphs ( males , sexual females and asexual females ) affect the location of sexually antagonistic mutations ( the term sexually antagonistic mutation when applied to aphids refers to a mutation beneficial for at least one of the three reproductive morphs and deleterious for at least one of the two others ) . Using a modeling approach , we show that: 1 ) sexually antagonistic mutations beneficial for males – detrimental to asexual females are expected to accumulate preferentially on the aphid X chromosome , 2 ) mutations beneficial for asexual females – detrimental for males are expected to accumulate preferentially on autosomes , 3 ) the selective effect of a mutation in sexual females has little effect upon its genomic location , and 4 ) in contrast with previous results derived for standard systems , these qualitative predictions are unaffected by the dominant or recessive character of mutations . Under the hypothesis that the evolution of sex-biased expression to restrict the product of sexually antagonistic allele to the sex it benefits might solve intra-locus sexual conflicts , one expects that male-biased genes should be enriched on the X while asexual female-biased genes should be enriched on autosomes . Using gene expression data ( RNA-Seq ) in males , sexual females and asexual females of the pea aphid , we confirm these theoretical predictions .
Using a general model in which a given mutation ( denoted B ) may have different effects on the fitnesses of asexual females , sexual females and males ( see Table 1 , Methods ) , one predicts that an autosomal mutation increases in frequency when rare ifwhere t is the number of clonal generations per cycle and wa , b/B , wf , b/B and wm , b/B are the fitnesses of heterozygous asexual females , sexual females and males , relative to the fitnesses of individuals homozygous for the ancestral allele . When the mutation occurs on the X chromosome , this condition becomeswhere wm , B/0 is the fitness of hemizygous males carrying the mutation . Using the notation of Table 1 , we have wm , b/B = 1+hm sm and wm , B/0 = 1+sm . Assuming that 0<hm<1 , the condition for invasion of a male-beneficial mutation ( sm>0 ) is therefore more stringent when this mutation occurs on an autosome than when it occurs on the X chromosome . Conversely , the condition for invasion of a male-detrimental allele ( sm<0 ) is more stringent when it occurs on the X chromosome than on an autosome . Among the mutations selected differentially between the sexes , one thus expects an excess of male-beneficial , female-detrimental mutations on the X chromosome , and an excess of female-beneficial , male-detrimental mutations on autosomes . Note that these expectations are not affected qualitatively by the dominance coefficients of mutations , and thus differ from those derived for standard XX/XY sex-determining systems [2] , for which opposite results are found for dominant or recessive mutations . Finally , it is important to note that selection coefficients of mutations in asexual females ( sa in Table 1 ) have a disproportionate effect on invasion criteria , due to the many asexual generations per cycle ( exponent t in the equations above ) . Therefore , when si ( where i stands for m or f ) has the same sign as sa , one expects that ( in most cases ) si has little effect on whether the mutation spreads or not . For this reason , sm should generally have little effect on conditions for the spread of mutations with sm<0 , sf>0 and sa<0 or with sm>0 , sf<0 and sa>0 ( the direction of selection is the same in males and asexual females , but different in sexual females ) . Although one would predict ( based on the arguments above ) that the first type of mutation ( sm<0 , sf>0 , sa<0 ) is found more often on autosomes and the second type ( sm>0 , sf<0 , sa>0 ) more often on the X , the bias should be rather small . Our simulations confirm that mutations rising in frequency on the X but not on autosomes correspond to sexually antagonistic mutations favorable for males but slightly deleterious for asexual females for all values of dominance h ( Figure 2 , see also Table S1 ) . In contrast , mutations rising in frequency on the autosomes but not on the X are deleterious for males but slightly beneficial for asexual females . Note that when selection coefficients in asexual females are too strong , the fate of mutations becomes independent of sf and sm ( again because of the larger number of asexual generations per cycle ) , and therefore also independent of their genomic location . When sf and sa have opposite signs , the overall effect of selection in females is attenuated ( the product ( wa , b/B ) t wf , b/B in the equations above becomes closer to 1 ) , which increases the parameter range where mutations are favored in one genomic location only . This effect is visible on Figure 2 only for high values of h , as the overall effect of selection on rare alleles in females is enhanced when h is high . For lower values of h , selection coefficients of mutations in sexual females have little effect on their preferred genomic location . In contrast with these results on the aphid-like system , simulating a standard XX/XY sex-determining system yields the classical prediction that the type of mutation invading preferentially the X chromosome depends on whether mutations are dominant or recessive ( Figure S1 , see also [2] ) . Additional simulations performed specifically for the aphid system showed that sexually antagonistic mutations beneficial to males – deleterious to asexual females accumulated on the X while those deleterious for males – beneficial to asexual females rose in frequency on autosomes under all tested conditions ( Figure S2 , Table S1 ) . These additional results include in particular a set of simulations run under a general model of dominance ( i . e . the dominance of an allele can differ between morphs , ha≠hm≠hf ) and another set where beneficial ( resp . deleterious ) mutations are dominant ( resp . recessive ) , as predicted by different models of stabilizing selection on quantitative traits [54] . Based on the previous arguments , one can deduce the preferred genomic location ( X versus autosomes ) of mutations characterized by different combinations of selective effects in males , sexual females and asexual females ( Table 2 , Predictions 1 ) . We also derived the expected pattern of expression of genes bearing such kind of sexually antagonistic mutations under the hypothesis that modifiers decreasing expression in the harmed sex will be selected to solve the conflict ( Table 2 , Predictions 2 ) . By combining Predictions 1 and 2 , we obtained the expected chromosomal location for genes with contrasted expression patterns in the three reproductive morphs ( Table 2 , Predictions 3 ) . More precisely , we expect 1 ) an enrichment of the X with genes overexpressed solely in males ( i . e . M+F−A− genes , where M , F and A refer to male , sexual female , asexual female , respectively , and the sign represents relative expression in each morph ) and with those overexpressed in both males and sexual females ( M+F+A− ) , 2 ) an enrichment of autosomes with genes overexpressed in asexual females ( M−F−A+ ) or in both asexual and sexual females ( M−F+A+ ) , 3 ) little chromosome bias for genes overexpressed in sexual females ( M−F+A− ) or in both males and asexual females ( M+F−A+ ) , with a slight autosomal bias for the former and a slight bias towards X for the latter . We studied eight RNA-Seq libraries ( three for males , two for sexual females and three for asexual females ) including two previously published datasets complemented by six new libraries specifically generated for this study . We found that 5706 out of the 36990 predicted genes on the pea aphid genome were differentially expressed ( p<0 . 05 after adjusting for multiple testing using the Benjamini-Hochberg method implemented in the R package DESeq ) between the three reproductive morphs . When considering the 3712 genes tagged either as X-linked or autosomal ( i . e . genes located within a 200 kb-window centered on the microsatellite markers tagged as X-linked or autosomal ) , we observed that M+F−A− genes with a 2-fold expression bias were overrepresented on the X chromosome ( f ( X ) = 0 . 24 , Chi-square-test: p<10−8 , Table 2 , Figure 3 ) compared to expected proportion ( f ( X ) = 0 . 12 ) . The bias further increased to f ( X ) = 0 . 31 and 0 . 34 when considering only M+F−A− genes at least 5-fold or 10-fold overexpressed , respectively ( Table 2 , Figure 3 ) . M+F+A− genes were also significantly overrepresented on the X at a 2-fold expression threshold ( f ( X ) = 0 . 26 , p = 0 . 03 ) , and at larger thresholds the effect became highly significant , the frequency of X-linkage reaching 0 . 50 ( p = 0 . 0001 ) and 0 . 57 ( p = 0 . 0002 ) for 5- and 10-fold overexpressed genes , respectively . By contrast , M−F−A+ and M−F+A+ genes were depleted on the X chromosome , the frequency ranging from 0 . 04 to 0 . 07 for M−F−A+ genes at different expression thresholds ( p<0 . 05 in all cases ) . For M−F+A+ genes , the deficiency on the X was significant only at the 2-fold threshold ( f ( X ) = 0 . 03 , p = 0 . 019 ) , presumably because of lack of power due to the low number of M−F+A+ genes satisfying the 5- and 10-fold expression criteria ( n = 22 and 13 , respectively ) . Despite a high statistical power , genes overexpressed only in sexual females ( M−F+A− ) showed no significant chromosome bias at any of tested thresholds ( p ranging from 0 . 21 to 0 . 63 , f ( X ) ranging from 0 . 07 to 0 . 10 , Table 2 , Figure 3 ) . M+F−A+ genes also showed no significant chromosome bias ( p ranging from 0 . 16 to 0 . 60 , and f ( X ) from 0 . 06 to 0 . 09 , Table 2 , Figure 3 ) . Overall , the observed genomic location for genes with contrasted patterns of expression fits Predictions 3 , derived under the hypothesis that the evolution of sex-biased gene expression might solve intra-locus sexual conflicts . When similar analyses were performed considering different window sizes to assign genes to the X or the autosomes , we found fairly similar results for the 100 kb and 200 kb window . However , as the size of the window increased , the contrast between X chromosome and autosomes regarding their sex-biased gene content decreased ( see Table S2 ) . Nevertheless , the X was still significantly enriched with genes overexpressed in males ( M+F−A− genes ) even when no window size restriction was applied ( p ranging from 10−4 to 10−15 depending on the fold-difference in expression considered , Table S2 ) . The frequency of low expressed genes ( <0 . 1 RPKM , Reads Per Kilobase of exon model per Million mapped reads ) on the autosomes was 5% while it reached 14% on the X ( Chi-square test: p<10−11 ) . X-linked genes were less expressed than autosomal ones in the three reproductive morphs , when considering all genes or those supported by >0 . 1 RPKM , Figure 4A–B , p<10−7 in all cases ) . In both cases , computationally doubling X-linked gene expression ( to account for the haploid state of the X in males ) resulted in a higher expression of X-linked genes compared to autosomal genes ( p<10−6 ) , suggesting partial dosage compensation . In contrast , X-linked and autosomal genes represented by more than 5 RPKM showed no significant difference in expression rate in males ( p = 0 . 13 ) , and computationally doubling X-linked genes expression in males resulted in significant higher gene expression for the X than for autosomes ( p<10−9 ) , suggesting a dosage compensation of these genes . X-linked genes were expressed at a much lower rate in both types of females ( Figure 4C , p<10−10 ) .
In this study , we demonstrate theoretically that the X chromosome of aphids is expected to accumulate sexually antagonistic alleles beneficial for males , while autosomes are expected to accumulate sexually antagonistic alleles beneficial for asexual females . We also identified a substantial masculinization of the aphid X , meaning that this chromosome is enriched with male-biased genes , and an “asexualization” of autosomes , that are enriched with asexual female-biased genes . Our model predictions , namely the enrichment of the X with sexually antagonistic mutations beneficial for males ( and the enrichment of autosomes with those favorable to asexual females ) regardless of the dominance values , differ markedly from those derived for standard XY sex-determining systems , whereby the X accumulates male-beneficial female-detrimental alleles that are recessive in males ( hm<0 . 5 ) , and female-beneficial male-detrimental alleles that are dominant in males ( hm>0 . 5 ) [2] , [3] , see also Figure S1 . The difference between aphid and standard systems arises from the peculiar inheritance of the X in aphids ( Figure 1 , see also [51] ) where the X is transmitted equally through males and sexual females . Male-beneficial , sexually antagonistic alleles rise in frequency on the X more easily when they are recessive in females ( because their deleterious effects in asexual females are rarely expressed as long as they are rare in the population ) . By contrast , when ha increases , only male–beneficial mutations having minor deleterious effects in asexual females can rise in frequency on the X , since mutations that are too deleterious for asexual females will be efficiently counter-selected during the several rounds of asexual reproduction . Symmetrical arguments explain the relative enrichment of autosomes with mutations favorable for asexual females - deleterious for males ( for a given sm , the effect of selection among males is weaker for autosomal loci , unless hm = 1 ) . Differences in effective population sizes between the asexual morph ( characterized by relatively small population sizes at the beginning of the asexual round of reproduction due to winter bottlenecks ) and sexual morphs ( characterized by large population sizes since they are generated after several rounds of clonal population growth ) - that might differently affect invasion probabilities of mutations in sexual and asexual morphs through drift- are not accounted for in our analytical model . Yet , stochastic and demographic effects are incorporated in our individual-based simulations , ensuring that these predictions are robust to such effects . Aphids also contrast with other XY , X0 or ZW systems because effective population sizes for X and autosomes are similar due to their peculiar inheritance of Xs and life-cycle [51] . This implies that the fate of an X-linked or autosomal mutation should not be differentially affected by drift in aphids , contrarily to other sex-determining systems [4] . Aphids are particularly valuable systems because as shown above , the qualitative predictions for the preferred genomic location of sexually antagonistic mutations is unaffected by dominance ( contrasting to standard XY systems ) , a parameter that is difficult to estimate for large gene sets , e . g . [55] . Furthermore , the occurrence of three different reproductive morphs in aphids ( males , sexual and asexual females ) creates conditions for the emergence of a three-way conflict , which allows making specific and precise predictions regarding the genomic location of sexually antagonistic genes ( Table 2 ) . Additionally , the fact that some aphid lineages have lost the ability to produce functional males and/or sexual females but still exchange genes with cyclically parthenogenetic populations [56] increases the potential for sexually antagonistic genetic variation . Indeed , deleterious alleles in sexual aphid individuals are not as strongly counter-selected ( e . g . [57] ) as they would be in a strictly sexual species where sexual individuals represent an obligate step for transmitting genes . A substantial proportion of the predicted genes of the pea aphid showed differential expression between morphs ( 21% of the 27003 genes represented by at least 5 reads over the 8 normalized libraries were biased ) . Different mechanisms have been proposed to account for sexually dimorphic gene expression , including the fixation of mutations in a regulatory sequence leading to a biased expression , without the prior increases in frequency of a sexually antagonistic allele [58]–[60] , epistatic interactions between sexually antagonistic alleles ( which may be in protein-coding sequences [2] or in regulatory sequences [11] ) and sex-limited modifiers of expression , gene duplication followed by divergence in expression of the two copies [61] , [62] or genomic imprinting on allele expression dependent on its parent of origin [63] . Of course , all these mechanisms may have contributed to some extent to differential gene expression between males , sexual females and asexual females in aphids . Our empirical analyses highlight however an important deviation from a random genomic distribution for genes differentially expressed between morphs , with male-biased genes being enriched on the X and asexual female-biased genes being enriched on autosomes . The evolution of sex-biased gene expression may not necessarily result from the spread of sexually antagonistic alleles: a mutation that increases ( or decreases ) the expression of a given gene in one sex only may be favored ( if it allows a better match with the optimum of that sex ) even without the previous existence of an intra-locus sexual conflict . According to our theoretical predictions , a cis-regulatory mutation affecting gene expression in males ( and thereby increasing the fitness of males ) should spread more easily when it occurs on the X chromosome . However , there is no obvious reason why such male-beneficial regulatory mutations should more often increase gene expression rather than decrease it: therefore , this scenario does not explain the enrichment of male-biased genes on the X , and the enrichment of asexual female-biased genes on autosomes . Models based on gene duplication and divergence [61] should not lead to expect an excess of male-biased genes on the X in aphids , because gene duplication dampens down expression of recessive alleles in males [61] , the key factor showed here to favor the spread of male-beneficial alleles on the X under a single-gene model in aphid . Sex-biased expression may also evolve by differential imprinting according to parental origin ( so that male-beneficial alleles transmitted by fathers are turned off in their daughters , while female-beneficial alleles transmitted by mothers are turned off in their sons ) [63] , but this process is unlikely to occur in aphids given the presence of many asexual generations between each sexual event . Contrastingly , epistatic interactions between sexually antagonistic alleles over protein-coding sequence and sex-limited modifiers of expression [2] should lead to an excess of male-biased genes on the X and an excess of asexual female-biased genes on autosomes in aphids , as observed on our empirical data . Furthermore , this scenario predicts only slight chromosomal bias for sexual female-biased genes or for those underexpressed in that morph only ( Table 2 ) , and accordingly , we did not detect significant deviation from equivalent frequencies on the X and autosomes . Our observations regarding the non-random genomic location of sex-biased genes thus suggest that sexual antagonism may have played a role in the evolution of sex-biased gene expression in aphids . Note however that we do not argue that all sex-biased genes evolved through this mechanism . We acknowledge that other models [11] , [58] , [61]–[66] must be invoked to explain the consequent fraction of X-linked genes ( respectively autosomal genes ) that are asexual female-biased ( respectively , male-biased ) in aphids . Hence , sex-biased gene expression does not necessarily imply a history of sexually antagonistic fitness effect . Moreover , sexual antagonism over protein-coding sequence should not systematically conduct to an evolution of sex-biased gene expression . Indeed , genes with large pleiotropic effects might be too constrained to evolve sex-biased expression [67] , [68] or some genes might be too essential for allowing to cease their expression in one of the sexes , hence additional steps such as duplications would be required to solve conflicts [61] , [62] . Several factors might contribute to the strong masculinization of the X in aphids , and to the differences between aphids and other invertebrate taxa . First , our models have shown that male-beneficial alleles can accumulate on the aphid X independently of dominance . Second , meiotic sex chromosome inactivation ( MSCI ) has been shown to drive late stage spermatogenesis genes out of the X in mammals , Drosophila and Caenorhabditis elegans [16] , [36] , [38] , [69] . One of the several hypotheses to explain MSCI is that it evolved to prevent the spread of sex-ratio distorters on sex chromosomes [50] , [70] . Since male aphids produce only X-bearing gametes ( i . e . haploid for the X and for autosomes , hence the progeny from sexual reproduction is 100% asexual female ) , there would be no reason for MSCI to evolve or to be maintained ( if this hypothesis explains MSCI ) . Finally , the X of Drosophila is not dosage-compensated in the male germline , further contributing to an apparent demasculinization of the X in this genus when relying on a 2-fold change in expression to identify male-biased genes [40] , [41] . Whether dosage compensation occurs or not in the aphid male germline is unknown , but if so , this would further increase the contrast between Drosophila and aphids . We examined whether our data support dosage compensation in the pea aphid whole body . We found an equal expression of X-linked and autosomal genes in males only for genes expressed at a relatively high expression ( RPKM>5 , Figure 4C ) which would be compatible with partial dosage compensation . This could be a relic of dosage compensation that would have evolved in ancestral reproductive system in which only males and sexual females were present ( i . e . before the acquisition of parthenogenesis ) – partial dosage compensation has indeed been found in most XY or X0 organisms studied so far [71]–[75] . However , a surprising pattern of expression was found for females , which have equal numbers of copies of Xs and autosomes , yet showing reduced expression levels for X-linked genes , for all categories of expression level ( Figure 4 ) . Such an observation does not fit with a scenario of dosage compensation whereby differences in expression among chromosomes types are expected to be deleterious , but is best explained by a scenario involving sexual antagonism: because the phenotype of both kinds of females is probably relatively close compared to males , it is likely that a consequent fraction of sexually antagonistic mutations would have similar fitness effects on sexual and asexual females . Should such an antagonism be solved by the evolution of sex-specific expression biases , this could explain the under-expression of X-linked genes ( compared to autosomal genes ) in both sexual and asexual females ( Figure 4 ) . Nevertheless , we strongly caution that our preliminary conclusions on dosage compensation - drawn from whole body ( i . e . including ovary and testis ) – need to be validated by transcriptomic data from tissues unaffected by sex-specific evolutionary forces . Here , by modeling the preferred genomic location of sexually antagonistic mutations in species characterized by: 1 ) an unconventional inheritance of the X chromosome and 2 ) the presence of different reproductive morphs ( males , sexual females , asexual females ) rather than just two sexes , we have been able to formulate several predictions regarding the genomic location of genes differentially expressed among morphs , under the hypothesis that the evolution of sex-biased expression to restrict the product of a sexually antagonistic allele to the sex it benefits might solve intra-locus sexual conflicts [2] . We then found a non-random genomic distribution of sex-biased genes that fits predictions derived from our model . Furthermore , we reported a strong masculinization of the X chromosome , contrasting with the general demasculinization of the X in all non-mammal species investigated so far and argue that it is likely due to its peculiar inheritance pattern . This study therefore highlights the relevance of organisms with peculiar modes of inheritance of sex chromosomes , such as aphids and some nematodes [76] , [77] , as complementary models to study the forces driving the evolution of sex chromosomes .
We used a one-locus , two-alleles model to track the spread of a sexually antagonistic mutation under an aphid life cycle . The first part of the life cycle consists in a number t of discrete , clonal generations; then , sexual females and males are generated and reproduce sexually ( we assumed that mating is random ) . Two alleles b and B segregate at a given locus , and have different effects on the fitnesses of asexual females , sexual females and males ( Table 1 ) . We assumed that selection occurs among asexual females at each clonal generation , while it occurs among females and among males during the sexual phase . In a very large , randomly mating population , the spread of allele B from rarity is determined by its effects in heterozygous ( or hemizygous ) individuals: wa , b/B , wf , b/B , wm , b/B ( locus on an autosome ) or wm , B/0 ( locus on the X chromosome ) - see Table 1 . Assuming that the frequency p of allele B is small , the change in frequency over the full life cycle is approximately ( to the first order in p ) :when the locus is located on an autosome , andwhen the locus is on the X chromosome . From these expressions , predictions on the preferred genomic location of different types of mutations can be derived ( see Results ) . We also used individual-based simulations written in R [78] to explore the spread of allele B in a more realistic model incorporating stochasticity and demographic effects . For each replicate of the simulation , the selective coefficients sf , sm and sa ( see Table 1 ) are randomly and independently drawn from a uniform distribution between [−0 . 5 , 0 . 5] . Depending on the sampled values , allele B can thus be 1 ) beneficial for all morphs ( i . e . sa , sf , sm>0 ) , 2 ) deleterious for all morphs ( i . e . sa , sf , sm<0 ) , 3 ) sexually antagonistic if the mutation is beneficial for at least one morph and deleterious for at least one other ( e . g . sm>0 but sa<0 ) . For simplicity , we assumed that the dominance coefficient h of allele B is the same in all three morphs and is drawn from a uniform distribution between [0 , 1]; however we relaxed the hypothesis of identical dominance coefficients in some simulations ( see below ) . For each combination of selection coefficients , two cases were simulated: ( i ) mutation B is carried by an autosome , ( ii ) mutation B is carried by the X chromosome . At generation 0 , N = 1000 asexual females are created: the number of individuals of each genotype ( b/b , b/B or B/B ) is drawn from a multinomial distribution assuming Hardy-Weinberg equilibrium and an initial frequency of allele B of 0 . 005 ( hence , on average 10 mutant alleles B are present at generation 0 ) . Then , females reproduce through apomictic parthenogenesis for t = 10 generations . At each round of asexual reproduction , the number of individuals Ii generated by each asexual female of genotype i is drawn from a Poisson distribution , with mean , where the term fa represents the fecundity of asexual females ( fa = 2 ) and the second term is the relative fitness of asexual female of genotype i . After these 10 generations , each female gives birth ( by parthenogenesis ) to one sexual female and one male ( which carry the same diploid autosomal genome as their asexual parent ) . The number of gametes generated by each sexual female with genotype i is then sampled from a Poisson distribution with parameter , where fs is fecundity ( set to 5 ) and the second term is the relative fitness of females with genotype i . If mutation B is located on an autosome , the number of gametes produced by each male is sampled from a Poisson distribution with parameter Ni , m , which takes the same form as Ni , f ( replacing f by m subscripts ) . If the mutation is carried by the X chromosome , Ni , m is given by . Finally , 1000 male and 1000 female gametes are randomly drawn from the pool of gametes to generate the 1000 asexual females of the next cycle . Each simulation runs for 100 cycles ( a cycle including 10 rounds of asexual reproduction followed by one event of sexual reproduction ) , and we recorded the frequency of the mutant allele B in asexual females after these 100 cycles . To obtain an accurate estimate of the frequency of the mutant allele at generation 100 , mutant allele frequency was averaged over 25 replicates ( run with identical selection and dominance coefficients ) . We tracked mutations that have opposite fates in the different types of chromosome . We considered a mutation B as rising in frequency on the X but not on autosomes if the frequency of the mutation at generation 100 ( averaged over 25 independent runs with identical selection and dominance coefficients ) increased at least ten-fold on the X ( i . e . reached an average frequency of 0 . 05 when on the X ) but was lower than 0 . 005 when on autosomes . Reciprocally , mutations that reached frequencies higher than 0 . 05 on autosomes but lower than 0 . 005 on the X were considered as specifically rising in frequency on the autosomes . The characteristics of such mutations ( i . e . sa , sf , sm , h ) were recorded . To explore a large panel of combinations of selection ( sa , sf , sm ) and dominance ( h ) coefficients , we repeated this procedure ( including the simulation of 25 replicates for both types of chromosomes ) by randomly drawing 200 , 000 sets of sf , sm , sa and h values . To contrast expectations for aphids with those for standard XX/XY or ZZ/ZW sex-determining systems [2] , we simulated the evolution of a newly appeared mutation B in standard systems 1 ) on the X and 2 ) on autosomes . In that case , the population consisted of 500 males and 500 females . The amount of gametes produced by males and females was proportional to their relative fitness values . Then 1000 male gametes ( 500 A/Y and 500 A/X ) and 1000 female gametes were randomly drawn to generate the 500 males and 500 females of the next generation . Mutations were defined by their selective effects in males and females ( sm and sf , respectively ) and by their dominance value h . Mutations invading X but not autosomes and vice versa were identified as previously . Finally , we ran additional simulations for the aphid system ( with identical settings as for the core set of simulations , except for specified parameters ) to extend our range of parameters . First , we relaxed the assumption of equal dominance value in the three reproductive morphs ( by allowing ha≠hf≠hm ) since the dominance coefficient of a mutation might differ between sexes or morphs [3] . Second , we introduced a constraint between hi and si ( where i stands for a , f , m ) , so that beneficial alleles are dominant ( hi = 0 . 75 for si>0 ) , and deleterious ones , recessive ( hi = 0 . 25 for si<0 ) . Third , we simulated a mechanism of dosage compensation similar to mammals , by assuming a dominance coefficient of ha = hf = 0 . 5 for X-linked mutations in sexual and asexual females to model the random inactivation of one of the Xs . Fourth , we tested the influence of similar selective effects in sexual females and asexual females ( i . e . sa = sf , and ha = hf = hm ) since the phenotype of these two morphs are more similar compared to males . Finally , we analyzed the effect of the asexual phase length . The annual cycle was reduced to just one generation ( instead of 10 ) of asexual reproduction directly followed by one sexual generation . Gene expression level in the three reproductive morphs was estimated from RNA-Seq data ( Illumina , Illumina RNA-Seq protocol ) collected on whole body of males , asexual and sexual females from the LSR1 pea aphid reference clone . For this , aphids were reared on broad bean Vicia faba at low density ( less than five individuals per plant ) to prevent the production of winged morphs . Parthenogenesis was maintained under a 16 h photoperiod and a temperature of 18°C . Twenty asexual females were then directly frozen into liquid nitrogen and kept for subsequent RNA extractions . The production of sexual individuals was initiated by transferring larvae from a 16 h to a 12 h photoperiod at the same temperature of 18°C [79] . Two generations later , sexual females and males were observed . A total of 20 adult sexual females and 20 adult males were then directly frozen into liquid nitrogen . RNA extractions were then performed using the SV Total RNA Isolation System ( Promega ) according to manufacturer's instructions . For each reproductive morph , 4 separate RNA extractions of 5 adult individuals were performed , for a total of 12 RNA samples . RNA quality was checked on Bioanalyzer ( Agilent ) and quantified on Nanodrop ( Thermo Scientific ) . For each morph , two samples made of a pool of 2 µg of two of the four independent RNA extractions were generated , so that six RNA samples ( two samples for each morph ) were subsequently sent to GATC Company for RNA paired-end sequencing . RNA sample for two additional samples of male and asexual female of the LSR1 clone previously obtained using the same protocol and sequenced at the Baylor College of Medecine , USA ( available in AphidBase [80] and NCBI ) were also used . We thus have a total of eight RNA-Seq libraries , corresponding to three libraries for males , three for asexual females and two for sexual females ( see Table S3 ) . Reads from each library were mapped to the V2 assembly of the pea aphid genome using GSNAP [81] , after filtering for rRNA . Then we recorded the number of reads as a proxy for gene expression levels in the three reproductive morphs for all 36 , 990 predicted genes ( gene predictions 2 . 1 [82] ) . The numbers of mapped reads per library ranged from 12 to 22 millions ( Table S3 ) . We used the R package DESeq [83] to normalize the libraries ( default parameters ) and to identify genes showing significant biased expression between the three morphs , considering the different libraries for each morph as replicates . Significance for biased expression between reproductive morphs for each gene was calculated in DESeq . This was done by comparing two Generalized Linear models ( GLM ) , considering or not an effect of the reproductive morph factor on expression level of the gene ( this factor having three levels: male , sexual female , asexual female ) . If the inclusion of reproductive morph improved the model fit for a focal gene , we concluded that the morph significantly affected expression . Genes differentially expressed ( p<0 . 05 after adjusting for multiple testing using the Benjamini-Hochberg method implemented in the R package DESeq ) were then classified according to their pattern of expression: M+F−A− ( respectively M−F+A− and M−F−A+ ) stands for genes at least n-fold overexpressed in males [M] ( respectively asexual females [A] and sexual females [F] ) compared to each of the two other morphs . M−F+A+ ( respectively M+F−A+ and M+F+A− ) stands for genes at least n-fold under expressed in males ( respectively sexual females and asexual females ) compared to each of the two other morphs and with less than 2-fold difference in the two morphs in which it is overexpressed . This classification was performed for different threshold n of fold-change in expression ( with n = 2 , 5 and 10 ) . Among genes showing a non-significant bias in expression between the different reproductive morphs , we differentiated between those supported by very few reads ( <5 reads in total over the eight normalized libraries ) from those expressed at higher rates . Note that we worked on normalized expression data ( but not on expression per chromosome copy ) . We then restricted the following analyses to the subset of genes assigned to the autosomes or X-linked , following the approach described in [51] . Briefly , the primer sequences of 396 microsatellite loci previously assigned to the X ( 52 loci ) or to autosomes ( 344 loci ) [51] , plus six new X-linked loci identified from a linkage analysis in a pedigree of 250 individuals from 5 families ( Table S4 ) were mapped to the V2 genome assembly ( ∼24 , 000 scaffolds ) of the pea aphid ( available on AphidBase [80] ) . This allowed assigning 37 scaffolds to the X and 247 to the autosomes . Eleven additional scaffolds contained at least one microsatellite locus identified as X-linked and one located on the autosomes , indicating errors in the genome assembly ( this was observed in large scaffolds ) . The average distance between the closest X-linked and autosomal microsatellite loci assigned to the same scaffold was 543 kb ( min: 183 kb , max: 1900 kb ) . Since the probability of assembly errors increases with the size of the scaffolds , we collected only the predicted genes located in a window of 200 kb centered on each of the 402 microsatellite loci . By doing so , we obtained a tentative collection of 497 X-linked and 3215 autosomal genes . Only 14 . 4% of the microsatellite markers mapped to chromosomes were X-linked ( though the X represents ∼1/3 of the genome size [2n = 8 , 84] ) , and a similar proportion of genes were X-linked ( 13 . 4% ) . Non-random chromosome association ( X versus autosomes ) for genes with biased expression patterns ( i . e . M+F−A− , M+F+A− , M−F−A+ , M−F+A+ , M−F+A− or M+F−A+ ) was tested with Chi-square tests by comparing observed counts of X-linked and autosomal genes for each category of gene to the proportion expected under random association . This proportion was computed as the frequency of X-linkage for genes supported by at least 5 reads ( rather than to the percentage of X-linkage for the 3712 genes assigned to chromosomes ) because the X is slightly enriched with genes with low RNA-Seq support ( see Results ) . Finally , we conducted similar analyses on the genes located within a smaller window ( 100 kb window ) around the 402 markers used to tag regions of scaffolds as X-linked or autosomal , but also at larger windows ( 400 kb , 800 kb , no limitation of the size of the window , i . e . the whole scaffold is used ) to test whether our conclusions remained unaffected by window size . All genes ambiguously tagged as X-linked and autosomal ( because located on one of the 11 chimerical scaffolds and close to two microsatellite markers tagged to different types of chromosome ) were removed from the analyses . These analyses performed with sets of genes collected at different window sizes around the X vs autosomal tagged markers revealed that the contrast between the X chromosome and autosomes in their sex-biased gene content decreased with increasing window size ( See Results , Table S2 ) . These results , in addition to the direct evidence that 11 scaffold are chimerical between the X and autosomes , argue for the occurrence of some errors in the V2 genome assembly for large scaffolds . Indeed , such errors would lead to an increased proportion of incorrectly assigned genes to the X and to the autosomes at larger window sizes , hence to a decrease in the contrast between the X and autosomes . While this highlights the need to improve the assembly of the pea aphid genome , this does not affect our conclusions . First , the analyses presented in the Results section were performed on genes “close” to the microsatellite markers ( max 100 kb ) , a threshold chosen to minimize error of gene assignment but allowing sufficient statistical power . Second , any error ( by falsely assigning X-linked genes to autosomes and vice versa ) should only decrease the contrast between X and autosomes , and thus be conservative regarding our conclusions . To investigate for possible dosage compensation , raw expression data for each library was transformed into RPKM ( Reads Per Kilobase of exon model per Million mapped reads ) . Expression per gene per reproductive morph was computed as the mean expression over the two or three replicate libraries for each morph , and these data were then log2+1 transformed . Non-random chromosomal distribution of genes expressed at low rate ( those with <0 . 1 RPKM in total over the eight libraries ) was tested with a Chi-square test by comparing observed counts for autosomes and X chromosome to the frequency of X-linked genes ( f ( X ) = 0 . 134 ) . A difference in expression [log2 ( RPKM+1 ) ] between X-linked and autosomal genes within each morph was tested with Wilcoxon Rank Sum tests , considering different minimal thresholds for gene expression ( no restriction , RPKM>0 . 1 , >5 in total over the eight libraries ) . We also computationally doubled X-linked genes expression in males ( because aphid males have one X but two autosomal copies ) and tested similarly if ( doubled ) X-linked gene expression differed from expression of autosomal genes .
|
Males and females differ in their optimal values for most phenotypic traits , which makes intra-locus genetic conflicts among sexes common . Sex chromosomes have a sex-biased transmission , a pattern which might create favourable conditions for the spread of sexually antagonistic alleles ( i . e . alleles beneficial for one sex but deleterious for the other ) . Yet , expectations for genetic systems with unusual inheritance of sex chromosomes may differ from those derived from standard systems ( e . g . XY ) . Here we demonstrate theoretically that in organisms such as aphids , which alternate sexual and asexual reproduction and display an unusual inheritance of the X chromosome , male-beneficial sexually antagonistic alleles accumulate preferentially on that chromosome , while asexual female-beneficial alleles accumulate on autosomes . Theoretical models suggest that the evolution of sex-biased gene expression may solve such sexual conflicts , by restricting the product of a sexually antagonistic allele to the sex it benefits . We show that in the pea aphid , the genomic location ( X versus autosomes ) of genes with a sex-biased expression fits predictions derived from this hypothesis . On top of reporting a strong masculinization of the aphid X chromosome , our study highlights the relevance of organisms with an alternative mode of sex chromosome inheritance to understanding the evolutionary forces shaping chromosome evolution .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"genome",
"expression",
"analysis",
"mutation",
"sexual",
"selection",
"chromosomal",
"inheritance",
"chromosome",
"biology",
"natural",
"selection",
"sexual",
"conflict",
"heredity",
"gene",
"expression",
"genetics",
"genome",
"evolution",
"biology",
"genomics",
"evolutionary",
"biology",
"evolutionary",
"processes"
] |
2013
|
Masculinization of the X Chromosome in the Pea Aphid
|
While high-frequency deep brain stimulation is a well established treatment for Parkinson’s disease , its underlying mechanisms remain elusive . Here , we show that two competing hypotheses , desynchronization and entrainment in a population of model neurons , may not be mutually exclusive . We find that in a noisy group of phase oscillators , high frequency perturbations can separate the population into multiple clusters , each with a nearly identical proportion of the overall population . This phenomenon can be understood by studying maps of the underlying deterministic system and is guaranteed to be observed for small noise strengths . When we apply this framework to populations of Type I and Type II neurons , we observe clustered desynchronization at many pulsing frequencies .
High frequency deep brain stimulation ( DBS ) , a medical treatment in which high-frequency , pulsatile current is injected into an appropriate brain region , is a well established technique for alleviating tremors , rigidity , and bradykinesia in patients with Parkinson’s disease [1 , 2] . While the underlying mechanisms of deep brain stimulation remain unknown , it is well documented that local field potential recordings recorded in the subthalamic nucleus of patients with Parkinson’s disease display increased power in the beta range ( approximately 13–35 Hz ) [3–5] . These findings have led to the hypothesis that pathological synchronization among neurons in the basal ganglia-cortical loop contribute to the motor symptoms of Parkinson’s disease [6–8] . This hypothesis has been supported by findings that when DBS is applied to the STN , abatement of motor symptoms is correlated with a decrease the power in the beta band of the local field potential recorded from STN [9–11] . This line of thinking has led researchers to develop new strategies for desynchronizing populations of pathologically synchronized oscillators , [12–14] , some of which have shown promise as new treatment options for Parkinson’s disease in animal and human studies [15 , 16] . While many factors including the location of the DBS probe , stimulus duration , and stimulus magnitude influence the efficacy of DBS , one factor that is difficult to explain is the strong dependency on stimulus frequency . Low-frequency stimulation ( ≤ 50 Hz ) is generally ineffective at reducing symptoms of Parkinson’s disease while high-frequency stimulation from 70 to 1000 Hz and beyond has been shown to be therapeutically effective [17–19] . However , not all high frequency stimulation is equally effective , and clinicians have generally settled on a therapeutic range at about 130–180 Hz . [20 , 21] . In an effort to provide insight into the frequency dependent effects of DBS , the authors of [22] postulated that specifically tuned pulse parameters might yield chaotic desynchronization in a network of neurons . If desynchronization is the goal of DBS , then achieving it chaotically is a worthwhile objective . However , this can generally only be seen in a small window of pulse parameters and frequencies which may make it difficult to observe in real neurons . Furthermore , in both brain slices and in vivo recordings , individual neuronal spikes have been found to be time-locked to the external high-frequency stimulation [23–28] which would be unlikely if the spike times were chaotic . Here we present a different viewpoint showing that with high frequency pulsatile stimulation , in the presence of a small amount of noise , a population of neurons can split into distinct clusters , each containing a nearly identical proportion of the overall population . We find that the number of clusters , and hence desynchronization , is highly dependent the pulsing frequency and strength . We provide theoretical insight into this phenomenon and show that it can be observed over a wide range of pulsing frequencies and pulsing strengths . This viewpoint merges two seemingly contradictory hypotheses , showing that the therapeutic effect of the periodic pulsing could be to replace the pathological behavior with a less synchronous pattern of activity , even if individual neuronal spikes are phase locked to the DBS pulses .
Consider a noisy , periodically oscillating population of thalamic neurons from [29]: C V i ˙ = f V ( V i , h i , r i ) + I b + u ( t ) + ϵ η i ( t ) , h i ˙ = f h ( V i , h i ) , r i ˙ = f r ( V i , r i ) , i = 1 , … , N . ( 1 ) Here Vi , hi , and ri represent the transmembrane voltage and gating variables of neuron i , respectively , with all functions and parameters taken to be identical to those found in [29] , DBS pulses are represented by an external current u ( t ) , taken to be identical for each neuron , ηi ( t ) is a Gaussian white noise process , C = 1μF/cm2 is the constant neural membrane capacitance , Ib = 1 . 93μA/μF is a baseline current chosen so that in the absence of external perturbations and noise the firing rate is 60 Hz , and N is the total number of neurons . Using phase reduction , [30 , 31] , we can study Eq ( 1 ) in a more convenient form: θ i ˙ = ω + f ( θ i ) δ ( mod ( t , τ ) ) + ϵ η i ( t ) Z ( θ i ) + O ( ϵ 2 ) , i = 1 , … , N , ( 2 ) where θ ∈ [0 , 1 ) is the phase of the neuron with θ = 0 defined to be the time the neuron fires an action potential , ω is the natural frequency of oscillation , f ( θ ) is a continuous function which describes the effect of the DBS pulse , τ is a positive constant that determines the period of the DBS input , and Z ( θ ) is the neuron’s phase response curve to an infinitesimal perturbation . Here we assume that ϵ is small enough so that higher order noise terms are negligible ( c . f . [32 , 33] ) . Fig 1 shows an example charge-balanced pulsatile stimulus . We take the positive portion to be five times larger than the negative portion , with the positive current applied for 100 μs . The bottom panel shows the function f ( θ ) for a given stimulus intensity , calculated using the direct method [34]: a pulsatile perturbation is applied to a neuron at a known phase θp so that f ( θp ) can be inferred by measuring the timing of the next spike . We note that even though the DBS pulse itself is not a δ-function , it is of short enough duration that Eq ( 2 ) is an accurate approximation to Eq ( 1 ) . We simulate Eq ( 1 ) with 1000 neurons , taking a pulse strength S = 110μA/μF , and noise strength ϵ = 0 . 05 , for various pulsing frequencies , with results shown in Fig 2 . After some initial transients , we find the network tends to settle to a state with different numbers of clusters for different pulsing frequencies . From the probability distributions of neural phases ρ ( θ ) , the bottom panels show somewhat surprisingly that once the network settles to a clustered state , each cluster contains a nearly identical portion of the overall population . Also , upon reaching the steady distribution , neurons can still transition between clusters , but on average , the amount that leave a given cluster must be identical to the amount that enter . Fig 3 shows individual voltage traces for 50 sample neurons from this population after the network settles to a clustered state . Highlighted traces represent neurons from each cluster . In general , increasing the number of clusters will decrease synchrony in the population . Furthermore , neurons are more likely to transition between clusters as the overall number of clusters becomes larger . For simplicity of notation , we will take ω = 1 for Eq ( 2 ) in the theoretical analysis , but note that any other value could be considered to obtain qualitatively similar results . In the absence of noise , one may integrate Eq ( 2 ) for a single neuron θ to yield θ ( t ) = θ ( 0 ) + t , for t < τ , θ ( t ) = θ ( 0 ) + f ( θ ( 0 ) + τ ) + t , for τ ≤ t < 2 τ . ( 3 ) In this work , we are interested in the state of the system immediately after each pulsatile input . By integrating Eq ( 2 ) , the system dynamics can be understood in terms of compositions of a map θ ( n τ ) = g n ( θ 0 ) , n = 1 , 2 , … , ( 4 ) where g ( s ) = s + f ( s + τ ) + τ and g ( n ) denotes the composition of g with itself n times , and θ0 is the initial state of a neuron . In Eqs ( 3 ) and ( 4 ) , θ ( t ) and the arguments of f and g are always evaluated modulo 1 . If g ( n ) has a stable fixed point , then any oscillator which starts within its basin of attraction will approach that fixed point as time approaches infinity [35] . With noise , the dynamics are more complicated . In this case , the phase of each neuron cannot be determined exactly from Eq ( 2 ) , but rather , follows a probability distribution . For a neuron with known initial phase θ0 , after mτ has elapsed , the corresponding δ-function distribution δ ( θ − θ0 ) will be mapped to the Gaussian distribution ρ ( θ ) = N ( μ , ν ) , ( 5 ) with mean μ = g ( m ) ( θ0 ) given by Eq ( 23 ) and variance ν given by Eq ( 24 ) . In order to study the infinite time behavior of Eq ( 2 ) it can be useful to calculate steady state probability distributions for the population of neurons . To simplify the analysis , we will study Eq ( 2 ) as a series of stochastic maps applied to an initial phase density ( c . f . [22 , 33] ) , ρ ( θ , t + m τ ) = P m τ ρ ( θ , t ) , ( 6 ) where Pmτ is the linear Frobenius Perron operator corresponding to evolution for the time mτ , and ρ ( θ , t ) is the probability distribution of phases at time t . We can approximate Pmτ by the matrix P m τ ∈ R M × M by using eq ( 5 ) to determine each column of the matrix for a set of discretized phases , Δθ = 1/M . In Fig 4 for instance , the map g ( 2 ) yields the stochastic matrix P 2 τ , shown in the panel on the right . The delta function distribution ( arrow ) is mapped to a Gaussian distribution ( solid line ) . The matrix P m τ has all positive entries , and since probability is conserved , the matrix is column stochastic ( i . e . the columns of P m τ sum to 1 ) . For this class of matrices , the Perron-Frobenius theorem allows us to write [36 , 37] , lim k → ∞ P m τ k = v w T , ( 7 ) where v and w are the right and left eigenvectors associated with the unique eigenvalue of 1 , and normalized so that wT v = 1 . Recalling that P m τ is column stochastic , its left eigenvector associated with λ = 1 is 1T . Therefore , as the map is applied repeatedly , any initial distribution will approach a steady state distribution determined by v . We find that the existence of m fixed points of the underlying map g ( m ) ( θ ) provide the basis for the clustered desynchronization seen in Fig 2 , with a more formal main theoretical result given below . Consider the map g ( m ) with the following properties: Then for a given choice of ϵ1 ≪ 1 , we may choose ϵ small enough in eq ( 2 ) so that the distribution of phases will asymptotically approach a state with m distinct clusters , each containing 1 / m + O ( ϵ 1 ) of the total probability density . A proof of this statement is given in the Methods Section . In this detailed proof , we find that desynchronization can still be guaranteed even when g ( m ) is not monotonic as long as a more general set of conditions is satisfied . Note that because Eq ( 2 ) does not contain any coupling terms , noise will drive the system to a uniform , desynchronized , distribution in the absence of DBS input . In the sections to follow , we give numerical and theoretical evidence that clustered desynchronization can emerge in a population of pathologically synchronized neurons when the DBS pulses overwhelm the terms responsible for the synchronization . Using the main theoretical result , we can calculate regions of parameter space where we expect clustered desynchronization . The top-left panel of Fig 5 gives regions of parameter space where clustering is expected , giving the appearance of Arnold tongues [35] . White regions in the graph represent regions where either clustering is not guaranteed , or where we expect more than five clusters . However , we do not include these regions in the figure because they only exist for very narrow regions of parameter space . At around 60 Hz , the natural unforced period of the neural population , exactly one cluster is guaranteed , corresponding to 1:1 locking ( one DBS pulse per neural spike ) . This locking corresponds to a highly synchronous state , which we found when forcing the population at 63 Hz in Fig 2 . For pulsing frequencies between 80 and 120 Hz , we see prominent tongues corresponding to states with three , four and five clusters , which correspond to the states in Fig 2 where we force at 83 Hz and 94 Hz . A very wide tongue corresponding to 2:1 locking ( two DBS pulses per neural spike ) exists at frequencies ranging from 120 to 200 Hz , which is where DBS is often seen to be effective . Pulsing in this region manifests in the behavior seen with 120 Hz forcing in Fig 2 . To make comparisons with [22] we calculate the average Lyapunov exponent of the resulting steady state distributions using Eq ( 12 ) . For Lyapunov exponents greater than zero ( resp . , less than zero ) , the pulsatile stimulus will , on average , desynchronize ( resp . , synchronize ) neurons which are close in phase , and this has been proposed as an indicator of the overall desynchronization that might be observed in a population of neurons receiving periodic DBS pulses . The Lyapunov exponent is calculated for multiple parameter values for a system with a noise strength ϵ = 0 . 1 . Results are given in the bottom-left panel of Fig 5 . We note that results are not qualitatively different for different noise strengths . Compared with the Arnold tongues in the top-left panel , we find very narrow regions where the Lyapunov exponent is positive at relatively high stimulus strengths . The top-right panels show the steady state distribution for a population with pulses of strength S = 50μA/μF at a rate of 119 Hz ( resp . , 180 Hz ) corresponding to a two ( resp . , three ) cluster steady state . The bottom-right panels show the steady state distribution for a pulsing strength of S = 208μA/μF at 120 Hz and S = 206μA/μF at 290 Hz corresponding to regions with positive Lyapunov exponents . Even though the clustered states have very negative Lyapunov exponents , they show similar clustering behavior to the states with a positive Lyapunov exponent . However , the clustered desynchronization in the top-right panels can be accomplished using a significantly weaker stimulus and can be observed at a much wider range of pulsing parameters . Results thus far have focused on populations of neurons receiving homogeneous pulsatile inputs . However , it is well established that voltage fields vary significantly with distance from an external voltage probe [38] . In computational models such heterogeneity has been shown to create complicated patterns of phase locking that are dependent on the stimulation strength [39] and can improve methods designed to desynchronize large populations of neurons [14] . To understand the emergence of clustered synchronization when external inputs are different among neurons , we can modify the stochastic differential eq ( 2 ) as follows θ i ˙ = ω + f i ( θ i ) δ ( mod ( t , τ ) ) + ϵ η i ( t ) Z ( θ i ) + O ( ϵ 2 ) , i = 1 , … , N . ( 8 ) Here , fi ( θ ) is calculated based on the pulsatile input to each neuron . For each neuron , we use Eq ( 7 ) to calculate its steady state probability distribution . The state of each neuron is independent of the others , so that the average of the individual distributions gives an overall probability distribution for the population . As an illustrative example , we model 1000 neurons of the form Eq ( 1 ) receiving a charge balanced input of the same shape as in Fig 1 with τ = 1/ ( 140 Hz ) and S drawn from a normal distribution with mean 168 μA/μF and standard deviation 20 , giving values of S between approximately 100 and 240 . From the top-left panel of Fig 5 , this range of stimulus parameters is mostly , but not completely , contained in a two cluster region . g ( 2 ) ( θ ) is plotted in black in the top left panel of Fig 6 for a randomly chosen subset of these neurons with the identity line plotted in red for reference . The top-right panel shows each neuron’s steady state probability distribution ( calculated from its associated stochastic matrix ) in black for a noise strength of ϵ = 0 . 4 . While the main clustering results are guaranteed when the noise strength is small enough , we find that clustering can still occur at higher levels of noise . The steady state probability distribution in corresponding simulations of Eq ( 1 ) with heterogeneous pulsing strengths ( blue dashed curve ) agrees well with the theoretical probability density ( red dashed curve ) calculated from the average of each black curve in the top-right panel . The bottom panel shows corresponding cumulative distributions for the theoretical ( red ) and computationally observed ( blue ) probability densities highlighting that similar numbers of neurons are contained in each cluster . Similar clustering results can be obtained for different heterogeneous stimulus parameters . For instance , from Fig 5 , three-cluster behavior will emerge for pulsing frequencies of 200 Hz and stimulus strengths between approximately 90 and 170 μA/μF . Our main clustering results are for single population of neurons which do not explicitly take interactions between multiple populations of neurons into account , as is the case for the brain circuit responsible for Parkinsonian tremor . Here , we provide evidence that clustered desynchronization can still emerge when additional forcing terms are much smaller in magnitude than the external DBS pulses . Populations of coupled oscillators subject to common external forcing have been widely studied in the form of the forced Kuramoto model [40–42] . Synchronization can be observed when either external forcing or intrinsic coupling dominate the system dynamics . For intermediate coupling and external forcing strengths , a complicated bifurcation structure emerges and the macroscopic order parameter , describing the overall synchronization of the population , can oscillate . These behaviors have been observed in chemical oscillator systems [43] and have implications to externally forced biological rhythms such as circadian oscillations and neural oscillations [44] . We simulate Eq ( 1 ) with an additional external sinusoidal forcing frequency which could represent an aggregate input from a separate , unperturbed neural population . We note that this is not meant to represent a physiologically accurate model of DBS , but instead is meant to illustrate clustered desynchronization in the presence of a common periodic perturbation . Here , we take u ( t ) = 0 . 1 sin ( ωext t ) + uDBS ( t ) , where ωext is chosen so that the frequency of oscillation is the same as the natural period of the unperturbed neurons ( 60 Hz ) and uDBS ( t ) represents the common pulsatile input . For these simulations , N = 1000 . As we show in the Methods Section we may write this system in an identical form as Eq ( 2 ) , for which the main theoretical result still holds . For this particular example , clustering results are identical to those in the top left panel of Fig 5 . Results are shown with a pulsing strength S = 110μA/μF and noise strength of ϵ = 0 . 02 in Fig 7 . We find that the presence of a sinusoidal external stimulus is sufficient to synchronize the network in the absence of DBS forcing . When the DBS is turned on at both 83 and 94 Hz , we see three and four cluster states , respectively , just as we observed in the simulation without external forcing . However , in this simulation , the mean phase of each cluster varies with the external sinusoidal stimulation . Note that 120 Hz stimulation in this network also leads to two cluster desynchronization but is not shown . When neurons are synchronized through forcing that is not periodic , clustered desynchronization may still emerge when the DBS pulsing overwhelms the stimulation responsible for synchronization . As a second example , we model a network of neurons Eq ( 1 ) with an additional synaptic current , with each neuron’s transmembrane voltage dynamics taking the form C V i ˙ = f V ( V i , h i , r i ) + I b + u ( t ) + ϵ η i ( t ) + I i syn ( t ) . Here , I i syn ( t ) = K N ∑ k = 1 N ( V i - V G ) s k ( t ) ( 9 ) where K determines the magnitude of the synaptic current , VG is the reversal potential of a given neurotransmitter , and sk an additional synaptic variable associated with neuron k that evolves according to ( c . f . [30] ) s ˙ k = α 2 ( 1 - s k ) ( 1 / ( 1 + exp ( - ( V k - V T ) / σ T ) ) ) - β 2 s k , where α2 = 2 , VT = -37 , σT = 2 , and β2 = 0 . 1 . We simulate the resulting network withVg = 60mV , K = 0 . 015 and a noise strength of ϵ = 0 . 02; neurons form a single synchronized cluster in the absence of DBS input shown in panel B of Fig 8 . starting at t = 0 . 5 ms , we apply 180 Hz stimulation with S = 200μA/μF , the pulsing quickly overwhelms the synchronizing influence of the coupling , and the population splits into two separate clusters as shown by the probability densities in Panels A and individual voltage traces in panel C . When DBS is applied , we see from the average probability distributions and cumulative distributions in panels F and E , respectively that there are nearly equal proportions of neurons in each cluster . Other computational modeling [29] has suggested that pulsatile DBS may help regulate neural firing patterns , and help alleviate strongly oscillatory synaptic inputs . Panel D shows a similar phenemenon , when DBS is on , high amplitude oscillations in synaptic current are replaced by oscillations with a higher frequency but smaller amplitude . The desynchronization results here can be observed for many choices of parameters provided the pulsatile stimulation is significantly stronger than the synchronizing stimulation and that clustering behavior is expected in the absence of coupling . Consider a two dimensional reduction of the classic Hodgkin-Huxley equations [45] which reproduce the essential dynamical behavior [46]: C V i ˙ = f V H ( V i , n i ) + I b + u ( t ) + ϵ η i ( t ) , n i ˙ = f n ( V i , n i ) , i = 1 , … , N . ( 10 ) Here Vi and ni represent the transmembrane voltage and gating variables , respectively . All functions and parameters are identical to those given in [47] . DBS pulses are represented by the external current u ( t ) , which is given identically to each neuron , ηi ( t ) is a white noise process , C = 1μF/cm2 is the constant neural membrane capacitance , Ib = 10μA/μF is a baseline current yielding a firing rate of 84 . 7 Hz in the absence of external perturbation , and N is the total number of neurons . Unlike the model for thalamic neurons used in the main text , the Hodgkin-Huxley neuron displays Type II phase response properties , i . e . , a monophasic pulsatile input can act to either significantly increase or decrease the phase of the neuron . The top panel of Fig 9 shows an example monophasic stimulus which will be applied to the network Eq ( 10 ) at different strengths , S and periods τ . For this example , the pulse duration will be 100 μs , approximately , one percent of the neural firing rate . For this model , using our main theoretical results , we can also calculate regions of parameter space in which we expect to observe clustered desynchronization , with results shown in the middle panel of Fig 10 . The Arnold tongues for clustering greater than five become quite narrow and are not included in this figure . We also calculate the average Lyapunov exponent for the steady state distribution using eq ( 12 ) from the main text for a noise strength of ϵ = 0 . 15 , with results shown in the bottom panel . We note that unlike for the thalamic neurons , the Lyapunov exponent for the Hodgkin-Huxley network is never positive . We find that regions with the lowest Lyapunov exponents tend to correlate with regions where clustered desynchronization is guaranteed . Even though the Lyapunov exponent might be quite negative , the steady state distribution can still be relatively desynchronized if there are a large number of clusters , as evidenced by the four cluster state in the top panel . Finally , we simulate Eq ( 10 ) with N = 1000 neurons with a pulse strength S = 10μA/μF and ϵ = 0 . 3 for pulsing frequencies that are expected to yield clustered desynchronization determined from Fig 10 . Results are shown in Fig 11 . We find clustered states begin to form almost immediately , and in the bottom panel , after the system has approached the steady state distribution , each cluster contains an approximately identical proportion of the population .
While deep brain stimulation is an important treatment for patients with medically intractable Parkinson’s disease , its fundamental mechanisms remain unknown . Making matters more complicated , experimental studies have shown that the symptoms of Parkinson’s can be alleviated using strategies that seek to desynchronize a population of pathologically synchronized oscillators [15 , 16] , while other seemingly contradictory studies have shown that neurons have a tendency to time-lock to external high-frequency pulses [23–27] , supporting the hypothesis that entrainment is necessary to replace the pathological neural activity in order to alleviate the symptoms of Parkinson’s disease . In this work we have have shown that these two phenomenon may be happening in concert: in the presence of a small amount of noise , high frequency pulsing at a wide range of frequencies is expected to split a larger population of neurons into subpopulations , each with a nearly equal proportion of the overall population . The number of subpopulations , and hence the level of desynchronization , is determined by phase locking relationships which can be found by analyzing the phase reduced system in the absence of noise . We note that other theoretical [12] and experimental [16][15] work has yielded control strategies that are specifically designed to split a pathologically synchronized neural population into distinct clusters . The theory presented in this paper suggests that clinical DBS may be accomplishing the same task with a single probe . The conditions we have developed guarantee clustered desynchronization for small enough noise , but we do not give any a priori estimate of how small the noise needs to be so that distinct clusters can be observed . If the noise is too large , the clusters may start to merge into one another , particularly when there are a large number of clusters ( see the bottom left panel of Fig 2 ) . Even in this case , however , we still have discernible clusters , throughout which the overall population of neurons is spread relatively evenly . We also note that this theory does not give any estimates on the time it takes for the system to achieve its steady state population distribution , but this can be calculated for a specific population by examining the second smallest eigenvalue in magnitude , λ2 , of a given stochastic matrix P m τ ( c . f . [33] ) . As λ2 becomes closer to 1 , more iterations of the map P m τ will be required for the transient dynamics to die out , and it will take longer for the system to approach the steady state distribution . In general , we find that for a given map , λ2 becomes closer to one as noise strength decreases , which is consistent with the notion that the average escape time between clusters will increase as the strength of the external noise decreases [48] . For the networks that we have investigated , regions of parameter space which are associated with either clustered desynchronization or positive Lyapunov exponents can display similar levels of desynchronization . However , numerics show that the regions with positive Lyapunov exponents are quite small and may be difficult to find without explicit calculation . In contrast , the regions of parameter space associated with clustered desynchronization are fairly large and are likely to be observed without knowledge of the system properties . If desynchronization of the overall population is an important mechanism of high-frequency DBS , doing so chaotically may be an overly restrictive objective if clustered desynchronization is sufficient to alleviate the motor symptoms of Parkinson’s disease . This study is certainly not without limitations . For instance , the computational neurons considered in this study are based on simple , low-dimensional models of neural spiking behavior . However , we have developed the theory to understand the clustered desynchronization phenomenon in such a way that it can easily be extended to more complicated neural models with more physiologically detailed dynamics provided the neural phase response properties can be measured experimentally in vivo[49] . Furthermore , while we only considered homogeneous populations in this study , the phase response properties and natural frequencies of a living population of neurons will surely have a heterogeneous distribution . In this context , we could still show that clustered desynchronization is expected by applying the theory developed in this work to a family of neurons with different phase response properties and natural frequencies . The expected steady state population could then be obtained as a weighted average of the individual steady state distributions . Numerical results presented here apply to networks for which external DBS perturbations overwhelm the intrinsic coupling between neurons . In this work , we have not considered the complicated interplay between multiple populations of neurons which give rise to the symptoms of Parkinson’s disease; more detailed modelling studies would be required to determine the effect of clustered desynchronization on the overall network circuit . Others have studied synchrony and clustering behavior in coupled populations of neurons [50–52] and it is possible that our results could be extended to describe clustering for weaker pulsatile stimuli when coupling cannot be neglected . Our results suggest that high-frequency external pulsing could have the effect of separating a neural population into equal subpopulations in the presence of noise . This viewpoint could help explain the frequency dependent nature of therapeutically effective DBS and could help merge competing hypotheses , as desynchronization and entrainment are not mutually exclusive when even small amounts of noise are present . If clustered desynchronization does provide a mechanism by which the motor symptoms associated with Parkinson’s disease can be mitigated , it could provide a useful control objective for designing better open-loop DBS stimuli in order to prolong battery life of the implantable device and to mitigate potential side effects of this therapy .
To make comparisons with [22] we calculate the average Lyapunov exponent of the resulting steady state distributions , giving a sense of whether , on average , the orbits of the trajectories oscillators from Eq ( 2 ) are converging or diverging . For instance , let ϕ denote the phase difference between oscillators θ1 and θ2 which are close in phase , i . e . ϕ ( t ) ≡|θ1 ( t ) − θ2 ( t ) | . Then from Eq ( 3 ) , immediately after a DBS pulse occuring at time τ , ϕ ( τ + ) = |f ( θ 1 ( τ - ) ) + θ 1 ( τ - ) - f ( θ 2 ( τ - ) ) - θ 2 ( τ - ) | , = |f ( θ 2 ( τ - ) ) + f ′ ( θ 2 ( τ - ) ) ϕ ( τ - ) + O ( ϕ ( τ - ) 2 ) + θ 1 ( τ - ) - f ( θ 2 ( τ - ) ) - θ 2 ( τ - ) | , = ϕ ( τ - ) |1 + f ′ ( θ 2 ( τ - ) ) | + O ( ϕ ( τ - ) 2 ) , ( 11 ) where ′ ≡ d/dθ and θ ( τ− ) ( resp . θ ( τ+ ) ) denotes the limit of θ ( t ) as t approaches τ from below ( resp . above ) . Note that in the second line , we have used a Taylor expansion of f about θ2 for small values of ϕ ( τ− ) . Therefore , the oscillators converge or diverge locally depending upon the derivative of f . For a population of neurons , the stochastic matrix P τ for a given pulsing rate can be used to determine the steady state distribution ρ* ( θ ) before each pulse , with an average Lyapunov exponent taken to be ( c . f . [22] ) : LE = ∫ 0 1 ρ * ( θ ) log [ 1 + f ′ ( θ ) ] d θ . ( 12 ) For LE > 0 ( resp . , LE < 0 ) , the pulsatile stimulus will , on average , desynchronize ( resp . , synchronize ) neurons which are close in phase , and this has been proposed as an indicator of the overall desynchronization that might be observed in a population of neurons receiving periodic DBS pulses . For a single neuron with a known initial phase θ evolving according to the stochastic differential Eq ( 2 ) , we calculate the expected value and variance with a strategy that is similar to the one employed in [33] . We first asymptotically expand θ ( t ) in orders of ϵ , θ ( t ) = θ 0 ( t ) + ϵ θ 1 ( t ) + … , ( 13 ) with θ0 ( 0 ) = θ ( 0 ) , and θ1 ( 0 ) = 0 . Substituting Eq ( 13 ) into Eq ( 2 ) and taking ω = 1 for simplicity of notation , allows us to write θ ˙ 0= 1 + f ( θ 0 ( t ) ) δ ( mod ( t , τ ) ) , ( 14 ) θ ˙ 1= η ( t ) Z ( θ 0 ( t ) ) + f ′ ( θ 0 ( t ) ) θ 1 ( t ) δ ( mod ( t , τ ) ) . ( 15 ) Integrating eq ( 14 ) for a time τ yields , θ 0 ( τ - ) = θ 0 ( 0 ) + τ , θ 0 ( τ + ) = θ 0 ( 0 ) + f ( θ 0 ( 0 ) + τ ) + τ . ( 16 ) This relationship can be used to write θ0 in terms of compositions of a map: θ 0 ( t ) = g ⌊ t τ⌋ ( θ 0 ( 0 ) ) + mod ( t , τ ) , ( 17 ) where g ( s ) = s + f ( s + τ ) + τ and g ( n ) denotes the composition of g with itself n times . In Eqs ( 16 ) and ( 17 ) , θ ( t ) and the arguments of f and g are always evaluated modulo 1 . We now focus our attention on θ1 . Integrating eq ( 15 ) yields θ 1 ( t ) = ∫ 0 t η ( s ) Z ( θ 0 ( s ) ) d s + ∑ m = 1 ∞ f ′ ( θ 0 ( m τ - ) ) θ 1 ( m τ - ) H ( t - τ m ) , ( 18 ) where H ( ⋅ ) is the Heaviside step function . Note here that θ1 ( mτ− ) denotes the limit of θ1 ( t ) as t approaches mτ from below . In the interval 0 ≤ t < τ , we note that θ 1 ( t ) = ∫ 0 t η ( s ) Z ( θ 0 ( s ) ) d s , so that θ 1 ( τ - ) = ∫ 0 τ η ( s ) Z ( θ 0 ( s ) ) d s . With this in mind , we can rewrite Eq ( 18 ) as θ 1 ( t ) = ∫ 0 t η ( s ) Z ( θ 0 ( s ) ) d s + f ′ ( θ 0 ( τ - ) ) ∫ 0 τ η ( s ) Z ( θ 0 ( s ) ) d s = 1 + f ′ ( θ 0 ( τ - ) ) ∫ 0 τ η ( s ) Z ( θ 0 ( s ) ) d s + ∫ τ t η ( s ) Z ( θ 0 ( s ) ) d s for τ ≤ t < 2 τ . ( 19 ) Likewise , using Eq ( 19 ) to determine θ1 ( 2τ− ) allows us to write θ 1 ( t ) = 1 + f ′ ( θ 0 ( τ - ) ) + f ′ ( θ 0 ( 2 τ - ) ) 1 + f ′ ( θ 0 ( τ - ) ) ∫ 0 τ η ( s ) Z ( θ 0 ( s ) ) d s + 1 + f ′ ( θ 0 ( 2 τ - ) ) ∫ τ 2 τ η ( s ) Z ( θ 0 ( s ) ) d s + ∫ 2 τ t η ( s ) Z ( θ 0 ( s ) ) d s for 2 τ ≤ t < 3 τ . ( 20 ) This process can be repeated indefinitely to find θ 1 ( m τ + ) = X 0 m ∫ 0 τ η ( s ) Z ( θ 0 ( s ) ) d s + X 1 m ∫ τ 2 τ η ( s ) Z ( θ 0 ( s ) ) d s + ⋯ + X m - 1 m ∫ ( m - 1 ) τ m τ η ( s ) Z ( θ 0 ( s ) ) d s , ( 21 ) where X i m is defined recursively so that X i m = 0 if m < i , 1 if m = i , X i m - 1 + f ′ ( θ 0 ( m τ - ) ) X i m - 1 if m > i . ( 22 ) Because Eq ( 21 ) is the sum of stochastic integrals which themselves are normally distributed random variables , θ1 ( mτ+ ) will also be a normally distributed random variable [53] . Ultimately , we are interested in calculating the expected value and variance of a neuron starting at θ ( 0 ) after the application of m DBS inputs . Using the asymptotic expansion Eq ( 13 ) , we can calculate the expected value of θ ( mτ+ ) as E [ θ ( m τ + ) ] = E [ θ 0 ( m τ + ) + ϵ θ 1 ( m τ + ) + O ( ϵ 2 ) ] . Using the relations Eqs ( 17 ) and ( 21 ) and noting that the noise η ( s ) has a mean of zero , E[θ1 ( mτ+ ) ] = 0 , and hence , the expected value , μ , is μ ≡ E [ θ ( m τ + ) ] = θ 0 ( m τ + ) = g m ( θ ( 0 ) ) + O ( ϵ 2 ) . ( 23 ) Again using Eqs ( 17 ) and ( 21 ) , we can calculate the variance , ν , of θ ( mτ+ ) to leading order ϵ2: ν ≡ E ( θ ( m τ + ) - E [ θ ( m τ + ) ] ) 2 = E [ ( ϵ θ 1 ( m τ + ) ) 2 ] = ϵ 2 ( X 0 m ) 2 ∫ 0 τ [ Z 2 ( θ 0 ( s ) ) ] d s + ( X 1 m ) 2 ∫ τ 2 τ [ Z 2 ( θ 0 ( s ) ) ] d s + ⋯ + ( X m - 1 m ) 2 ∫ ( m - 1 ) τ m τ [ Z 2 ( θ 0 ( s ) ) ] d s . ( 24 ) Note that in the last line , when squaring θ1 ( mτ+ ) , the fact that any noise processes that do not overlap in time are uncorrelated allows us to eliminate terms nonidentical noise processes . Also note that for identical white noise processes , E[η ( s ) η ( s ) ] = δ ( 0 ) . Suppose that conditions 1 and 2 from our main theoretical results are satisfied for g ( m ) with m stable fixed points . Consider the corresponding stochastic matrix P m τ . Denote the stable and unstable fixed points as θsi and θui , respectively . To begin , it will be convenient to define g ( m ) ( θ ) = θ + F ( θ ) ( 25 ) so that F ( θ ) ≡ g ( m ) ( θ ) - θ . With this definition , F ( θ ) = 0 corresponds to the fixed points of the map g ( m ) ( θ ) . We subdivide θ ∈ [0 , 1 ) into into 4m disjoint subregions with the following procedure: Choose α > 0 and define the region si near each stable fixed point so that s i = [ θ s i - δ s i - , θ s i + δ s i + ] , where α = F ( θ s i - δ s i - ) = - F ( θ s i + δ s i + ) . Likewise , define the region ui near each unstable fixed point so that u i = [ θ u i - δ u i - , θ u i + δ u i + ] , where α = - F ( θ u i - δ u i - ) = F ( θ u i + δ u i + ) . Define the remaining regions n i + = ( θ s i + δ s i + , θ u i - δ u i - ) and n i - = ( θ u i - 1 + δ u i - 1 + , θ s i - δ s i - ) . See Figs 12 and 13 for an example of how the matrix P m τ is partitioned into submatrices according to this procedure . In the analysis to follow we will show that the steady state probability distribution v ≡ lim k → ∞ P m τ k ρ ( θ , 0 ) exists and is invariant to ρ ( θ , 0 ) . We define subregions of v a , a = s i , u i , n i + , n i - to represent the subset of v contained in the region a . We have carefully defined the matrix partition in Fig 13 so that , for instance , | | P m τ s i → n i + v s i | | 1 corresponds to the amount of probability that is mapped from the region si into to the region n i + when P m τ is applied to v . This relation results because all entries of P and v are positive . See Fig 14 for a visual representation of this probability mapping process which is central to the proof to follow . Suppose that there exists α for the resulting partition such that , Note here that for a given set p , p ˚ denotes its interior , and p ¯ denotes its closure . Furthermore , we will also assume , without loss of generality , that in the absence of noise , upon successive iterations of the map g ( 1 ) the period m orbit is θs1 → θs2 → ⋯ → θsm → θs1 . Let γ = mod ( i , m ) + 1 . Suppose then that Then for any choice of ϵ1 ≪ 1 , we may choose ϵ ( the noise strength ) small enough in eq ( 2 ) , so that as time approaches infinity , regardless of initial conditions , the population will be split into m distinct clusters , and to leading order in ϵ1 , each cluster will contain an equal portion of the population . We note that if conditions 1 and 2 from our main theoretical result are satisfied and g ( m ) is monotonic , then we will be guaranteed to be able to choose and α so that the remaining conditions 3a-d are satisfied . In order to prove the main theoretical result presented earlier , we will first show that for ϵ small enough , as time tends toward infinity , the probability of finding an oscillator far from any of the stable fixed points is O ( ϵ 1 ) . We will then show that as time approaches infinity , the chance of finding a randomly chosen oscillator near any of the stable fixed points is identical to leading order ϵ1 . Throughout this proof , we are interested in the unique steady state solution which solves v = P τ v . We note that for any positive integer k , P k τ = P τ k so that v = P k τ v , i . e . the unique steady state solution of P k τ is the same for any choice of k . Using Eqs ( 23 ) and ( 24 ) , we will assume that ϵ is taken small enough so that errors in the approximations of all necessary stochastic matrices P k τ are negligible . We consider a modified version of Eq ( 2 ) where each neuron feels a small , common periodic perturbation θ ˙ i = ω + ϵ f ( θ i ) δ ( mod ( t , τ ) ) + ϵ p ( ω 1 t ) Z ( θ i ) + ϵ η i ( t ) Z ( θ i ) + O ( ϵ 2 ) , i = 1 , … , N . ( 41 ) Here , p ( ω1 t ) is a periodic perturbation with period T1 = 1/ω1 common to each oscillator . This perturbation may represent the effect of coupling from an external population of neurons or coupling between neurons in the population under study . Here we give conditions for which Eq ( 41 ) exhibits clustered desynchronization . Defining ϕi ≡ θi − ω1 t allows us to selectively average the term associated with p ( ω1 t ) from Eq ( 41 ) [54] , c . f . [55]: ϑ ˙ i = ω + ϵ f ( ϑ i ) δ ( mod ( t , τ ) ) + ϵ G ( ϑ i - ω 1 t ) + ϵ η i ( t ) Z ( ϑ i ) + O ( ϵ 2 ) , ( 42 ) where G ( φ ) = ∫ 0 T 1 [ p ( ω 1 t ) Z ( φ i + ω 1 t ) ] d t and φi = ϑi − ω1 t . Here ϑi is a close approximation to θi so that φi ≈ ϕi . Noting that G is also T1 periodic , we selectively average Eq ( 42 ) to yield Θ ˙ i = ω + ϵ K + ϵ f ( Θ i ) δ ( mod ( t , τ ) ) + ϵ η i ( t ) Z ( Θ i ) + O ( ϵ 2 ) , ( 43 ) Here , Θi ≈ ϑi and K = ∫ 0 T 1 G ( φ i - ω 1 t ) d t . Notice that Eq ( 43 ) is in an identical form as Eq ( 2 ) for which our main theoretical result still holds .
|
While high-frequency deep brain stimulation ( DBS ) is a decades old treatment for alleviating the motor symptoms Parkinsons disease , the way in which it alleviates these symptoms is not well understood . Making matters more complicated , some experimental results suggest that DBS works by making neurons fire more regularly , while other seemingly contradictory results suggest that DBS works by making neural firing patterns less synchronized . Here we present theoretical and numerical results with the potential to merge these competing hypotheses . For predictable DBS pulsing rates , in the presence of a small amount of noise , a population of neurons will split into distinct clusters , each containing a nearly identical proportion of the overall population . When we observe this clustering phenomenon , on a short time scale , neurons are entrained to high-frequency DBS pulsing , but on a long time scale , they desynchronize predictably .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[] |
2015
|
Clustered Desynchronization from High-Frequency Deep Brain Stimulation
|
Recent advances in single-cell time-lapse microscopy have revealed non-genetic heterogeneity and temporal fluctuations of cellular phenotypes . While different phenotypic traits such as abundance of growth-related proteins in single cells may have differential effects on the reproductive success of cells , rigorous experimental quantification of this process has remained elusive due to the complexity of single cell physiology within the context of a proliferating population . We introduce and apply a practical empirical method to quantify the fitness landscapes of arbitrary phenotypic traits , using genealogical data in the form of population lineage trees which can include phenotypic data of various kinds . Our inference methodology for fitness landscapes determines how reproductivity is correlated to cellular phenotypes , and provides a natural generalization of bulk growth rate measures for single-cell histories . Using this technique , we quantify the strength of selection acting on different cellular phenotypic traits within populations , which allows us to determine whether a change in population growth is caused by individual cells’ response , selection within a population , or by a mixture of these two processes . By applying these methods to single-cell time-lapse data of growing bacterial populations that express a resistance-conferring protein under antibiotic stress , we show how the distributions , fitness landscapes , and selection strength of single-cell phenotypes are affected by the drug . Our work provides a unified and practical framework for quantitative measurements of fitness landscapes and selection strength for any statistical quantities definable on lineages , and thus elucidates the adaptive significance of phenotypic states in time series data . The method is applicable in diverse fields , from single cell biology to stem cell differentiation and viral evolution .
Selection is a process in which the interaction of organisms with their environment determines which types of individuals thrive and proliferate more than others . Genetic information encoded in the genome is a primary determinant of reproductivity , but epigenetic and fluctuating phenotypic traits can also strongly influence selection [1–4] . Recent single-cell measurements revealed the existence of phenotypic heterogeneity within clonal populations , including cases in which heterogeneity has been shown to have a clear functional role [5 , 6] such as bacterial persistence [7–9] , infection [10] , and competence and sporulation [11] . Quantifying reproductivity of phenotypic traits and revealing how strongly selection acts within a clonal population are thus of crucial importance for understanding the biological significance of phenotypic heterogeneity . To experimentally evaluate reproductivity of a unicellular organism , one usually measures bulk growth rate ( Malthusian parameter [12] ) of a cellular population in batch or uses a competition assay between a genotype of interest and a reference genotype [13] . These methods are only valid when the time-scale of genotypic changes is sufficiently long compared with that of the measurements . However , the time-scale of phenotypic changes is often comparable to cellular generation time ( [14] and S1 Fig ) , and only in certain cases is it orders of magnitude longer , e . g . when a phenotypic state is stabilized by specific epigenetic and/or positive-feedback regulations . As a result , bulk population growth rates of sub-populations fractionated based on initial phenotypic traits , e . g . by fluorescence-activated cell sorting , do not necessarily represent reproductivity of initial phenotypic traits because phenotypic traits are diversified rapidly by complex dynamical processes that occur during measurements . An alternative approach is necessary to measure reproductivity for heterogeneous and fluctuating cellular phenotypes . Using time-lapse microscopy and fluorescent reporters , it has become possible to follow full individual cell histories recording all division events and instantaneous expression levels of reporters within cellular populations [8 , 15–20] . Several theoretical studies have demonstrated the utility of history-based analysis of growing populations , regarding individual histories rather than single cells as the basic replicating entity [21–23] . For example , Leibler and Kussell introduced a time-integrated instantaneous reproduction rate , termed historical fitness [21] , and defined a measure of selection using the response of mean historical fitness over all histories within a population . However , empirically determining the instantaneous reproduction rate of an individual cell can be difficult in general , e . g . due to the fact that cell size , age , elongation rate , and division timing are a subset of possible observables all of which contribute to reproduction . Evaluating the fitness value of a certain phenotypic trait such as expression level of a specific gene results in additional complications . To address these difficulties , we introduce an empirically measurable quantity associated with phenotypic states , which we call the phenotypic fitness landscape . This quantity , which reports how cellular reproductivity is correlated with phenotypic states , extends the definition of historical fitness so that it becomes meaningful in a general setting without requiring any assumptions . Our approach allows one to assign a fitness value to any statistical quantities observed over cellular lineages , and to evaluate the selection strength acting on different phenotypic states . Although it does not imply causal relationships driving selection , our measure of selection strength quantifies correlations between phenotypic states and fitness . To formulate our framework , we leverage a fundamental property of selection processes: the retrospective probability of observing a certain phenotypic trait value by moving backward in time from the present to the ancestral parts of a lineage is different from its counterpart , the chronological probability to observe the trait value moving forward in time along a lineage as individuals grow and divide . We show that these two probabilities can be evaluated directly using single-cell lineage tree data , leading to natural definitions of the fitness landscape and selection strength . We apply this framework to analyze proliferation processes of simulated and experimental cellular populations , demonstrating the utility of our measures to reveal phenotype dependent fitness and its response to environmental change .
We first present an overview of the type of biological data that is examined in this study . Fig 1A shows an example of time-lapse images of E . coli growing on an agarose pad . Analyzing the images provides information on the proliferation dynamics and phenotype transitions ( Fig 1B–1D ) . For example , one can reveal the lineage tree structure , which shows genealogical relationships among the individual cells that originated from an ancestor cell ( Fig 1B ) . One can also extract the information on the transitions of cell size along cell lineages ( Fig 1C ) and of other cellular phenotypes such as intracellular concentration of a particular protein if it can be probed with an appropriate fluorescence reporter ( Fig 1D ) . Here we regard any measurable quantity or set of quantities observed along cell lineages as a quantifiable phenotype in the broadest sense . We note that this type of information is now available for many biological processes including embryogenesis and stem cell differentiation [19 , 20 , 24–27] . As shown in Fig 1B , division intervals ( i . e . cell cycle durations ) of individual cells are usually heterogeneous , yielding variability in the number of divisions observed along different cell lineages . Moreover , cellular phenotypes also fluctuate along cell lineages ( Fig 1C and 1D ) . To understand the role of such phenotypic heterogeneity for population growth , one must know whether the difference of the phenotypic states are correlated with the growth rate heterogeneity within a population . Below we present the theoretical results and the analytical procedure that allow us to reveal the quantitative relations between phenotypes and growth . We consider a binary division process as depicted in Fig 2 , where t0 , t1 are the start and end times of a lineage tree , and we define τ = t1 − t0 as the duration of observation . To illustrate our view of lineage statistics , we first consider a single fixed lineage tree denoted by T derived from a single ancestor cell ( Fig 2A ) . Let N ( t , T ) be the number of cells in the tree T at time t and we label and distinguish each lineage by i = 1 , 2 , . . . , N ( t 1 , T ) . For the tree T in Fig 2A , N ( t 1 , T ) = 13 . We consider two different ways of randomly sampling single-cell lineages on the tree . We could sample each lineage with equal weight , where the probability of choosing lineage i is π i rs = 1 / N ( t 1 , T ) , which we call the retrospective probability because it corresponds to the probability that the past history of the last cell on lineage i is chosen . For the tree T , π i rs = 1 / 13 for all i . Alternatively , letting Di be the number of cell divisions on lineage i , we could sample lineage i with probability π i cl = 2 - D i , which we call the chronological probability because it is the probability that lineage i is chosen by descending the tree from the ancestor cell at t0 randomly at each branch point with equal probability 1/2 . For example , π 3 cl = 2 - 3 = 1 / 8 and π 9 cl = 2 - 5 = 1 / 32 for the tree T ( Fig 2A ) . The probability distribution π i cl is determined solely by the number of divisions on lineage i , being unaffected by the reproductive performance of the other lineages . In contrast , π i rs strongly depends on the reproductive performance of the other lineages , which enters into the total number of cell lineages N ( t 1 , T ) . Generally , the ratio π i rs / π i cl is positively correlated with the relative reproductive performance of lineage i . In fact , π 9 rs / π 9 cl > π 3 rs / π 3 cl for T . In addition , we note that the inconsistencies between π i rs and π i cl reflect the variability in the numbers of divisions among the cell lineages . It can be confirmed that π i rs = π i cl for all i if the same number of divisions occur for all the lineages on a tree ( Fig 2B ) . The consideration above indicates that the cell lineages with more divisions are over-represented in the retrospective probability relative to its chronological probability . This idea can be further extended to the general situation where a large number of lineage trees are contained in the population . We now consider the set of lineages within a large collection of independent trees initiated from a large number of progenitor cells N ( t0 ) ≫ 1 ( Fig 2C ) . For each lineage , we record a phenotypic trait x and the number of divisions D , where x can be any random variable representing a phenotypic trait of a single cell lineage , e . g . a time-averaged gene expression level , average cell length , number of divisions D , or any variety of other possibilities . We consider the joint statistics of D and x across all possible trees , letting n ( D , x , T ) denote the number of lineages with values D and x within tree T , and we denote the sum of this quantity over trees as n ( D , x ) . The total number of lineages observed across all trees , N ( t1 ) , is given by summing n ( D , x ) over D and x ( see S1 Text ) . In analogy with the single tree quantities , we define the retrospective probability of choosing a lineage with D and x as P rs ( D , x ) ≡ n ( D , x ) N ( t 1 ) , ( 1 ) and the chronological probability as P cl ( D , x ) ≡ 2 - D N ( t 0 ) n ( D , x ) . ( 2 ) Defining Λ to be the population growth rate , Λ ≡ 1 τ ln N ( t 1 ) N ( t 0 ) , ( 3 ) we obtain using Eqs 1 and 2 the relation P rs D , x = e τ h ˜ ( D ) - Λ P cl D , x , ( 4 ) where h ˜ ( D ) ≡ τ - 1 D ln 2 . We see from Eq 4 that h ˜ ( D ) is the natural measure of fitness for a lineage , since lineages for which this quantity is greater than Λ will be exponentially over-represented in retrospective probability relative to chronological probability . We call h ˜ ( D ) the lineage fitness . We now measure how quickly the number of lineages with a given phenotype x grow between times t0 and t1 according to their chronological and retrospective probabilities . We denote by Prs ( x ) ≡ ∑D Prs ( D , x ) and Pcl ( x ) ≡ ∑D Pcl ( D , x ) the retrospective and chronological marginal probability distributions of x . We define the phenotypic fitness landscape h ( x ) as h x ≡ 1 τ ln N ( t 1 ) P rs ( x ) N ( t 0 ) P cl ( x ) = Λ + 1 τ ln P rs ( x ) P cl ( x ) , ( 5 ) noting that N ( t0 ) Pcl ( x ) and N ( t1 ) Prs ( x ) are the effective numbers of cell lineages with a phenotypic trait x from the chronological and retrospective perspectives , respectively . We can rewrite Eq 5 as P rs x = e τ h ( x ) - Λ P cl x , ( 6 ) which shows that if h ( x ) is greater than Λ the phenotypic state x will be exponentially over-represented in retrospective relative to chronological probability . Thus , h ( x ) provides a natural extension of fitness for lineage-based phenotypic traits . We point out that both Prs ( D , x ) and Pcl ( D , x ) ( hence , Prs ( x ) and Pcl ( x ) as well ) are obtainable directly from the set of lineage trees ( Fig 2C ) . Thus , h ( x ) can also be determined directly from the lineage tree data using Eq 5 . In Fig 3 , we schematically show how the fitness landscapes look depending on the deviation between chronological and retrospective probability distributions . When the deviation is small , the fitness landscape h ( x ) is flat over the phenotypic space x ( Fig 3A ) ; when the deviation is large , h ( x ) changes greatly depending on x ( Fig 3B ) . In the next section , we quantify the amount of heterogeneity in h ( x ) using the selection strength , S[x] , defined below . Specific states of the phenotypic trait x can be selected if x and D are correlated . In general , the strength of this correlation could differ significantly among different phenotypes . In the conventional framework of natural selection known as Fisher’s fundamental theorem , selection strength is measured by the gain of mean fitness due to the change of probability distribution of a phenotype [12] . Inspired by this idea , we define the strength of selection acting on a phenotypic trait x as S x ≡ ⟨ h ( x ) ⟩ rs - ⟨ h ( x ) ⟩ cl , ( 7 ) where 〈h ( x ) 〉rs = ∑x h ( x ) Prs ( x ) and 〈h ( x ) 〉cl = ∑x h ( x ) Pcl ( x ) are the mean fitness in retrospective and chronological perspectives , respectively . This simple measure of selection strength has rich underpinnings . First , S[x] is also a measure of fitness variation on the landscape h ( x ) because S x ≈ τ Cov h ˜ ( D ) , h ( x ) ≈ τ Var h ( x ) , ( 8 ) where the variance and covariance can equivalently be taken over either chronological or retrospective distributions , and the approximation is accurate to the order of second cumulants of h ˜ ( D ) and h ( x ) ( see S1 Text ) . Therefore , S[x] ≈ 0 if h ( x ) is uniform over the range of observed value of x , but > 0 if h ( x ) changes significantly with x ( Fig 3 ) . Secondly , S[x] also represents the statistical deviation between the probability distributions Pcl ( x ) and Prs ( x ) because S [ x ] = 1 τ J P cl ( x ) , P rs ( x ) , ( 9 ) where J [ p ( x ) , q ( x ) ] ≡ ∑ x ( p ( x ) - q ( x ) ) ln p ( x ) q ( x ) is the Jeffereys divergence [28–30] , a non-negative quantity that measures the dissimilarity between two probability distributions ( see S1 Text ) . From the properties of Jeffreys divergence , we can prove that 0 ≤ S [ x ] ≤ S [ D ] . ( 10 ) The strength of selection acting on any phenotypic state is therefore bounded by the strength of selection acting on D , i . e . the maximal possible value . As described in S1 Text , S[x] can be interpreted as an amount of information representing to what extent variation of D can be explained by phenotype x . Therefore , when S[x] is large , phenotype x is strongly correlated with lineage fitness . In fact , we prove that the relative selection strength , defined by S rel [ x ] ≡ S [ x ] S [ D ] , ( 11 ) is approximately equal to the squared correlation coefficient between h ˜ ( D ) and h ( x ) to the order of second cumulants ( Eq . S1 . 55 in S1 Text ) . We now introduce an explicit dependence of all quantities on an environment variable E , and using this notation Eq 7 becomes S [ x ] ( E ) ≡ ⟨ h ( x ; E ) ⟩ rs , E - ⟨ h ( x ; E ) ⟩ cl , E . ( 12 ) Let us denote the changes of mean fitness and selection strength due to an environmental shift from E 1 to E 2 as Δ〈h ( x ) 〉cl , Δ〈h ( x ) 〉rs and ΔS[x] . Then Δ ⟨ h ( x ) ⟩ rs = Δ ⟨ h ( x ) ⟩ cl + Δ S [ x ] . ( 13 ) Δ〈h ( x ) 〉rs represents the response of mean fitness in retrospective histories due to the change of the environments . Eq 13 indicates that this term can be decomposed into two terms: Δ〈h ( x ) 〉cl , which represents the intrinsic response to the environmental change; and ΔS[x] , the change of selection strength . Thus , this framework allows us to distinguish and evaluate the contributions of individual response and selection to the total change of retrospective mean fitness . In S1 Text , we apply the above framework to several analytically tractable models , and directly calculate the fitness landscape and selection strength in each model . We also provide examples of the fitness decomposition in S1 Text . To demonstrate the utility of our lineage-based analysis , we first applied it to simulation data of a cell proliferation model . In this model , we consider a population in which cells divide according to division probability f ( yt ) Δt , where Δt is time increment , and yt is a variable that represents an instantaneous state of a certain phenotype at time t ( Fig 4A ) . For example , yt could be the intracellular concentration of a protein of interest . In the simulation , we assume that ln yt follows the Ornstein-Uhlenbeck process so that the stationary distribution of yt in chronological cell histories follows the log-normal distribution with mean 1 and standard deviation 0 . 3 ( Fig 4B ) . We set f ( y ) to be a Hill function , f ( y ) = y n 1 + y n f max , where n is the Hill coefficient , and fmax is the maximum division rate ( Fig 4B ) . We fixed fmax = 1 . 2 h−1 and ran the simulation under different values of n . The initial state of a cell lineage at t0 was randomly sampled from the stationary log-normal distribution . In each condition , we repeated the simulation 100 times , i . e . N ( t0 ) = 100 , which is a realistic sample size of single-cell time-lapse experiments . To calculate the fitness landscape and selection strength , we used the lineage tree data between t0 = 0 min and t1 = 250 min ( thus τ = 250 min ) . Additional details of the simulation are described in Materials and Methods . We tested our methodology using the time-averaged expression level as a simple phenotypic trait x of cell lineages , i . e . x = y ¯ τ ≡ 1 τ ∫ t 0 t 1 y t d t . ( 14 ) We found that the fitness landscape h ( y ¯ τ ) calculated from the simulated lineage trees and the time-series of yt recovers f ( y ) accurately despite the non-linearity of this function ( Fig 4C and S2A and S2B Fig ) . The chronological mean fitness 〈 h ( y ¯ τ ) 〉 cl is unchanged by the change of n , but the retrospective mean fitness 〈 h ( y ¯ τ ) 〉 rs increases significantly with n ( Fig 4D ) . As a result , selection strength S [ y ¯ τ ] = 〈 h ( y ¯ τ ) 〉 rs - 〈 h ( y ¯ τ ) 〉 cl as well as relative selection strength S rel [ y ¯ τ ] increase with n as expected from the fact that larger n introduces greater fitness variation ( Fig 4D and 4E ) . Reducing the autocorrelation time of yt decreases selection strength ( S2C Fig ) , since faster fluctuations of the phenotype decrease the variation of the time average , y ¯ τ . In this case , h ( y ¯ τ ) deviates slightly from f ( y ) when the non-linearity is strong ( n = 10 , S2A and S2B Fig ) , which results from the fact that the time-average of f ( yt ) is not equivalent to f ( y ¯ τ ) , an effect that becomes pronounced when n is large . We also examined a bell-shaped fitness landscape , confirming that h ( y ¯ τ ) recovered f ( y ) to good precision ( see Materials and Methods and S3 Fig in detail ) . These results show that our lineage-based analysis allows us to probe fitness and selection strength of heterogeneous cellular phenotypes from realistic sample sizes of single-cell lineage trees . We point out that S [ y ¯ τ ] can be zero when f ( y ) = const . , but S[D] still becomes positive even in such circumstances due to the stochastic occurrence of cell division . In fact , we can analytically calculate that S[D] = f0 ln 2 > 0 when f ( y ) = f0 ( obtained by substituting ρ = 0 in Eq . S3 . 64 in S1 Text ) . We emphasize that it is relatively rare to find cases in which population growth is driven by the instantaneous value of a single measurable phenotype , such as yt above . That is , one should generally not equate the fitness landscape extracted for a single phenotype , h ( y ¯ τ ) , with the overall physiological fitness landscape of cells , a much more complex , multi-dimensional quantity . Instead , h ( y ¯ τ ) constitutes the effective fitness landscape for the phenotype of interest , y ¯ τ , and despite the underlying complexity of cellular physiology , it remains well defined and experimentally measurable . Next , we apply the analytical framework to analyze single-cell time-lapse data of E . coli cells that express an antibiotic resistance gene smR [31] and a fluorescent reporter venus-yfp [32] . We constructed and used two strains in the experiments: F3/pTN001 , in which venus and smR are transcribed together under the control of a common promoter PLlacO-1 [33] on a low copy plasmid pTN001 ( pSC101 ori ) , but translated separately ( Fig 5A ) ; and F3NW , in which the fusion protein , Venus-SmR is expressed from the intC locus on the chromosome under the control of PLlacO-1 promoter ( Fig 5B ) . The SmR protein confers resistance to a ribosome-targeting antibiotic drug , streptomycin , by direct inactivation [34 , 35] . We conducted fluorescent time-lapse measurements of cells proliferating on agarose pads that contain either no drug ( −Sm ) or a sub-inhibitory concentration of streptomycin ( +Sm ) ( 200 μg/ml for F3/pTN001; and 100 μg/ml for F3NW ) ( Fig 5C ) . The minimum inhibitory concentrations ( MIC ) of streptomycin for F3/pTN001 and F3NW were 1000 μg/ml and 250 μg/ml , respectively , which were significantly higher than the MIC of the parental strain F3 ( 8 μg/ml ) ( Fig 5C ) . Thus , SmR protein is functional in the constructed strains . We extracted the information of lineage trees along with time-series of cell size v ( t ) and of fluorescence intensity c ( t ) ( Fig 5D ) . Since c ( t ) is a proxy for protein concentration in a cell , c ( t ) v ( t ) can be regarded as the quantity that scales with the total amount of protein in a cell . c ( t ) and v ( t ) are correlated very weakly in all the experiments as shown in S4 and S5 Figs . Based on these quantities , we analyzed three different time-averaged phenotypes along a single-cell lineage: elongation rate λ ¯ τ , protein production rate p ¯ τ , and protein concentration c ¯ τ , which are defined as λ ¯ τ ≡ 1 τ ∫ t 0 t 1 d d t ln v ( t ) d t , ( 15 ) p ¯ τ ≡ 1 τ ∫ t 0 t 1 1 v ( t ) d d t c ( t ) v ( t ) d t , ( 16 ) c ¯ τ ≡ 1 τ ∫ t 0 t 1 c ( t ) d t . ( 17 ) We calculated these phenotypic quantities for all the lineages spanning from t0 to t1 , and obtained the chronological probability distribution Pcl ( ⋅ ) , fitness landscape h ( ⋅ ) , and selection strength S[⋅] of these phenotypes . We first analyzed the growth of F3/pTN001 . Population growth kinetics revealed that the growth rate difference between −Sm and +Sm conditions was small and became noticeable only after t = 200 min ( Fig 6A ) . Therefore , we focused on the time window between t0 = 200 min and t1 = 400 min ( see S6 Fig for the results when t0 = 0 min and t1 = 200 min ) . The population growth rates during this period were 0 . 45±0 . 01 h−1 for −Sm and 0 . 39±0 . 01 h−1 for +Sm , respectively ( p < 0 . 05 ) ( Fig 6B ) . Consistently , the mean of lineage fitness in the chronological perspective 〈 h ˜ ( D ) 〉 cl in +Sm condition was 0 . 35±0 . 01 h−1 , which is smaller than that in −Sm condition , 0 . 41±0 . 01 h−1 ( p < 0 . 05 ) ( Fig 6C ) . Despite the decrease in the mean lineage fitness , we did not detect the difference in intra-population lineage heterogeneity measured by maximum selection strength S[D] ( p = 0 . 5 ) ( Fig 6D ) . The three lineage phenotypes had distinct characteristics in their response to the drug ( Fig 6E–6I ) . The fitness landscapes of elongation rate were nearly identical between −Sm and +Sm conditions , and increased approximately linearly with λ ¯ τ ( Fig 6E and S7A and S7B Fig ) . This agrees with the natural assumption that fast elongation should lead to proportionately high division rate . The chronological distribution P cl ( λ ¯ τ ) shifted to the left in +Sm condition ( Fig 6E ) , which is also consistent with the fact that 〈 h ˜ ( D ) 〉 cl is slightly lower in +Sm condition . Nevertheless , we did not detect the difference in selection strength S [ λ ¯ τ ] ( Fig 6H ) . These results confirm that λ ¯ τ behaves coherently with D under these conditions . The fitness landscape of protein production rate were likewise nearly identical between +Sm and −Sm conditions ( Fig 6F and S7C and S7D Fig ) . The landscape is a more saturating function rather than linear with the kink around 0 . 5 a . u . The fact that h ( p ¯ τ ) is an increasing function even in the absence of the drug is presumably because overall cellular metabolism couples to all production rates and cells growing faster generally have higher production rates in most genes . The chronological distribution P cl ( p ¯ τ ) shifted significantly toward the left in +Sm condition . Interestingly , we detected an increased selection strength S [ p ¯ τ ] in +Sm condition ( 1 . 7 × 10−2 h−1 ) compared with that in −Sm condition ( 0 . 5 × 10−2 h−1 , p < 0 . 05 ) ( Fig 6H ) . The relative selection strength S rel [ p ¯ τ ] was also significantly different ( Fig 6I ) . Because S rel [ p ¯ τ ] is a measure of correlation between h ( p ¯ τ ) and lineage fitness h ˜ ( D ) , this result indicates that the heterogeneity in SmR production rate becomes more strongly correlated with fitness in +Sm condition than in −Sm condition . This change in the selection strength largely comes from the shift of the chronological distribution P cl ( p ¯ τ ) : A large portion of the probability distribution resides in the plateau region of the fitness landscape in −Sm condition , whereas its shift in +Sm condition causes a significant overlap with the linear region , resulting in a larger fitness heterogeneity in the phenotypic space of p ¯ τ . The fitness landscapes of protein concentration decrease linearly with c ¯ τ in both +Sm and −Sm conditions; protein expression levels and fitness are thus anti-correlated ( Fig 6G and S5E and S5F Fig ) . Surprisingly , we did not detect any advantages of high expression level even in the presence of the drug ( Fig 6G ) . The chronological distribution P cl ( c ¯ τ ) and selection strength S [ c ¯ τ ] were nearly identical between the two conditions ( Fig 6G and 6H ) . This indicates that , unlike production rate p ¯ τ , the strength of correlation between SmR expression level and fitness is unchanged even if the drug is added . The results therefore suggest that the protein production rate of SmR is a more responsive phenotype to drug than protein expression level in this strain . The response characteristics of selection strength are unchanged even if the relative selection strengths were compared between the two conditions ( Fig 6I ) . Applying fitness decomposition in Eq 13 to the experimental data revealed that the changes of mean fitness in retrospective perspective due to the environmental change from −Sm to +Sm ( Δ〈h ( x ) 〉rs ) mostly came from the changes in Δ〈h ( x ) 〉cl , not from the changes in selection strengths ΔS[x] , for all the phenotypes ( Table 1 ) . Therefore , the contribution of ΔS[x] to Δ〈h ( x ) 〉rs were marginal at least in the environmental difference used in this study . We found that the relative selection strengths of x = λ ¯ τ , p ¯ τ , and c ¯ τ were approximately equal to the squared correlation coefficients between h ˜ ( D ) and h ( x ) evaluated by both chronological and retrospective probabilities ( Fig 6J ) . This validates the simple interpretation that Srel[x] represents the correlation between h ˜ ( D ) and h ( x ) , though the small differences of the squared correlation coefficients between the chronological and retrospective probabilities suggest the contribution of higher-order cumulants ( S1 Text ) . We next examined how the difference in the expression scheme between F3/pTN001 and F3NW affected the phenotype distributions , fitness landscapes , and selection strength ( Fig 7 ) . For F3NW , we focused on the time window between t0 = 100 min and t1 = 300 min , where the difference in population growth rate was significant ( 0 . 52± 0 . 02 h−1 in −Sm condition , and 0 . 48 ± 0 . 01 h−1 in +Sm condition ) ( Fig 7A and 7B ) . We did not detect statistically significant differences in 〈 h ˜ ( D ) 〉 cl ( 0 . 49± 0 . 03 h−1 in −Sm , and 0 . 45 ± 0 . 01 h−1 in +Sm , Fig 7C ) and in S[D] ( 0 . 06 ± 0 . 01 h−1 in −Sm , and 0 . 057 ± 0 . 004 h−1 in +Sm , Fig 7D ) . Comparing the fitness landscapes between the two strains revealed that the overall shapes of the fitness landscapes were unchanged by the difference of the expression schemes ( Fig 7E–7G and S8A–S8F Fig ) : For elongation rate of F3NW strain , fitness landscapes increased with λ ¯ τ almost linearly; for production rate , fitness increases monotonically with p ¯ τ with a kink; and for protein concentration , fitness decreases monotonically with c ¯ τ . One important difference is that the fitness landscape of protein production rate h ( p ¯ τ ) in +Sm condition shifted significantly toward the left along with the distribution ( Fig 7F ) . Consequently , the main part of the distribution of p ¯ τ remained in the range where the fitness is fairly uniform ( Fig 7F ) , and the selection strength S [ p ¯ τ ] of F3NW did not increase in +Sm condition ( Fig 7H and 7I ) . The selection strength of λ ¯ τ and c ¯ τ of F3NW was also unchanged between −Sm and +Sm conditions ( Fig 7H and 7I ) as seen in F3/pTN001 ( Fig 6H and 6I ) . The measured selection strength of the three phenotypes was close to the squared correlation coefficients between h ( x ) and h ˜ ( D ) ( Fig 7J ) , which again validates the interpretation that Srel[x] is a measure of correlation between phenotype x and fitness . Interestingly , we found that the chronological distributions of the three phenotypes of F3NW were all narrower than those of F3/pTN001 ( Fig 7E–7G , and S9 Fig ) . This indicates that expressing SmR and Venus from the plasmid induced additional heterogeneity in all the phenotypes including elongation rate . Such non-trivial effects of different gene expression schemes can be also probed by this method quantitatively . We note that the ranges where we can assess the fitness landscapes became narrower in F3NW than in F3/pTN001 simply because of the lower levels of phenotypic heterogeneity of this strain . We remark that the measured selection strength for all the phenotypes ( x = D , λ ¯ τ , p ¯ τ , and c ¯ τ ) is significantly greater than the values calculated after randomly shuffling the combination of D and x of the lineages , which indicates that the experimentally measured S[x] reports the true selection levels ( S10 Fig ) . The details on computing selection strength for shuffled phenotypes are described in S1 Text .
Phenotypes of individual cells are intrinsically heterogeneous , and phenotypic heterogeneity is ubiquitously seen across taxa from microbes to mammalian cells . Different phenotypic states among genetically identical cells can be selected within a population when they are correlated with fitness . Therefore , unraveling the unique phenotypic characteristics that allowed a certain set of cell lineages to outperform in a population is important for understanding the biological roles of the phenotypic heterogeneity of interest . We have presented a method to quantify fitness differences and selection strength for heterogeneous phenotypic states of individual cells within a population . Our framework shares a basic idea with the method for measuring selection strength developed in evolutionary biology in that we evaluate phenotype-dependent fitness [36–38] . The key novelty of our approach is that we consider individual lineages or histories as the basic units of proliferation . An important advantage of this history-based formulation of fitness landscapes and selection strength is that it is applicable even to cellular phenotypes that fluctuate in time , such as gene expression levels in single cells . Indeed , we demonstrated by simulation that the pre-assigned fitness landscape could be recovered from the single-cell lineage trees and the associated dynamics of cellular phenotypic states despite the stochastic transitions of internal , cellular states . Though a number of single cell studies have suggested the functional roles of phenotypic fluctuation in a genetically uniform cell population [5 , 6] , our framework provides the first procedure for the rigorous quantification of the fitness values of such fluctuating cellular states . In this framework , we can use any statistical quantities that are measurable on cell lineages as the ‘phenotype’ . Although we exclusively evaluated the time-averages of cellular phenotypes along cell lineages in the analysis , other statistical quantities such as variance and coefficient of variation can also be evaluated as lineage phenotypes , which might reveal e . g . the fitness value of “noisiness” of gene expression level . Conversely , the flexibility imposes a technical challenge to select a suitable quantity that correctly reports cellular functions . We emphasize that the fitness landscapes and selection strengths quantified in this study report only correlation between the lineage phenotypes and cell division , not causality . To address causality , one must carefully choose appropriate lineage phenotypes that take detailed time-series of phenotypic states into account . One of the key features of our analysis is that we measure fitness by cell divisions , and not by cell size growth such as elongation rate , which is widely used as a proxy for fitness in single-cell analysis on bacterial growth . There are two important reasons for this: ( 1 ) population growth is ultimately driven by cell divisions , not by cell elongation; and ( 2 ) selection strength S[D] imposes the fundamental upper limit on the strength of selection for any phenotype . We are interested in determining which single-cell variables are under the strongest selection , which can be assessed by our measure of relative selection strength , Srel[x] = S[x]/S[D] . Thus , evaluating the fitness and selection through the correlation with cell divisions has a fundamental importance for studying selection in a population . We applied our method to the clonal proliferation processes of E . coli , and quantified the fitness landscape and the selection strength for different phenotypes with and without an antibiotic drug . First , we found that the elongation rate was the phenotype with largest relative selection strength , across conditions and strains , with values of S [ λ ¯ τ ] that ranged from 30% to 50% of the maximum possible value . This indicates that elongation rate behaves like a trait that is under strong phenotypic selection within clonal populations of E . coli . As mentioned previously , this analysis on its own cannot determine the causal relations , i . e . whether elongation rate is directly under selection , or indirectly by correlation with another trait . All other things being equal , however , cells that elongate faster are likely to divide sooner , and their lineages will thus be amplified with respect to cells that elongate slower and divide later , yielding a simple mechanism for the selection we detect . Second , we made an interesting observation concerning the selection strength for the time-averaged protein concentration of SmR , which was indistinguishable between the two environments with and without the drug , whereas that for time-averaged protein production rate increased significantly by drug exposure in F3/pTN001 . This result indicates that , at least for this particular strain and experimental condition , the production rate is a more responsive phenotype that increases its correlation with fitness in +Sm condition . This does not mean that SmR protein concentration is less important for the fitness in +Sm condition . The correlation between phenotype and fitness in each condition is represented by Srel[x] itself , not by the change in Srel[x] . S rel [ c ¯ τ ] of F3/pTN001 remains at a high level both in −Sm and +Sm conditions ( Fig 6I ) , thus its heterogeneity is significantly correlated with the lineage fitness . It is , however , surprising that the heterogeneity in SmR protein concentration is not correlated with fitness in the +Sm condition any more than that in the −Sm condition , considering the known functional role of SmR protein in inactivating the drug . It would be important for the future studies to examine the fitness landscapes and selection strength for broader sets of drug conditions and resistance proteins . Our method characterized the similarities and differences of phenotype distributions , fitness landscapes , and selection strength between the closely related E . coli strains ( F3/pTN001 and F3NW ) ( Fig 7 ) . Interestingly , the results revealed that the phenotypes of F3NW were all less heterogeneous than those of F3/pTN001 ( S9 Fig ) . This suggests that even a small difference of expression scheme could affect the heterogeneity levels of a large set of phenotypes . Even if one knows that different strains have different levels of phenotypic heterogeneity , the consequences for fitness are usually difficult to evaluate rigorously . Our method extracts such information from experimentally obtainable lineage trees and the measured transition of phenotypes along cellular lineages . We emphasize that our method evaluates net results , i . e . the fitness landscapes and selection strength for each phenotype whose heterogeneity can be caused by many possible noise sources . The fitness landscapes and selection strength of F3/pTN001 and F3NW are themselves valid for describing the properties of the measured phenotypes in each strain , and comparing the results among the strains would provide the contributions of different noise sources such as plasmid copy number variations . The same is true for cases where multiple cell types coexist in a population due to cellular differentiation or bet-hedging; even when the differences between these cell types are not easily apparent , our method can evaluate the overall fitness landscapes and selection strength of the phenotypes of interest . When one can clearly identify differences between cell types using markers , such a parameter can be directly incorporated into the analysis as a phenotype , and the contribution of coexisting cell types to the overall population growth can then be unraveled . Recently , several groups have demonstrated that the heterogeneity of division intervals in clonal cellular populations increases population growth rate [15 , 39] . Cerulus , et al . showed that the levels of variability and epigenetic inheritance of division intervals are changeable to a large extent depending on the environments and the genetic backgrounds in Saccharomyces cerevisiae . In general , larger variations and stronger epigenetic inheritance of division times cause stronger selection in the population , and our method allows us to quantify the contribution of these factors to population growth by the selection strength S[D] . Since S[D] is directly measurable using lineage trees , and has a clear meaning as the upper bound of selection strength for any phenotype , the statistics of division counts have a fundamental importance in our lineage analysis framework . Conventional genetic perturbation methods such as gene knock-out , overexpression , and gene suppression only associate a population-level gene expression state with population fitness; they are unable to report whether different expression states of single cells in the same population are correlated to their fitness . Our new analytical framework , however , allows us to reveal the impact of different expression levels and dynamics on cellular fitness without modifying population-level expression states , and might open up a new field in genetics that connects different expression states to cellular fitness without applying the genetic perturbation . The application of this method is not restricted to the analysis of clonal proliferation in unicellular organisms . An important application would be in the analysis of embryogenesis and stem cell differentiation of multicellular organisms , in which cellular reproduction rates diversify among the branches of lineage trees as the differentiation process goes forward [40] . Recently , large-scale cell lineage trees along with detailed quantitative information on cellular phenotypes ( gene expression , cell position , movement , etc . ) have been available [19 , 20 , 24 , 25] . Quantifying fitness and selection strength for different phenotypes at the single-cell level in differentiation processes might reveal key phenotypic steps and events leading to cell fate diversification . Additionally , fruitful applications may be found in the analysis of evolutionary lineages in viral populations , such as influenza [41] and HIV [42] , where lineage trees have been obtained using temporal sequencing data . Quantifying the strength of selection on viral traits , such as antigenic determinants , and inferring their fitness landscape is an important challenge in the field [43–45] which the method presented here could address . The application of this new lineage analysis tool to broader biological contexts may unravel the roles of phenotypic heterogeneity in diverse cellular and evolutionary phenomena .
We simulated clonal cell proliferation processes using a custom C program . We determined phenotypic state yt+Δt by randomly sampling the value of ln yt+Δt from the normal distribution with mean μ + e−γΔt ( ln yt − μ ) and variance σ2 ( 1 − e−2γΔt ) assuming that the transition of ln yt follows the Ornstein-Uhlenbeck process . We set Δt = 5 min , μ = −0 . 5 ln ( 1 . 09 ) , σ2 = ln ( 1 . 09 ) , and γ = ( −0 . 6 ln rg ) h−1 with rg = 0 . 8 . In this setting , yt follows the log-normal distribution with mean 1 . 0 and standard deviation 0 . 3 in the stationary state without selection ( i . e . Hill coefficient n = 0 ) . We assumed that cells divide with the probability of f ( yt ) Δt where f ( y ) = y n 1 + y n f max with fmax = 1 . 2 h−1 at each time point , and the initial states of two daughter cells ( yt+Δt ) were determined independently of each other from the last state ( yt ) of their mother cell . Without selection , the division rate is f0 = fmax/2 = 0 . 6 h−1 and thereby the mean interdivision time along a lineage is f 0 - 1 = 0 . 6 - 1 h - 1 . Without selection , since the normalized autocorrelation function of ln yt at stationary sate is ϕ ( τ ) = e−γτ , r g = e - γ / f 0 is the autocorrelation of ln yt after a single generation . Fitness landscapes of y ¯ τ with faster fluctuation conditions ( rg = 0 . 5 and 0 . 2 ) were shown in S2 Fig . We also ran the simulations with another type of instantaneous reproduction rate f ( y ) = f 0 1 + s - 2 exp ( - ( y - 1 ) 2 / 2 ( 0 . 3 s ) 2 ) with f0 = 0 . 6 h and with s = 0 . 5 , 1 , 2 ( S3 Fig ) . We produced a dataset that contains 100 lineage trees ( i . e . N ( t0 ) = 100 cells ) with the length of τ = 250 min in each condition , which is comparable to the data size of the real experiments ( S1 Table ) . For each condition , we repeated the simulation 10 times , and the average and standard deviation were shown in Fig 4 and in S2 Fig . We used F3 , F3/pTN001 , F3/pTN002 , and F3NW E . coli strains in the experiments . F3 is a W3110 derivative strain in which three genes ( fliC , fimA , and flu ) are deleted . pTN001 and pTN002 are low copy plasmids constructed from pMW118 ( Nippon Gene , Co . , LTD ) . We constructed pTN001 by introducing the PLlacO-1 promoter [33] , venus gene [32] , smR gene , t1t2 rrnB terminator , and frt-franked kanamycin resistance cassette [46] into the multi-cloning site of pMW118 . We also placed ribosome-binding sites in front of both venus and smR genes; these two genes are transcribed together , but translated separately . pTN002 is a control plasmid that lacks the smR gene from pTN001 . F3NW expresses the fusion protein of Venus-SmR from the intC locus of the chromosome under the control of the PLlacO-1 promoter . We also introduced mcherry-rfp into the galK locus on the chromosome under the PLtetO-1 promoter [33] to facilitate the microscopic observation and image analysis . See S1 Text and S3 Table for the details on how we constructed these plasmids and strains . We cultured the cells in M9 minimal medium ( M9 minimal salt ( Difco ) + 2 mM MgSO4 ( Wako ) + 0 . 1 mM CaCl2 ( Wako ) + 0 . 2% glucose ( Wako ) ) at 37°C . 0 . 1 mM Isopropyl β-D-1 thiogalactopyranoside ( IPTG ) ( Wako ) was added to the cultures of F3/pTN001 , F3/pTN002 , and F3NW to induce the expression of the genes under the control of the PLlacO-1 promoter . For single-cell time-lapse experiments , we solidified M9 medium with 1 . 5% ( w/v ) agarose ( Gene Pure Agarose , BM Bio ) . We adjusted the IPTG concentration in M9 agarose to 0 . 1 mM by adding ×1 , 000 concentrated IPTG solution to the melted M9 agarose before solidification . Approximately 5 mm ( W ) ×8 mm ( D ) ×5 mm ( H ) piece of M9 agarose gel was mounted onto cell suspension on a glass-bottom dish ( IWAKI ) . For +Sm condition , we added 200 μg/mL streptomycin when solidifying M9 agarose gel . Overnight cultures of the four E . coli strains in M9 medium at 37°C from glycerol stock were diluted ×100 into 2-ml fresh M9 medium and cultured for three hours at 37°C . 100 μl exponential phase culture was mixed with 100 μl fresh M9 medium containing streptomycin in a 96-well plate . We prepared 10 different conditions of streptomycin concentration for each strain with the concentration increased in two-fold stepwise . The optical density of the cell cultures after mixing was ca . 0 . 05 at 600 nm . The cell cultures in a 96-well plate were incubated by shaking at 37°C for 20 hours . We determined the MICs of the four strains with a microtiter plate ( FilterMax F5 , Molecular Devices ) by absorbance at 595 nm . To prepare a sample for time-lapse microscopy , we first cultured the cells from glycerol stock in M9 medium at 37°C by shaking overnight . Next , we diluted the overnight culture ×100 in 2 ml fresh M9 medium , and cultured it for another three hours at 37°C by shaking . We adjusted the OD600 of the culture to 0 . 05 , and 1 μl of the diluted culture was spread on a 35-mm ( ϕ ) glass-bottom dish ( IWAKI ) by placing M9 agarose pad onto the cell suspension . To avoid drying the M9 agarose pad , water droplets ( total 200 μl ) were placed around the internal edge of the dish . The dish was sealed by parafilm to minimize water evaporation . Fluorescent time-lapse images were acquired every 5 minutes with Nikon Ti-E microscope equipped with a thermostat chamber ( TIZHB , Tokai Hit ) , 100x oil immersion objective ( Plan Apo λ , N . A . 1 . 45 , Nikon ) , cooled CCD camera ( ORCA-R2 , Hamamatsu Photonics ) , and LED excitation light source ( DC2100 , Thorlabs ) . The temperature around the dish was maintained at 37°C . The microscope was controlled by micromanager ( https://micro-manager . org/ ) . Time-lapse images were analyzed with a custom macro of ImageJ ( http://imagej . nih . gov/ij/ ) . This macro produces the results file , which contains the information of mean fluorescence intensity , cell size ( area ) , and geneaological position of individual cells . We analyzed the results file with a custom C program . To evaluate fitness landscapes and selection strengths both in the simulation and the experiments , we determined the bin width based on the interquartile range of each phenotypic state ( S11 , S12 and S13 Figs ) . The details are explained in S1 Text .
|
Selection is a ubiquitous process in biological populations in which individuals are endowed with heterogeneous reproductive abilities , and it occurs even among genetically homogeneous cells due to the existence of phenotypic noise . Unlike genotypes , which can remain stable for many generations , phenotypic fluctuations at the single cell level are often comparable to cellular generation times . For this reason , quantifying the contribution of specific phenotypic states to cellular fitness remains a major challenge . Here , we develop a method to measure the fitness landscape and selection strength acting on diverse cellular phenotypes by employing a novel conceptual framework in which cellular histories are regarded as a basic unit of selection . With this framework , one can tell quantitatively whether a population adapts to environmental changes by selection or through individual responses . This new analytical approach to genetics reveals the roles of heterogeneous expression patterns and dynamics without directly perturbing genes . Applications in diverse fields including stem cell differentiation and viral evolution are discussed .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"antimicrobials",
"medicine",
"and",
"health",
"sciences",
"cell",
"division",
"analysis",
"cell",
"cycle",
"and",
"cell",
"division",
"drugs",
"cell",
"processes",
"microbiology",
"streptomycin",
"protein",
"expression",
"probability",
"distribution",
"mathematics",
"antibiotics",
"bioassays",
"and",
"physiological",
"analysis",
"pharmacology",
"population",
"biology",
"molecular",
"biology",
"techniques",
"research",
"and",
"analysis",
"methods",
"cell",
"analysis",
"molecular",
"biology",
"probability",
"theory",
"molecular",
"biology",
"assays",
"and",
"analysis",
"techniques",
"gene",
"expression",
"and",
"vector",
"techniques",
"population",
"metrics",
"cell",
"biology",
"phenotypes",
"natural",
"selection",
"genetics",
"microbial",
"control",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"evolutionary",
"biology",
"evolutionary",
"processes",
"population",
"growth"
] |
2017
|
Inferring fitness landscapes and selection on phenotypic states from single-cell genealogical data
|
In order to maintain genome information accurately and relevantly , original genome annotations need to be updated and evaluated regularly . Manual reannotation of genomes is important as it can significantly reduce the propagation of errors and consequently diminishes the time spent on mistaken research . For this reason , after five years from the initial submission of the Entamoeba histolytica draft genome publication , we have re-examined the original 23 Mb assembly and the annotation of the predicted genes . The evaluation of the genomic sequence led to the identification of more than one hundred artifactual tandem duplications that were eliminated by re-assembling the genome . The reannotation was done using a combination of manual and automated genome analysis . The new 20 Mb assembly contains 1 , 496 scaffolds and 8 , 201 predicted genes , of which 60% are identical to the initial annotation and the remaining 40% underwent structural changes . Functional classification of 60% of the genes was modified based on recent sequence comparisons and new experimental data . We have assigned putative function to 3 , 788 proteins ( 46% of the predicted proteome ) based on the annotation of predicted gene families , and have identified 58 protein families of five or more members that share no homology with known proteins and thus could be entamoeba specific . Genome analysis also revealed new features such as the presence of segmental duplications of up to 16 kb flanked by inverted repeats , and the tight association of some gene families with transposable elements . This new genome annotation and analysis represents a more refined and accurate blueprint of the pathogen genome , and provides an upgraded tool as reference for the study of many important aspects of E . histolytica biology , such as genome evolution and pathogenesis .
Although many infectious diseases receive little attention in today's world , the pathogenic intestinal parasite E . histolytica occupies a major place in the list of ignored illnesses . The parasite is the causative agent of amoebiasis , causes a significant level of morbidity and mortality in developing countries , and affects at least 50 million people every year , causing over 100 , 000 deaths [1] . Yet , a lot is there to be learned about this important protozoan . Genome information allows for better understanding of pathogenic processes and consequently helps improve the prevention , diagnosis , and treatment of the disease . Therefore , accurate and up to date data is fundamental to generate a reliable tool for both research and medical use . The E . histolytica genome was automatically annotated and published in 2005 [2] . This draft genome provided the scientific community with the first blueprint of this pathogen , its gene organization and content . However , genome annotation was performed in an automated way , leading to a very raw dataset to work with . Here , in an effort to improve the structural and functional annotation for this organism , we have reviewed , re-assembled and re-annotated the E . histolytica genome . The ultimate goal was to generate a high-quality annotation dataset to be used as gold standard by the scientific community and to carry on comparative analysis with the closely related species Entamoeba dispar and Entamoeba invadens . Using a combination of manual and automated methods we significantly improved the E . histolytica assembly . In addition , we generated a new training set of genes for training gene finders , created new gene models and reevaluated and refined previous gene structures based on up to date information , reassessed gene functions , and mapped transposable elements to remove overlapping predicted genes . Here we present an overview of the methods employed for this task and protocols followed , summarizing the contents of the latest data release , with special emphasis on our final assembly and annotation release .
Reads were obtained directly from the Sanger Institute and JCVI databases . Reads were filtered based on similarity to an E . histolytica plasmid sequence [3] or to tRNA models [4] . Reads were assembled with UMD Overlapper [5] and Celera Assembler [6] . See Text S1 for assembly details . The re-assembled sequence was deposited at the National Center for Biotechnology Information ( NCBI ) with the accession number AAFB02000000 . A set of 20 , 192 ESTs and 71 full-length cDNAs were downloaded form GenBank . ESTs were assembled and aligned to the newly assembled genome using PASA [7] . A training set consisting of 300 genes supported by 60 full length cDNAs and 240 assembled ESTs was created to train the following gene finders: Genezilla [8] , and GlimmerHMM [8] . EVidenceModeler ( EVM ) [9] was used to generate the new gene dataset , as a weighted consensus of all available evidence , including proteins and conserved protein-domains alignments , cDNAs/ESTs and gene finder output predictions . The new datas[6]et was manually inspected in areas covered by transposable elements ( see below ) . Coding regions shorter than 300 bp supported by no evidence other than Gene Finders were eliminated from the gene dataset . To generate more accurate gene structures in our new dataset , we focused on structural reannotation by improving the accuracy of existing gene models , validating intron/exon boundaries , incorporation of UTRs when available ( using PASA ) , identifying pseudogenes and eliminating spurious genes . First , we created a comprehensive custom database containing all reported E . histolytica repetitive elements: LINEs , SINEs , EhERE1 and EhERE2 [10] . Then , we ran RepeatMasker ( http://www . repeatmasker . org/ ) on the current assembly to map and quantify the elements . Regions of the genome that match the repeats were masked to avoid gene prediction on these regions . Any gene predicted on masked regions was removed from the annotation . Predicted gene models from the previous assembly were mapped to the new assembly using a combination of methods ( Fig . 1 ) . First we identified the correspondence between the scaffolds in the first assembly and the new assembly . Once this correspondence was identified , gene models from the old annotation were mapped onto the new assembly in a multistep fashion . During the first mapping iteration performed with an in-house tool , annotation_transfer , based on the software Mummer [11] , not all models were transferred as expected due to small sequence variation resulting from a new , independent assembly . In a second mapping round , unmapped genes were aligned to the new assembly using GeneWise [12] an algorithm that combines protein alignment and gene prediction into a single statistical model as a paired Hidden Markov Model ( HMM ) and provides a gene prediction based on protein homology . Then , genes that failed to map by the previous methods were positioned on the new assembly by tblastn , using a coverage of at least 80% identity , 80% coverage , and an e-value <1×10−20 . Finally , structural changes between OGA and NGA predictions was assessed using GSAC ( Gene Structure Annotation Comparison , unpublished ) , a JCVI in-house tool that evaluates coordinate differences between two gff3 ( generic feature format version 3 ) files ( http://www . sequenceontology . org/gff3 . shtml ) . To evaluate the structural improvement of gene models in the new annotation we selected a dataset of 1024 pairs of genes . Each pair was composed of an OGA and a NGA gene that only map to each other ( i . e . they represent the same gene in each annotation ) but are structurally different . This dataset was used to perform two types of analyses . First we ran HMM-searches on each pair against the Pfam HMM database and then , we evaluated NGA HMM-searches statistic ( e-value , score or number of a particular Pfam domain ) compared to their OGA counterparts . In addition , we performed local blastp searches against our internal non-redundant protein database , PANDA . db ( ftp . jcvi . org/pub/data/panda ) and identified pairs that shared the same top-hit to run stretcher , a global pairwise alignment tool ( bioweb2 . pasteur . fr/docs/EMBOSS/stretcher . html ) , between each gene and its corresponding top-hit . Pairs having hits with percent identity below 30% were removed from the results to eliminate false positive hits and results for each pair were analyzed according to their alignment statistics ( score , percent identity , percent similarity and percent of gaps ) to determine the level of improvement between the annotations . For measuring functional annotation improvement , we estimated the number of genes in the NGA that acquired a descriptive name or an improved name with respect to the OGA only for those genes that did not undergo structural changes to discard functional improvements associated with drastic structural changes , such as incorporation of new exons and changes in coding frame . Gene level searches were performed against protein , domain and profile databases including JCVI in-house non-redundant protein database Panda-AllGroup . niaa , Pfam [13] and TIGRfam [14] HMMs , Prosite [15] , and InterPro [16] . In addition , programs to predict membrane localization such as SignalP [17] , TMHMM and TargetP [17] were run . After the working gene set had been assigned function , predicted proteins were organized into protein families as previously described [7] with the purpose of refining the annotation in the context of related genes in the genome . Predicted genes were assigned informative names and classified using Gene Ontology ( GO ) [18] . GO assignments were attributed automatically , based on other assignments from closely related organisms using Pfam2GO , a tool that allows automatic mapping of Pfam hits to GO assignments as well as manually by expert annotators . All assignments were reviewed manually for consistency , on a family based approach , via Manatee , a web-based genome annotation tool that can view , modify , and store annotation for prokaryotic and eukaryotic genomes . Names between OGA and NGA were compared by simple query for corresponding genes to determine the level of change and improvement . Annotation of transporter proteins was performed using TransportDB ( http://www . membranetransport . org/ ) [19] . Segmental genome duplications along the E . histolytica genome were identified with DAGchainner [20] , a program that looks for chains of syntenic genes within complete genome sequences , using default parameters . Briefly , we performed all-vs-all blastp searches using the E . histolytica proteome . The blastp output was then filtered out to remove repetitive matches that could potentially contribute noise to the data . Finally , all segmental genome duplications containing five or more duplicated set of genes were further analyzed .
Close examination of the initial assembly of E . histolytica strain HM-1:IMSS revealed multiple problems . Sequence analysis using intra-scaffold dot-plots exposed 161 artifactual tandem duplications ( Figure S1 , panel A ) located at the boundaries between neighboring contigs ( a contiguous assembled sequence ordered together to form a scaffold ) . Tandem duplications spanned 364 , 707 bp of genomic sequence with a median length of 892 bp . In the previous assembly , genes predicted on these regions and on unmasked repetitive regions caused an over-estimation of genes by approximately 18% . Indeed , of the 399 genes located in those regions , 61 hit transposable elements ( TEs ) or were likely pseudogenes , while most of the remaining 338 coding sequences were artifactually duplicated and so collapsed into 206 individual genes in the new annotation ( Figure S1 , panel B ) . A comparative description of the features of the original and the new E . histolytica assemblies is summarized in Table 1A . The new genome assembly consists of ∼20 Mb of sequence organized into 1 , 496 scaffolds . To generate a “core” assembly for functional annotation , scaffolds lacking predicted genes were not considered . The resulting core assembly consisted of 818 non-redundant scaffolds with a total of 19 , 220 , 345 bp . All scaffolds that were excluded from the core assembly as well as degenerate contigs and singleton reads , although not annotated , were considered to survey the presence or absence of genes when necessary , and all sequences were deposited in GenBank ( see Methods ) . The results of the new assembly show higher fragmentation and a reduction in genome size with respect of the published assembly . However , our comparative analysis between the two annotations shows that there is no loss of coding information from one assembly to the other . The new assembly contains 8 , 201 de novo predicted protein coding genes , 1 , 784 fewer than previously reported for this genome ( Table 1 ) [2] . To determine the origin of these differences and to evaluate changes in gene structure between the original ( OGA ) and new ( NGA ) annotation , genes from OGA were mapped onto the new assembly and structural differences were estimated using GSAC ( see Methods and Fig . 1A ) . Mapping results indicated that the main reason for gene number reduction is the elimination of genes within repetitive regions and artifactual tandem duplications , and the removal of genes smaller than 300 bp without any supporting evidence ( Fig . 1B ) . Noteworthy , less than 0 . 2% of the genes from the original annotation do not map onto the new assembly , despite the fact that the assembly is 2 , 562 , 911 bp smaller than the published one . These missing OGA genes contained no supporting evidence and were originally annotated as hypothetical protein coding genes . This analysis also showed that 51% of the OGA genes keep the same structure in the new annotation ( same isoform in Fig . 1B ) , while 36% underwent structural change ( different isoform in Fig . 1B ) . As part of the curation process , the structure of 740 genes was manually reviewed and curated based on supporting evidence such as ESTs . An important hallmark of this work is the concerted effort from scientists of the Entamoeba community that contributed to the curation of the genome by direct communication with the authors as well as participation via specific workshops held at JCVI . To evaluate whether structural changes in the new annotation reflect an overall improvement of gene structures we selected a group of 1 , 024 OGA-NGA pairs of genes that map to each other but are structurally different . Then , we ran HMM-searches and global pairwise alignments on each pair of proteins against Pfam HMMs and our PANDA database ( see Methods ) . Finally , we compared the resulting statistics between OGA and NGA peptides from each pair ( Fig . 2 ) . These analyses showed that translated products from NGA genes consistently give better hits against Pfam and PANDA databases when compared to OGA genes , demonstrating an overall improvement in gene structures for the new annotation . In those cases where NGA genes gave worse hits compared to their OGA counterparts , we manually inspected and corrected gene structures in the new annotation . Structural improvements in the new annotation were also reflected by ( 1 ) the appearance of new Pfam/TIGRfam domain hits not present in the original protein dataset and ( 2 ) the identification of genes coding for additional members of different protein families . Noteworthy , among novel protein domains are a domain typically found in some subunits of several DNA polymerases ( PF04042 ) , a domain found in phospholipid methyltransferases ( PF04191 ) and another present in panthotenate kinase proteins ( PF03630 , see section below ) . On the other hand , point ( 2 ) is very well exemplified by the subunits of the Gal/GalNAc lectins . In E . histolytica these lectins are composed of three different subunits: a 170 kDa heavy subunit ( Hgl ) , a 150 kDa intermediate subunit ( Igl ) and a 31–35 kDa light subunit ( Lgl ) [21] , [22] . In agreement with the current number of Hgl and Lgl genes in the new annotation , studies of pulse-field gel electrophoresis have shown that there are five hgl and six lgl genes in the genome [22] . However , only four Hgl genes , one of them truncated , and four Lgl genes are part of the old dataset . Particular effort was directed towards the improvement of functional annotation ( summarized in Table 1B ) by the incorporation of additional 974 enzyme commission ( EC ) numbers and 531 Pfam/TIGRfam domains . Gene ontology ( GO ) terms were automatically assigned from Pfam HMM searches refreshing and updating the assignments from InterPro evidence used in the old annotation . The advantage of using hits from Pfam HMM searches is that results can then be filtered not just by e-value but also by trusted cutoff scores , giving a more accurate estimation than InterPro searches and therefore , a more confident GO assignment . In addition to automatic EC number and GO term assignments , functional annotation has been manually curated for 2 , 130 genes . A total of 3 , 468 genes have been assigned GO terms , of which 3 , 216 have a molecular function term . We have distributed the specific terms in a total of 30 molecular function GO-Slim categories ( Table S1 ) . No difference was observed in the representation of GO categories in the protein families with respect to that of singletons . E . histolytica predicted proteins were organized into protein families to facilitate the review of their functional annotation , visualizing relationships between proteins and allowing annotators to examine related genes as a group . Our family clustering method produces groups of proteins sharing protein domains conserved across the proteome , and consequently , related biochemical function , as described in Methods [23] , [24] . For example , based on our clustering criteria , all proteins containing a single RhoGAP domain ( PF00620 ) fall within the same family irrespectively of their length . A total of 897 protein families containing 4 , 564 proteins ( 56% of the proteome ) were identified from the 8 , 201 predicted polypeptides in the new annotation , leaving 3 , 637 “orphan” proteins . Among the families , 247 clusters ( 479 proteins ) have no homology to any known Pfam or TIGRfam domain , and harbor potentially novel domains ( 91 of these families contain five members or more ) . On average , E . histolytica families contain five proteins , ranging from two to 149 members ( Fig . 3A ) . We identified seven families with more than 50 members encoding proteins such as small GTP binding proteins , BspA-like leucine-rich repeat proteins , kinase domain-containing proteins , WD domain-containing proteins , a large family of uncharacterized hypothetical proteins , a RNA recognition motif domain-containing protein family and a RhoGAP domain-containing protein family ( see Table S2 for the complete list of families ) . Interestingly , a number of protein families appear to be physically linked to transposable elements . Table 2 shows the top 27 families that present this type of association ( for the entire repertoire of genes see Table S3 ) . For example , a cluster of 31 members of the Hsp70 protein family appears associated 30% of the time with TEs within 1 kb of the gene context . Hsp70 proteins are molecular chaperones that assist a large variety of protein folding processes in the cell by the transient association between their substrate-binding domain and the short hydrophobic peptide segments present in their target proteins . Hsp70 s are highly conserved and are known to be induced by a variety of stresses [25] . It has been previously reported that multiple natural TE insertions in Drosophila reduce the level of expression of hsp70 genes by insertion nearby gene promoter regions [26] . The characteristics of the hsp70 promoter in the fly may make it a suitable target for transposition leading to the generation of novel alleles . In this sense , TEs could be playing an adaptive role in microevolution by gene amplification and also manipulating the expression of genes critical for the parasite fitness [27] . Another family showing a high correlation with transposable elements is the large BspA-like surface protein family [28] , [29] . Initially , Davis et al . identified 89 genes coding for BspA-like proteins in the genome of E . histolytica , containing a leucine-rich repeat motif ( LRRs ) . LRRs serve as recognition motifs for surface proteins in bacteria and other eukaryotes [30] and have been shown to be involved in binding to fibronectin . E . histolytica BspA-like proteins have unique LRR-like repeats that resemble , to certain extent , to the Treponema pallidum LLRs ( LrrA proteins ) [28] , that appear to have a role in attachment and penetration to host tissues [31] , suggesting they may be involved in attachment to the host cells . Our analysis identified 116 BspA-like genes in the genome , 41 of them associated with transposable elements . The core domain of the BspA-like proteins contains 23 amino acids with the consensus P[T/S][T/S][V/I/L]xx[I/L]GxxCFxxCxxLxx[I/L]x[I/L] , and these units form tandem blocks that can contain two or more core motifs represented from 1 to 21 times in a single molecule , leading to a great variability in the protein length in the family . Most of the proteins in the family contain a novel 50 amino acids N-terminal domain that is preserved in 85 members of this cluster . A closer examination of those genes encoding proteins lacking the N-terminal domain showed they are probably truncated by the insertion of transposable elements , primarily SINE and LINE elements at their 5′ end . BspA-like proteins are located on the surface of E . histolytica [28] however no classic membrane-targeting signal is present in the proteins . Therefore , it is tempting to speculate that the conserved N-terminal domain of these proteins might function as either an export signal or serve as a membrane-anchor domain or that export involves a non-classical transport mechanism , independent of the ER–Golgi pathway , similar to those that have been detected in yeast and mammalian cells [32] . Details on the motifs and domain structure are shown in Figure S2 . A third worthy of note family associated with TEs is the AIG family of proteins , comprising 29 members distributed in 3 clusters , of which 18 genes are in close proximity to repetitive elements ( Table 2 ) . AIG1 proteins are associated with resistance to bacteria [33] . Interestingly , comparative gene expression studies have shown that AIG1 proteins as well as heat shock proteins have significantly reduced expression levels in E . dispar [34] , when compared to E . histolytica . This observation leads us to speculate that transposable elements inserted in the neighborhood of these genes could lead to the enhanced expression of these genes and ultimately could be related to the increased virulence . Indeed we have previously shown that LINEs and SINEs are involved in genome rearrangements driving in consequence genomic evolution [10] . It is tempting to speculate that the amplification of the AIG family was mediated by the close association of TEs , but the observation that non-virulent E . dispar contains the same number of genes without the TE association seems to indicate that this is not the case . We are currently analyzing all gene family/transposable element associations in the context of comparative genomics with other Entamoeba species ( manuscript in preparation ) . Close examination of the functional annotation of protein families and singleton proteins revealed that a total of 2 , 981 ( 65% ) genes within the families were annotated as encoding proteins with putative functions and 1 , 583 genes are hypothetical proteins ( 34% , Fig . 3B ) . Of a total of 1 , 088 genes that have EST support in the whole genome , 705 are genes within protein families . In contrast , singletons had a larger proportion of hypothetical genes ( 76% ) and a smaller portion of genes with a known or putative function ( 24% ) , and half the number of genes supported by EST evidence ( 383 ) . As mentioned above , about 20% of the E . histolytica genome consists of transposable elements . These repeats show a tendency to insert close to each other forming large TE clusters . We have previously shown that these repeat clusters are frequently found at syntenic breakpoints between E . histolytica and E . dispar suggesting that they could contribute to parasite genome instability and , consequently , to the evolution of these species [10] . It is also possible that the highly repetitive nature of this genome led to genome duplications . In order to evaluate this possibility we analyzed the presence of additional rearrangements within the genome by searching for segmental duplications using DAGchainer as explained in Methods [20] . We observed the presence of four different types of segmental duplications , named D1-D4 , spanning seven to ten genes each ( Fig . 4 ) . The first duplication ( D1 , Fig . 4A ) spans a 16 . 6 kb region containing up to 8 hypothetical protein coding genes . These duplications are approximately 94% identical at the nucleotide level . All D1-type duplications are flanked by 2 . 3 kb inverted repeats ( IR ) not found in the rest of the genome . Nucleotide composition analysis revealed that D1-IRs are highly AT-rich ( 84 . 3% ) compared to the average content of those regions 71 . 4% and they are 95% identical at the nucleotide level . A genome wide survey of D1-duplications led to the identification of four complete and two partial copies of this element in the genome . It is interesting to mention that all the scaffolds containing the four complete duplications have similar sizes ( 16 . 6 kb on average ) and are spanned almost in their entire length by their respective segmental duplications . The two partial D1-duplications are located in shorter scaffolds of 14 . 4 kb and 6 . 6 kb , respectively . The second duplication ( D2 , Fig . 4B ) is 12 . 5 kb long and contains up to eight duplicated hypothetical protein coding genes depending on the duplication . Comparative analysis showed that these duplications are more than 80% identical at the nucleotide level with an average of 92% . Similar to D1-type duplications , D2 are frequently flanked by 1 . 2 kb IRs , composed of two fragments derived from the TEs EhERE1 and EhLINE2 . D2-IRs share 92 . 6% identity at the nucleotide level and are also very AT-rich ( 85%AT ) . The organization of the duplications is not conserved in all copies across the genome , with some copies flanked by IRs composed of either EhERE1 or EhLINE2 fragments , while in others we could not identify any IR . D3-type duplications are 7 . 4 kb long and 83% identical at the nucleotide level . Although frequently found nearby TEs ( mostly EhLINE1 ) , none of the eight identified genome duplications are flanked by IRs as D1- and D2-type duplications . D3 presents a very unique gene content that suggest that the segment could present a unique functionality , represented in Fig . 4C . A total of seven protein coding genes are arranged in the same orientation , and include a putative serine-threonine kinase similar to ARK1 , a human protein that participates in cell cycle regulation; an endonuclease V domain-containing protein coding gene potentially involved in DNA repair; a putative secreted hypothetical protein coding gene; a tandem duplicated gene coding for a putative protein containing a type-1 glutamine amido transferase-like domain and a GDSL-like lipase/acylhydrolase domain-containing protein coding gene . Interestingly , D3-type duplications are found at or in close proximity to the end of scaffolds , and therefore , they could potentially be located at subtelomeric regions . However , in spite of a thorough analysis we could not identify any repetitive telomeric/subtelomeric motif in these regions . Lastly , the 10 kb long D4 ( Fig . 4D ) shares more than 85% identity at the nucleotide level and spans up to 9 hypothetical protein coding and one putative dUTP hydrolase-coding genes . Most D4-type duplications have TEs inserted nearby , but no flanking IRs were identified . The presence of these duplications is not likely to be an artifact of the assembly due to the fact that they are also appear duplicated in E . dispar . It is possible that some of these duplications , that in some cases span full scaffolds represent different copies of one of the several extrachromosomal elements known to exist in Entamoeba species , as described by Dhar et al [35] . Our work has led to the identification of 460 novel putative protein coding genes not present in the OGA , 16% of which have some functional annotation . One of these genes codes for a putative pantothenate kinase ( EHI_183060 ) the first enzyme in the biosynthesis of coenzyme A from pantothenate . Although the coding genomic region was present in the original assembly , the gene had not been predicted and therefore , it was missing from the previous annotation . Only the enzymes phophopantothenoyl-cysteine decarboxilase ( EC 4 . 1 . 1 . 36 ) , phosphopantothenoyl-cysteine synthase ( EC 6 . 3 . 2 . 5 ) , and dephospho-CoA kinase ( EC 2 . 7 . 1 . 24 ) , responsible for the second , third and last of the five consecutive enzymatic reactions , had been previously identified in the OGA ( EHI_164490 , EHI_092330 , EHI_040840 ) . However , the lack of candidate enzymes for the remaining two biochemical reactions of this pathway raised the question whether E . histolytica can synthesize coenzyme A from pantothenate [36] . Our de novo gene prediction for a putative pantothenate kinase plus the identification of a candidate gene for the forth step of this pathway , a putative pantetheine-phosphate adenylyltransferase ( EC 2 . 7 . 7 . 3 ) , indicates that the whole set of metabolic reactions required to synthesize coenzyme A from pantothenate is present in this amoeba . Interestingly , the enzymes that participate in this pathway resemble those from eubacteria but not higher eukaryotes . Indeed , the second and third sets of reactions are catalyzed by a single enzyme present in two copies ( EHI_164490 , EHI_092330 ) , while the fourth and fifth steps are carried out by independent enzymes , EHI_006680 and EHI_040840 , respectively . In higher eukaryotes the last two reactions are carried out by the same enzyme [37] . Another gene not present in the OGA ( EHI_141410 ) codes for a protein with a predicted molecular weight of 44 . 6 kDa similar to subunit p50 of the DNA polymerase delta , a key enzyme for chromosomal DNA replication in higher eukaryotes . In mammals , it has been shown that p50 is tightly associated with p125 , the catalytic subunit of these types of DNA polymerases . Accordingly , a gene coding for a putative 124 . 4 kDa catalytic subunit of the DNA polymerase delta ( EHI_006690 ) , is also present in the NGA . These results are in agreement with a previous work showing that the sensitivity to different inhibitors of the DNA polymerase activity of E . histolytica resembles that of mammalian DNA alpha , delta and epsilon polymerases [38] . In addition , a gene coding for a protein containing a Yos1-like Pfam domain is also absent from OGA ( EHI_178640 ) . This putative protein has similarity to other members of the Yos1 family , involved in protein transport between the endoplasmic reticulum and the Golgi apparatus [39] . Comparative analysis between the two annotation datasets also allowed us to identify genes present in their complete form in NGA but truncated in OGA . Example of these genes are two copies of a gene coding for a putative pyridine nucleotide transhydrogenase , EHI_055400 and EHI_014030 , the latter identical to a gene previously cloned by Clark et al . [40] , which exists as a single truncated copy in the OGA . Another example is a 605 bp gene coding for a putative phospholipid methyltransferase protein ( EHI_153710 ) similar to Schizosaccharomyces pombe cho1 ( 35% identity; e-value = 4×10−21 ) , an enzyme that participates in the synthesis of phosphatidylcholine via the methylation of phosphatidylethanolamine . A coding sequence containing only the last 222 bp of this gene is present in the OGA . Our reannotation effort has focused mostly on the improvement of the assembly and the gene content and structure of the E . histolytica genome . The new assembly , annotation and analysis of the genome has incorporated many updates and enhancements to the structural and functional assignments of the original gene predictions , including an improved assembly , removal of spurious genes , improved gene structures and functional assignments , and generation of gene families . Regardless of the advancement of the computational methods and of the exponentially growing amount of data that could be used for automated genome annotation , only experimental evidence from expression data will conclusively validate the accuracy of computationally assigned functions done at the genome-wide level . Nevertheless , in order to provide a sound bases to drive research , genome annotations have to be maintained and revised , either by expert annotators in the field and/or community involvement . Additional sequence information will allow the further refinement of gene structures and a deeper understanding of the genome architecture , while the functional annotation will be enriched both by the availability of new experimental data and from expression and other kinds of analyses to characterize each gene and its function fully . This reannotation effort will be the base for the future analysis and annotation of new E . histolytica genomes from patient isolates , a project recently approved under the NIAID supported program Genome Sequence Centers for Infectious Disease , GSCID ( http://gsc . jcvi . org/ ) .
|
Entamoeba histolytica is an anaerobic parasitic protozoan that causes amoebic dysentery . The parasites colonize the large intestine , but under some circumstances may invade the intestinal mucosa , enter the bloodstream and lead to the formation of abscesses such amoebic liver abscesses . The draft genome of E . histolytica , published in 2005 , provided the scientific community with the first comprehensive view of the gene set for this parasite and important tools for elucidating the genetic basis of Entamoeba pathogenicity . Because complete genetic knowledge is critical for drug discovery and potential vaccine development for amoebiases , we have re-examined the original draft genome for E . histolytica . We have corrected the sequence assembly , improved the gene predictions and refreshed the functional gene assignments . As a result , this effort has led to a more accurate gene annotation , and the discovery of novel features , such as the presence of genome segmental duplications and the close association of some gene families with transposable elements . We believe that continuing efforts to improve genomic data will undoubtedly help to identify and characterize potential targets for amoebiasis control , as well as to contribute to a better understanding of genome evolution and pathogenesis for this parasite .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"and",
"Discussion"
] |
[
"genetics",
"and",
"genomics/genome",
"projects",
"genetics",
"and",
"genomics/bioinformatics",
"genetics",
"and",
"genomics/genomics"
] |
2010
|
New Assembly, Reannotation and Analysis of the Entamoeba histolytica Genome Reveal New Genomic Features and Protein Content Information
|
Pattern classification of human brain activity provides unique insight into the neural underpinnings of diverse mental states . These multivariate tools have recently been used within the field of affective neuroscience to classify distributed patterns of brain activation evoked during emotion induction procedures . Here we assess whether neural models developed to discriminate among distinct emotion categories exhibit predictive validity in the absence of exteroceptive emotional stimulation . In two experiments , we show that spontaneous fluctuations in human resting-state brain activity can be decoded into categories of experience delineating unique emotional states that exhibit spatiotemporal coherence , covary with individual differences in mood and personality traits , and predict on-line , self-reported feelings . These findings validate objective , brain-based models of emotion and show how emotional states dynamically emerge from the activity of separable neural systems .
Functional neuroimaging offers unique insight into how mental representations are encoded in brain activity [1 , 2] . Seminal cognitive neuroscience studies demonstrated that distributed patterns of cortical activity measured with functional magnetic resonance imaging ( fMRI ) contain information capable of differentiating among visual percepts , including object categories [3] and basic visual features [4] . Extending findings from these studies , subsequent work demonstrated that machine learning models trained on stimulus-evoked brain activity , termed “decoding” or “mind-reading” [5] , can be used to predict the contents of working memory [6–8] and mental imagery [9 , 10] , even during sleep [11] . Thus , pattern recognition approaches can identify defining features of mental processes , even when driven solely on the basis of endogenous brain activity . The approach was further shown to accurately discriminate among multiple cognitive processes ( e . g . , decision-making , working memory , response inhibition , among others ) in independent subjects [12] , establishing the efficacy of assessing diverse mental states with fMRI across individuals . Paralleling cognitive studies decoding task-evoked brain activity , multivariate decoding approaches have recently been used to map patterns of neural activity evoked by emotion elicitors onto discrete feeling states [13 , 14] . However , a key piece of missing evidence is whether categorically distinct emotional brain states occur intrinsically [15 , 16] in the absence of external eliciting stimuli . If so , then it should be possible to classify the emotional status of a human being based on analysis of spontaneous fluctuations of brain activity during rest . Successful classification would validate multivariate decoding of unconstrained brain activity and provides insight into the nature of emotional brain activity during the resting state . Adapting the logic of other cognitive imaging studies [16 , 17] , we postulate that the presence of spontaneous emotional brain states should be detectable using multivariate models derived from prior investigations of emotion elicitation . We previously developed decoding algorithms to classify stimulus-evoked responses to emotionally evocative cinematic films and instrumental music [13] . These neural models ( Fig 1 ) accurately classify patterns of neural activation associated with six different emotions ( contentment , amusement , surprise , fear , anger , and sadness ) and a neutral control state in independent subjects , generalizing across induction modality . Importantly , these neural biomarkers track the subjective experience of discrete emotions independent of differences in the more general dimensions of valence and arousal [18] . By indexing the extent to which a pattern of neural activation to extrinsic stimuli reflects a specific emotion , these models can be used to test whether intrinsic spatiotemporal patterns of brain activity correspond to stimulus-evoked emotional states . Here , we evaluate whether these neural models of discrete emotions generalize to spontaneous brain activation measured via fMRI in two experiments . The first experiment assesses if model predictions are convergent with individual differences in self-reported mood and emotional traits . Because individual differences are linked to mental health and subjective well-being [19–21] , this evaluation provides insight into the potential clinical utility of quantifying spontaneous emotional states , as they may be associated with risk factors for mental illness . The second experiment employs an experience sampling procedure to evaluate whether model predictions based on brain activity during periods of rest are congruent with on-line measures of emotional experience . Together , these studies probe how brain-based models of specific emotion categories quantify changes in extemporaneous affect both between and within individuals .
We applied the multivariate models of emotional experience to brain activation acquired from young adults during resting-state fMRI ( n = 499; Fig 2A ) . Two consecutive runs of resting-state scans were acquired , spanning a total duration of 8 . 53 min . Following preprocessing of data , we computed the scalar product of the resting-state signal and emotion category-specific model weights at every time point of data acquisition . This procedure yielded scores that reflect the relative evidence for each of seven emotional states across the full scanning period . A confirmatory analysis revealed that voxels distributed across the whole brain informed this prediction , as opposed to activity in a small number of brain regions ( S1 Fig ) . If emotional brain states occur spontaneously , the frequency of classifications from our decoding models should be more varied than the uniform distribution that would be expected by chance . To test this hypothesis , we sought to identify whether the total time ( or absolute frequency ) in each state differed across emotion categories . Such an analysis informs the degree to which discrete emotional brain states may spontaneously occur and , by extension , could contribute to the identification of individual differences that map onto the likelihood of experiencing specific spontaneous states . To perform this comparison , we identified the single model with the maximum score at each time point ( one-versus-all classification ) and summed the number of time points assigned to each category . The frequency of emotional states clearly differed across categories ( Fig 2B , χ2 = 1491 . 52 , P < . 0001 , Friedman test ) , in contrast to the uniform distribution that would be expected if emotional brain-states did not occur in spontaneous activity ( see S2 Fig ) . Follow-up comparisons revealed that neutral states occurred more frequently than chance rates ( 20 . 1 ± 3 . 59% [s . d . ] , z = 20 . 50 , Punc = 2 . 03E-93 ) , followed by states of surprise ( 18 . 37 ± 3 . 87% [s . d . ] , z = 16 . 38 , Punc = 2 . 47E-60 ) and amusement ( 14 . 71 ± 3 . 78% [s . d . ] , z = 1 . 25 , Punc = 0 . 21 ) . States of sadness ( 13 . 49 ± 3 . 76% [s . d . ] , z = -3 . 31 , Punc = 9 . 24E-4 ) , fear ( 13 . 26 ± 3 . 42% [s . d . ] , z = -5 . 28 , Punc = 1 . 28E-7 ) , and anger ( 11 . 31 ± 3 . 62% [s . d . ] , z = -13 . 07 , Punc = 4 . 78E-39 ) occurred with lower frequency , while states of contentment occurred the least often ( 8 . 74% ± 3 . 42% [s . d . ] , z = -19 . 61 , Punc = 1 . 33E-85; see Table 1 ) . Although patterns of neural activation were most often classified as neutral as a whole , it is possible that consistent fluctuations in the time course of emotional states occur against this background . Research on MRI scanner-related anxiety has shown that self-report [22 , 23] and peripheral physiological [24] measures of anxiety peak at the beginning of scanning , when subjects first enter the scanner bore . This literature predicts that brain states indicative of fear should be most prevalent at the beginning of resting-state runs , and that neutral states should emerge over time , given their overall high prevalence ( Fig 2B ) . To assess gradual changes in the emotional states over time , we performed Friedman tests separately for each emotion category , all of which revealed significant effects of time ( see S1 Table ) . Next , we quantified the direction of these effects using general linear models to predict classifier scores using scan time as an input . We found the scores for fear decreased over time ( β^=−0 . 001 , t498 = -4 . 92 , Punc = 1 . 20E-006 , Fig 3 gray lines ) , whereas neutral states exhibited an increasing trend throughout the scanning period ( β^=0 . 0017 , t498 = 7 . 36 , Punc = 7 . 66E-013 ) , consistent with predictions ( additional effects were observed for scores for contentment [β^=0 . 0017 , t498 = 7 . 37 , Punc = 7 . 05E-13] , surprise [β^=0 . 0010 , t498 = 4 . 07 , Punc = 5 . 51E-05] , anger [β^=−0 . 0007 , t498 = -3 . 36 , Punc = 0 . 00085] , and sadness [β^=−0 . 0034 , t498 = -15 . 59 , Punc < 2 . 52E-038] ) . To determine whether emotional states exhibited consistent dynamics over the course of the scanning period , we fit smoothing spline models [25] for each subject and assessed the correlation between each subject and the average time course of other subjects in a cross-validation procedure . This analysis showed that there is substantial moment-to-moment variability in the time course of emotional states across subjects ( which cannot simply be explained by scaling differences in the emotion models or resting-state data; see S3 Fig ) . Consistent with the linear models using time as a predictor , evidence for neutral brain states was most prevalent in the second scanning session , especially during a peak at the beginning of the run , whereas the time course for fear peaked at the beginning of the first run and decreased throughout the scanning session . The model for surprise exhibited a similar time course as neutral states but peaked at the end of the second run . Additionally , this analysis showed that evidence for sad classifications peaked in the middle of the first run and decreased over time . Overall , these time series revealed a gradual change in evidence from negative emotions ( fear and sadness in run 1 ) to non-valenced or bi-valenced emotions ( neutral and surprise in run 2 ) . To ensure that our emotion-specific brain states are not proxies for more general resting-state networks thought to subserve other functions , we examined the spatial overlap between our models and those commonly derived by connectivity-based analysis of resting-state fMRI data [26] . On average , we observed little overlap ( Jaccard index = 13 . 1 ± 1 . 97% [s . d . ]; range 10 . 8%–16 . 7% ) with the seven most prominent networks found in resting-state data , implicating a substantial degree of independence . To further establish the construct validity of the spontaneous emotional brain states , we reasoned that their incidence should vary with individual differences in self-reported mood and personality traits associated with specific emotions . We assayed depressive mood with the Center for Epidemiologic Studies Depression Scale ( CESD ) [27] and state anxiety using the State-Trait Anxiety Inventory State Version ( STAI-S ) [28] , instructing participants to indicate how they felt during the resting-state scan itself . Binomial regression models revealed that higher depression scores were associated with increases in the frequency of sadness ( β^=0 . 0025 , t497 = 2 . 673 , Punc = . 0075 , Fig 4A , see S4 Fig for scatter plots of predictions ) and no other emotional state ( all Punc > . 24 ) . State anxiety was associated with increasing classifications of fear ( β^=0 . 0033 , t497 = 2 . 608 , Punc = . 0091 ) and decreasing frequency of contentment ( β^=−0 . 0031 , t497 = -2 . 015 , Punc = . 0439 ) . Viewing these beta estimates as odds ratios ( computed as eβ^ ) reveals how a one-unit increase in self-reported mood is associated with differences in the occurrence of spontaneous emotional states . Applying this approach to CESD scores reveals that individuals with a score of 16 ( the cutoff for identifying individuals at risk for depression ) have 5 . 92% increased odds of being in a sad state compared to those with a score of 0 . In more practical terms , this corresponds to approximately seven extra minutes a day of exhibiting a brain state that would be classified as sadness . Drawing from the Revised NEO Personality Inventory ( NEO-PI-R ) [29] , we focused personality trait assessment on the specific Neuroticism subfacets of Anxiety , Angry Hostility , and Depression , due to their discriminant validity [30] , heritability [31] , universality [32] , and close theoretical ties to the experience of fear , anger , and sadness . We found that increasing Anxiety scores were associated with more frequent classification of fear ( β^=0 . 003 , t497 = 1 . 978 , Punc = 0 . 0479 , Fig 4B ) and fewer classifications of anger ( β^=−0 . 004 , t497 = -2 . 407 , Punc = 0 . 0161 ) . Angry Hostility scores were positively associated with the number of anger classifications ( β^=0 . 0042 , t497 = 2 . 400 , Punc = 0 . 0164 ) . Depression scores were positively associated with the frequency of fear ( β^=0 . 003 , t497 = 2 . 058 , Punc = 0 . 0396 ) and sadness ( β^=0 . 0037 , t497 = 2 . 546 , Punc = 0 . 0109 ) . These results provide converging evidence across both state and trait markers that individual differences uniquely and differentially bias the spontaneous occurrence of brain states indicative of fear , anger , and sadness . Finally , we examined whether the predictions of our decoding models were consistent with self-report of emotional experience during periods of unconstrained rest . We conducted a separate fMRI experiment in which an independent sample of young adult participants ( n = 21 ) performed an experience sampling task in the absence of external stimulation ( Fig 5A ) . Participants were instructed to rest and let their mind wander freely with their eyes open during scanning . Following intervals of rest of at least 30 s , a rating screen appeared during which participants moved a cursor to the location on the screen that best indicated how they currently felt . If spontaneous emotional states are accessible to conscious awareness , then scores should be greater for emotion models congruent with self-report relative to scores for models incongruent with self-report . Contrasting emotion models in this manner is advantageous from a signal detection standpoint because it minimizes noise by averaging across emotions , as some were reported infrequently or not at all in some subjects ( see [33] for an analogous approach to predict the contents of memory retrieval during similarly unconstrained free-recall ) . To test our hypothesis , we extracted resting-state fMRI data from the 10-s interval preceding each self-report query and applied multivariate models to determine the extent to which evidence for the emotional brain states in this window predicted the participants’ conscious emotional experience . Consistent with our hypothesis , we found that scores for models congruent with self-report were positive ( 0 . 016 ± 0 . 0093 [s . e . m . ] , z = 2 . 068 , Punc = 0 . 0386; Wilcoxon signed rank test ) , whereas scores for incongruent models were negative ( -0 . 0048 ± 0 . 0017 [s . e . m . ] , z = -3 . 041 , Punc = 0 . 0024 ) . Classification of individual trials into the seven emotion categories exhibited an overall accuracy of 27 . 9 ± 2 . 1% ( s . e . m . ) of trials , where chance agreement is 21 . 47% ( Punc = 0 . 001; binomial test ) . Not only do these results demonstrate that classification models are sensitive to changes in emotional state reported by participants , but also that there is selectivity in their predictions , as negative scores indicate evidence against emotion labels that are incongruent with self-report . Establishing both sensitivity and selectivity is important for the potential use of these brain-based models as diagnostic biomarkers of emotional states . As an additional validation of our decoding models , we examined the correspondence between the prevalence of individual emotional brain states as detected via pattern classification and participant self-report . Classifications based on self-report and multivariate decoding yielded similar frequency distributions ( Fig 5C ) , in which neutral and amusement were the most frequent . We found a positive correlation between the frequency of classifications based on participant ratings and multivariate decoding ( r = . 3876 ± 0 . 102 [s . e . m . ] , t20 = 2 . 537 , Punc = . 0196; one sample t test ) , further demonstrating a link between patterning of brain states and subjective ratings of emotional experience in the absence of external stimuli or contextual cues .
Converging findings from our experiments provide evidence that brain states associated with distinct emotional experiences emerge during unconstrained rest . Whereas prior work has decoded stimulus-evoked responses to emotional events , our study demonstrates that spontaneous neural activity dynamically fluctuates among multiple emotional states in a reliable manner over time . Observing such coherent , emotion-specific patterns in spontaneous fMRI activation provides evidence to support theories that posit emotions are represented categorically in the coordinated activity of separable neural substrates [34 , 35] . Validating the neural biomarkers in the absence of external stimulation suggests that they track information of functional significance , and do not merely reflect properties of the stimuli used in their development . It is possible that these classifiers detect the endogenous activity of distributed neural circuits , consistent with recent views that emotions are not represented in modular functional units [36 , 37] . However , the extent to which such activity is the result of innate emotion-dedicated circuitry , a series of cognitive appraisals , or constructive processes shaped by social and environmental factors remains to be determined ( for a review of these viewpoints , see [38] ) . Regardless of the relative influence of such factors , the present findings suggest that the emotion-specific biomarkers track the expression of functionally distinct brain systems , as opposed to idiosyncrasies of the particular machine-learning problem . Our findings complement recent studies demonstrating that a variety of emotion manipulations have lasting effects on resting brain activity [39–41] . For instance , one study revealed elevated striatal activity following gratifying outcomes in a decision-making task—an effect that was diminished in individuals with higher depressive tendencies [39] . Because these effects immediately followed emotional stimulation , they could plausibly reflect regulatory processes or lingering effects of mood . The present results , on the other hand , show that resting brain activity transiently fluctuates among multiple emotional states and that these fluctuations vary depending on the emotional status of an individual . Thus , emotional processes unfolding at both long and short time scales likely contribute to spontaneous brain activity . Findings from our resting state experiment stand in contrast to recent work investigating emotion-specific functional connectivity [42] . In this study , whole-brain resting-state functional connectivity was assessed using seeds identified from a meta-analytic summary of emotion research [43] . This latter approach failed to reveal unique patterns of resting-state connectivity for individual emotions but showed that seed regions were commonly correlated with domain-general resting-state networks , such as the salience network [44] . In light of the present results , it is important to consider methodological differences between studies . Seed-based correlation highlights connectivity between brain regions whose time course of activation is maximally similar to the activity of a small number of voxels ( which are averaged together to create a single time series ) , whereas pattern classification identifies combinations of voxels that maximally discriminate among mental states . Because individual voxels sample diverse neural populations [45] , it is plausible that seed-based correlation is biased towards identifying networks that have large amplitudes in seeded regions as opposed to exhibiting specificity ( e . g . , see [46] ) . Thus , our approach may have greater sensitivity to detect discriminable categorical patterns . Results of the experience sampling study provide external validation of our emotion-specific biomarkers [13] . Consistent with the resting-state study , the overall distribution of emotional states was clearly non-uniform , and classifications of neutral states occurred with high frequency . Beyond these commonalities , the inclusion of behavioral self-report led to differences in emotion-related brain activity . States of contentment and amusement were more frequently predicted during experience sampling compared to resting-state ( 46 . 31% versus 23 . 45% ) , a finding that was corroborated by higher ratings for these emotions in the self-report data . It is possible that this difference in the frequency of positive brain states is the result of a self-presentation bias [47] , wherein participants may have employed emotion regulation in order to project a more positive image . Alternatively , it is possible that the self-reporting task requirement elicited more introspection between trials , which contributed to the pattern of altered emotional states [48] . Future work will be necessary to fully characterize how such cognitive-emotional interactions shape the landscape of emotional brain states [36 , 49] . We found that individual differences in mood states and personality traits are associated with the relative incidence of brain states associated with fear , anger , and sadness . These findings further establish the construct validity of our brain-based models of emotion and link subfacets of Neuroticism to the expression of emotion-specific brain systems . Given their sensitivity to individual differences linked to the symptomology of anxiety and depression , spontaneous emotional brain states may serve as a novel diagnostic tool to determine susceptibility to affective illness or as an outcome measure for clinical interventions aimed at reducing the spontaneous elicitation of specific emotions . This tool may be particularly useful to objectively assess the emotional status of individuals who do not have good insight into their emotions , as in alexithymia , or for those who cannot report on their own feelings , including patients in a vegetative or minimally conscious state .
All participants provided written informed consent in accordance with the National Institutes of Health guidelines as approved by the Duke University IRB . The resting state experiment was approved as part of the Duke Neurogenetics Study ( Pro00019095 ) with an associated database ( Pro00014717 ) . The experience sampling project was approved separately ( Pro00027404 ) . Classification of emotional states was performed using neural biomarkers that were developed based on blood oxygen level dependent ( BOLD ) responses to cinematic films and instrumental music [13] . This induction procedure was selected because it reliably elicits emotional responses over a 1 to 2 min period , as opposed to longer-lasting moods . These models were developed to identify neural patterning specific to states of contentment , amusement , surprise , fear , anger , and sadness ( in addition to a neutral control state ) . These particular emotions were modeled to broadly sample both valence and arousal , as selecting common sets of basic emotions ( e . g . , fear , anger , sadness , disgust , and happiness ) undersamples positive emotions . In selecting these particular emotions , we verified that the accuracy of these models tracked the experience of specific emotion categories ( average R2 across emotions = . 57 ) independent of subjective valence and arousal . Thus , the models offer unique insight into the emotional state of individuals and characterize the likelihood they would endorse each of the seven emotion labels , independent of general factors such as valence or arousal . A total of 499 subjects ( age = 19 . 65 ± 1 . 22 years [mean ± s . d . ] , 274 women ) were included as part of the Duke Neurogenetics Study ( DNS ) , which assesses a wide range of behavioral and biological traits among healthy , young adult university students . For access to this data , see information provided in S1 Text . This sample was independent of that used to develop the classification models . This sample size is sufficient to reliably detect ( β = . 01 ) a moderate effect ( r = . 2 ) with a type-I error rate of . 05 , which is particularly important when studying individual differences in neural activity . All participants provided informed consent in accordance with Duke University guidelines and were in good general health . The participants were free of the following study exclusions: ( 1 ) medical diagnoses of cancer , stroke , head injury with loss of consciousness , untreated migraine headaches , diabetes requiring insulin treatment , chronic kidney or liver disease , or lifetime history of psychotic symptoms; ( 2 ) use of psychotropic , glucocorticoid , or hypolipidemic medication; and ( 3 ) conditions affecting cerebral blood flow and metabolism ( e . g . , hypertension ) . Diagnosis of any current DSM-IV Axis I disorder or select Axis II disorders ( antisocial personality disorder and borderline personality disorder ) , assessed with the electronic Mini International Neuropsychiatric Interview [50] and Structured Clinical Interview for the DSM-IV subtests [51] , were not an exclusion , as the DNS seeks to establish broad variability in multiple behavioral phenotypes related to psychopathology . No participants met criteria for a personality disorder , and 72 ( 14 . 4% ) participants from our final sample met criteria for at least one Axis I disorder ( 10 Agoraphobia , 33 Alcohol Abuse , 3 Substance Abuse , 25 Past Major Depressive Episode , 5 Social Phobia ) . However , as noted above , none of the participants were using psychotropic medication during the course of the DNS . Participants were scanned on one of two identical 3 Tesla General Electric MR 750 system with 50-mT/m gradients and an eight channel head coil for parallel imaging ( General Electric , Waukesha , Wisconsin , USA ) . High-resolution 3-dimensional structural images were acquired coplanar with the functional scans ( repetition time [TR] = 7 . 7 s; echo time [TE] = 3 . 0 ms; flip angle [α] = 12°; voxel size = 0 . 9 × 0 . 9 × 4 mm; field of view [FOV] = 240 mm; 34 contiguous slices ) . For the two 4 min , 16 s resting-state scans , a series of interleaved axial functional slices aligned with the anterior commissure—posterior commissure plane were acquired for whole-brain coverage using an inverse-spiral pulse sequence to reduce susceptibility artifact ( TR = 2000 ms; TE = 30 ms; α = 60°; FOV = 240 mm; voxel size = 3 . 75 × 3 . 75 × 4 mm; 34 contiguous slices ) . Four initial radiofrequency excitations were performed ( and discarded ) to achieve steady-state equilibrium . Participants were shown a blank gray screen and instructed to lie still with their eyes open , think about nothing in particular , and remain awake . Preprocessing of all resting-state fMRI data was conducted using SPM8 ( Wellcome Department of Imaging Neuroscience ) . Images for each subject were slice-time-corrected , realigned to the first volume in the time series to correct for head motion , spatially normalized into a standard stereotactic space ( Montreal Neurological Institute template ) using a 12-parameter affine model ( final resolution of functional images = 2 mm isotropic voxels ) , and smoothed with a 6 mm FWHM Gaussian filter . Low-frequency noise was attenuated by high-pass filtering with a 0 . 0078 Hz cutoff . A total of 22 subjects ( age = 26 . 04 ± 5 . 16 years [mean ± s . d . ] , 11 women ) provided informed consent and participated in the study . Data from one participant was excluded from analyses because of excessive head movement ( in excess of 1 cm ) during scanning . While no statistical test was performed to determine sample size a priori , this sample size is similar to those demonstrating a correspondence between self-report of affect and neural activity [13 , 52 , 53] . Participants engaged in an experience sampling task in which they rated their current feelings during unconstrained rest . Participants were instructed to keep their eyes open and let their mind wander freely and that a rating screen [54] would occasionally appear , which they should use to indicate the intensity of the emotion that best describes how they currently feel . This validated assay of emotional self-report consists of 16 emotion words organized radially about the center of the screen . Four circles emanate from the center of the screen to each word ( similar to a spoke of a wheel ) , which were used to indicate the intensity of each emotion by moving the cursor about the screen . During four runs of scanning , participants completed 40 trials ( 10 per run ) with an inter-stimulus interval ( ISI ) of 30 s plus pseudo-random jitter ( Poisson distribution , λ = 4 s ) . Self-report data were transformed from two-dimensional cursor locations to categorical labels . Polygonal masks were created by hand corresponding to each emotion term on the response screen . A circular mask in the center of the screen was created for neutral responses . Because terms in the standard response screen did not perfectly match those in the neural models , the item “relief” was scored as “content , ” whereas “joy” and “satisfaction” were scored as “amusement . ” The items “surprise , ” “fear , ” “anger , ” “sadness , ” and “neutral” were scored as normal . Scanning was performed on a 3 Tesla General Electric MR 750 system with 50-mT/m gradients and an eight channel head coil for parallel imaging ( General Electric , Waukesha , Wisconsin , USA ) . High-resolution images were acquired using a 3D fast SPGR BRAVO pulse sequence ( TR = 7 . 58 ms; TE = 2 . 936 ms; image matrix = 2562; α = 12°; voxel size = 1 × 1 × 1 mm; 206 contiguous slices ) for coregistration with the functional data . These structural images were aligned in the near-axial plane defined by the anterior and posterior commissures . Whole-brain functional images were acquired using a spiral-in pulse sequence with sensitivity encoding along the axial plane ( TR = 2000 ms; TE = 30 ms; image matrix = 64 × 64; α = 70°; voxel size = 3 . 8 × 3 . 8 × 3 . 8 mm; 34 contiguous slices ) . Four initial radiofrequency excitations were performed ( and discarded ) to achieve steady-state equilibrium . Processing of MR data was performed using SPM8 ( Wellcome Department of Imaging Neuroscience ) . Functional images were slice-time-corrected , spatially realigned to correct for motion artifacts , coregistered to high resolution anatomical scans , and normalized to Montreal Neurologic Institute ( MNI ) space using high-dimensional warping implemented in the VBM8 toolbox ( http://dbm . neuro . uni-jena . de/vbm . html ) . Low-frequency noise was attenuated by high-pass filtering with a 0 . 0078 Hz cutoff . To rescale data for classification , preprocessed time series were standardized by subtracting their mean and dividing by their standard deviation . Maps of partial least squares ( PLS ) regression coefficients from stimulus-evoked decoding models [13] were resliced to match the voxel size of functional data . These coefficients are conceptually similar to those in multiple linear regression , only they are computed by identifying a small number of factors ( reducing the dimensionality of the problem ) that maximize the covariance between patterns of neural activation and emotion labels ( for specifics on their computation , see [55] ) . Classifier scores were computed by taking the scalar product of functional data at each time point and PLS regression coefficients from content , amusement , surprise , fear , anger , sad , and neutral models . Individual time points were assigned categorical labels by identifying the model with the maximal score . In order to determine if relatively focal or diffuse patterns of resting-state activity informed classification , we computed importance maps for each subject ( S1 Fig ) . This was accomplished by calculating the voxel-wise product between PLS regression coefficients for each emotion model and the average activity of acquisition time points labeled as the corresponding emotion . We made inference on these maps by conducting a mass-univariate one-sample t test for each of the seven models , thresholding at FDR q = . 05 . To address the potential overlap of the emotion classification models and canonical resting-state networks of the brain , we computed the maximal Jaccard index for each emotion model and the seven most prominent resting-state networks identified in Yeo et al [26] . This index is computed as the intersection of voxels in the two maps ( voxels above threshold in both maps ) relative to their union ( the number of voxels above threshold in either map ) . Thresholds for classification models were adaptively matched to equate the proportion of voxels assigned to each resting state network . When conducting inferential tests on classification frequency ( count data ) , non-parametric tests were conducted . To test whether classifications were uniformly distributed across the emotion categories , a Friedman test was performed ( n = 499 subjects , k = 7 emotions ) . Wilcoxon signed-rank tests were performed to test for differences in frequency relative to chance rates ( 14 . 3% ) in addition to pairwise comparisons between emotion models , and corrected for multiple comparisons based on the false-discovery rate . Because the models have different levels of accuracy when used for seven-way classification [13] , we additionally conducted wavelet resampling of classifier scores in the time domain [33 , 56] over 100 iterations to ensure that differences in the sensitivity of models did not bias results . This procedure involved scrambling the wavelet coefficients ( identified using the discrete wavelet transform ) of classifier scores ( time series in Fig 3 ) to generate random time series with similar autocorrelation as the original data . Classifications were then made on these surrogate time series , and Friedman tests were performed to test for differences in frequencies across categories . This procedure yielded a null distribution for the chi-square statistic against which the observed statistic on unscrambled data was compared . To test whether classifier scores changed over time , Friedman tests were conducted on the outputs of the emotion models separately ( concatenating the time series across runs ) , as classifier scores were found to violate assumptions of normality . Follow-up tests on the direction of these changes ( either as increases or decreases ) were conducted using general linear models with one constant regressor and another for linearly increasing time for each subject . Inference on the parameter estimate for changes over time was made using a one-sample t test ( 498 degrees of freedom ) . In addition to testing gradual changes over time , smoothing spline models [25] were used to characterize more complex dynamics of emotional states . Because spline models are flexible and may include a different number of parameters for each subject , cross-validation was conducted to assess the coherence of spline fits across subjects . In this procedure , a smoothing spline model was fit for each subject , and its Pearson correlation with the mean fit for all other subjects was computed . The average of resulting correlations accordingly reflects the coherence of nonlinear changes in emotional states across all subjects . The influence of individual differences in mood and personality was assessed using generalized linear models with a binomial distribution and a logit link function . Multiple models were constructed , each using a single measure from either the CESD , STAI , or facets from the NEO-PI-R to predict the frequency of classifications for the seven emotion categories ( seven models per self-report measure ) . Inference on parameter estimates ( characterizing relationships between individual difference measures and classification frequency ) was made using a t distribution with 497 degrees of freedom . To control for multiple comparisons , FDR correction ( q = . 05 ) [57 , 58] was applied for targeted predictions . For individual differences in mood , this procedure included correction for positive associations between the frequency of sad classifications and CESD scores and between fear classification and STAI values ( Pthresh = . 0091 ) . For differences in emotional traits , correction was applied to models predicting the frequency of fear classification on the basis of Anxiety scores , anger classification using Angry Hostility scores , and sad classifications on the basis of Depression scores ( Pthresh = . 0479 ) . Scatterplots and predicted outcomes for these regression analyses are displayed in S4 Fig . To assess concordance in the experience sampling study , classifier scores were averaged for trials congruent and incongruent with self-report for each subject . For instance , all trials in which a participant self-reported “fear , ” the classifier outputs from the neural model predicting fear were considered congruent , whereas the remaining six models were averaged as incongruent . Because the frequency of self-report varied across emotions ( e . g . , endorsement of fear and sadness were very infrequent ) , scores were averaged across all trials to reduce noise . In a supplemental analysis , scores were extracted separately for all trials and classified by identifying the model with the highest score . Accuracy was assessed on data from all subjects , using self-reports of emotion as ground truth . Because the frequency of self-reported emotions was non-uniform , chance agreement between self-report and neural models was calculated based on the product of marginal frequencies , under the assumption of independent observer classifications [59] . Inference on the observed classification accuracy was tested against this value using the binomial distribution B ( 480 , 0 . 2147 ) . Due to infrequent self-reports of surprise , fear , and anger , accuracy on individual models was not computed . Scores were initially assessed by averaging the 10 s preceding each rating . Subsequent analyses increasing the window length up to 20 s did not alter results . Because the scores for congruent ( p = 0 . 0186 , Lilliefors test against normal distribution ) and incongruent ( p = 0 . 0453 ) trials exhibited non-normal distributions , Wilcoxon signed rank tests were used to test each sample against zero mean rank . The correspondence between the frequencies of classification labels from self-report and neural decoding was assessed by computing the Pearson correlation for each subject . The correlation coefficients were Fisher transformed and tested against zero using a one-sample t test . To ensure that population differences ( i . e . , inclusion of individuals with psychopathology ) did not contribute to differences in the prevalence of emotions in the resting-state and experience sampling studies , we re-calculated the frequency of classifications using repeated random subsampling of healthy participants in the resting-state sample ( 1 , 000 iterations , sampling 21 participants without replacement ) . The average correlation between the healthy subsamples and the full sample was very high ( ravg = . 981 , s . d . = . 013 ) , making it unlikely that clinical status accounts for differences in the frequency of classifications across studies .
|
Functional brain imaging techniques provide a window into neural activity underpinning diverse cognitive processes , including visual perception , decision-making , and memory , among many others . By treating functional imaging data as a pattern-recognition problem , similar to face- or character-recognition , researchers have successfully identified patterns of brain activity that predict specific mental states; for example , the kind of an object being viewed . Moreover , these methods are capable of predicting mental states in the absence of external stimulation . For example , pattern-classifiers trained on brain responses to visual stimuli can successfully predict the contents of imagery during sleep . This research shows that internally mediated brain activity can be used to infer subjective mental states; however , it is not known whether more complex emotional mental states can be decoded from neuroimaging data in the absence of experimental manipulations . Here we show that brain-based models of specific emotions can detect individual differences in mood and emotional traits and are consistent with self-reports of emotional experience during intermittent periods of wakeful rest . These findings show that the brain dynamically fluctuates among multiple distinct emotional states at rest . More practically , the results suggest that brain-based models of emotion may help assess emotional status in clinical settings , particularly in individuals incapable of providing self-report of their own emotional experience .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"diagnostic",
"radiology",
"functional",
"magnetic",
"resonance",
"imaging",
"personality",
"traits",
"social",
"sciences",
"biomarkers",
"neuroscience",
"magnetic",
"resonance",
"imaging",
"fear",
"anxiety",
"cognition",
"brain",
"mapping",
"personality",
"neuroimaging",
"research",
"and",
"analysis",
"methods",
"imaging",
"techniques",
"emotions",
"biochemistry",
"psychology",
"radiology",
"and",
"imaging",
"diagnostic",
"medicine",
"biology",
"and",
"life",
"sciences",
"cognitive",
"science"
] |
2016
|
Decoding Spontaneous Emotional States in the Human Brain
|
Calorie restriction ( CR ) , the only non-genetic intervention known to slow aging and extend life span in organisms ranging from yeast to mice , has been linked to the down-regulation of Tor , Akt , and Ras signaling . In this study , we demonstrate that the serine/threonine kinase Rim15 is required for yeast chronological life span extension caused by deficiencies in Ras2 , Tor1 , and Sch9 , and by calorie restriction . Deletion of stress resistance transcription factors Gis1 and Msn2/4 , which are positively regulated by Rim15 , also caused a major although not complete reversion of the effect of calorie restriction on life span . The deletion of both RAS2 and the Akt and S6 kinase homolog SCH9 in combination with calorie restriction caused a remarkable 10-fold life span extension , which , surprisingly , was only partially reversed by the lack of Rim15 . These results indicate that the Ras/cAMP/PKA/Rim15/Msn2/4 and the Tor/Sch9/Rim15/Gis1 pathways are major mediators of the calorie restriction-dependent stress resistance and life span extension , although additional mediators are involved . Notably , the anti-aging effect caused by the inactivation of both pathways is much more potent than that caused by CR .
The effect of restricting calorie intake on life span extension has been known for more than 70 years [1 , 2] . Although many hypotheses on how calorie restriction ( CR ) modulates aging have been proposed , the underlying mechanism for CR is still elusive [3] . Evidence from genetic studies utilizing model organisms ranging from yeast to mammals points to an important role of nutrient-sensing/insulin/insulin growth factor I ( IGF-I ) pathways in life span modulation , suggesting a common evolutionary origin of aging regulation [4] . Furthermore , these signaling pathways have been implicated in mediating CR-induced life span extension in yeast , flies , and mice [4–6] . In yeast , the conserved Ras , Tor , and Sch9 signaling pathways integrate the nutrient and other environmental cues to regulate cell growth/division [7 , 8] . Deletion of SCH9 , a homolog of mammalian AKT and S6K [9 , 10] , enhances cellular protection against thermal and oxidative challenges , and extends yeast chronological life span ( CLS , defined as the survival of non-dividing cells ) as well as replicative life span ( RLS , defined as the number of daughter cells produced by a mother cell ) [11 , 12] . Similarly , the RAS2-null strain shows increased stress resistance and survival [13–15] . Recently , evidence has been presented that deficiency in TORC1 signaling also promotes longevity in both the replicative and chronological model systems [6 , 16 , 17] . Rim15 is a glucose-repressible protein kinase and a key integrator of signals transduced by the Sch9 , Ras , and Tor pathways in response to nutrients [18–20] . Nutrient depletion activates Rim15 , which in turn upregulates the expression of a variety of genes involved in G0 entry and stress response through the transcription factors Msn2/4 and Gis1 [21] . We have previously reported that life span extension associated with deficiencies in Sch9 and Ras2/cAMP/PKA is partially mediated by enhanced cellular protection against oxidative stress through the activation of SOD2 [13] . Both the stress response element ( STRE ) and post-diauxic shift motif ( PDS ) are present in the promoter region of SOD2 , suggesting the involvement of stress response transcription factors Msn2/4 and Gis1 [22 , 23] . In fact , deletion of MSN2/4 in ras2Δ and of RIM15 in sch9Δ mutants reverses or reduces life span extension [11] . Lack of Rim15 also abolishes the life span extension associated with a reduced activity of adenylate cyclase [13] , which is found downstream of Ras2 in the Ras/PKA nutrient sensing pathway . Moreover , Msn2/4 and Rim15 are negatively regulated by the TORC1 signaling , which promotes the cytoplasmic retention of Msn2/4 and Rim15 through the interaction with the 14-3-3 protein BMH2 [24 , 25] . Genetic data also suggest that Tor inhibits protein phosphatase 2A-dependent nuclear accumulation of Msn2 in response to stresses [26] . CR delays aging and prolongs chronological and replicative life span in yeast [27–30] . For RLS studies , CR can be modeled by maintaining yeast cells on reduced glucose concentration but otherwise complete ( rich ) medium [28 , 29] . CR fails to further extend the RLS of either sch9Δ or tor1Δ mutants , indicating that down-regulation of the Tor and Sch9 pathways may mediate CR effect in dividing yeast [6] . In liquid culture , yeast cells growing in glucose containing medium release and accumulate ethanol , which promotes cell death in wild-type cells during chronological aging [30] . Switching non-dividing yeast cells from ethanol-containing medium to water , which models the extreme CR/starvation condition that yeast encounter in the wild , extends not only the mean life span of wild-type cells but also that of sch9Δ mutants , indicating the presence of additional mechanism ( s ) controlled by CR [27 , 30] . Here we present results showing that the serine/threonine kinase Rim15 and the downstream stress resistance transcription factors Msn2/4 and Gis1 are required for chronological life span extension in mutants with defects in Ras/cAMP/PKA or Tor/Sch9 signaling as well as in calorie restricted cells . In addition , we show that calorie restriction/starvation doubles the chronological life span of the extremely long-lived mutants lacking both RAS2 and SCH9 , and that this 10-fold life span extension is only partially dependent on Rim15 . Our findings are consistent with the existence of a longevity regulatory network centered on the Ras/cAMP/PKA/Rim15/Msn2/4 and Tor/Sch9/Rim15/Gis1 pathways which play important roles in the mediation of CR-dependent stress resistance and life span extension . However , our results also indicate that mutations in Tor , Sch9 , and Ras signaling in long-lived mutants do not recapitulate the full effect of CR , and both Rim15/Msn2/4/Gis1-dependent and -independent mechanisms are required to achieve maximum life span extension .
Previously , we have shown that deficiencies in Ras and Sch9 signaling pathways extend yeast chronological life span through , in part , the activation of the stress response transcription factors Msn2/4 and protein kinase Rim15 , respectively [11 , 13] . Since Rim15 has also been shown as the integrating point of the Tor and Ras/PKA nutrient-sensing pathways and an important regulator for G0 entry [21 , 25 , 31] , we examined its role in yeast chronological life span extension caused by mutations in tor1Δ and ras2Δ mutants . The mean life span of rim15Δ mutant was slightly reduced ( 12% ) compared to that of wild-type ( DBY746 ) ( Figure 1A; Table S1 ) . Deletion of RIM15 abolished life span extension associated with deficiencies in Tor1 , Ras2 , or Sch9 ( Figure 1C and 1D; Table S1 ) , suggesting that the longevity regulatory network controlled by Tor , Sch9 , and Ras converges on Rim15 . Activation of cellular protection mechanisms represents an important survival strategy in yeast [32] . We tested the role of Rim15 in cellular protection in tor1Δ , sch9Δ , and ras2Δ mutants . Cells lacking Rim15 were hypersensitive to thermal and oxidative challenges ( Figure 1E ) . Deletion of Rim15 not only abolished protection against hydrogen peroxide , and to a lesser extent to heat , in sch9Δ ( Figure 1F ) , it also abolished any beneficial effect associated with attenuated Tor signaling ( Figure 1E ) . However , Rim15-mediated stress resistance only accounted for part of the stress resistance phenotype observed in ras2Δ mutant ( Figure 1F ) . Rim15 activates Gis1 , a transcription factor that binds to the PDS element ( AWAGGGAT ) , and induces a variety of stress response genes when cells enter stationary phase [23] . To determine the contribution of Gis1 to chronological survival and cellular protection , we monitored CLS of the gis1Δ mutant as well as cells lacking GIS1 in the long-lived genetic backgrounds . gis1Δ mutant had a mean life span similar to that of wild-type yeast ( Figure 1A; Table S1 ) . In contrast , the survival of the msn2Δ msn4Δ gis1Δ triple mutant was shorter than that of wild-type and resembled that of rim15Δ ( Figure 1A; Table S1 ) , in agreement with the gene expression profile data suggesting that Msn2/4 and Gis1 cooperatively mediate the Rim15 response to glucose limitation [19 , 21] . Deficiency in Gis1 almost completely abolished the mean life span of sch9Δ mutant ( Figure 1B ) , in agreement with our earlier finding regarding the role of Rim15 in mediating the effect of sch9Δ mutation in stress resistance and life span [11] . In the RAS2-null background , the enhanced survival effect was not fully dependent on Gis1 ( Figure 1D; Table S1 ) . This observation may be explained by the fact that Msn2/4 play an important role in the life span extension associated with ras2Δ [13] . With respect to cellular protection , 1-d-old msn2Δ msn4Δ mutant was hypersensitive to both heat and oxidative stresses as expected ( Figure 2A and unpublished data ) . At day 3 , however , the mutant showed more than 10-fold increase in resistance to heat , but not to hydrogen peroxide ( Figure 2A and unpublished data ) . This phenotype was not due to an adaptive mutagenesis , as the frequency of canavanine-resistant ( canR ) mutation did not differ significantly between msn2Δ msn4Δ mutant and that of wild-type ( Figure S1 ) . Furthermore , the day 3 heat resistant msn2Δ msn4Δ cells were still sensitive to stress challenges 1 d after being re-inoculated in fresh medium ( unpublished data ) . We showed that this compensatory activation of additional cellular protection in msn2Δ msn4Δ mutant at day 3 was Rim15/Gis1-dependent since it was abolished by deletion of either RIM15 or GIS1 ( Figure 2A ) . The enhanced thermal resistance of msn2Δ msn4Δ seen at day 3 was also abolished by the overexpression of Sch9 or , to a lesser extent , the constitutively active Ras2 ( ras2val19 ) , both of which inhibit Rim15/Gis1 ( Figure 2B ) . These results depict a Ras- , Tor- , and Sch9-controlled longevity regulatory network with Rim15 in the center transducing the signals to activate stress response genes and positively regulating life span ( Figure 5B ) . It is notable that the degree of dependence on stress response transcription factors downstream of Rim15 is quite different in sch9Δ and ras2Δ mutants , with the former depending primarily on Gis1 and the latter on both Msn2/4 and Gis1 . Tor , Sch9 , and Ras/cAMP/PKA control a dynamic transcriptional network that regulates the balance between cell growth and division [7 , 8] . Whereas cells lacking SCH9 are small in size ( ∼60% of that wild-type in volume ) and display a slow growth phenotype , tor1Δ mutants are only slightly smaller than wild-type cells ( ∼86% ) and grow at a normal rate ( Figure 3A ) . This may be due to the fact that Tor2 can function , in redundancy to Tor1 , in the TORC1 complex [33] . RAS2-null cells show a small increase in cell size ( by 10% in volume ) compared to wild-type . The combination of the ras2Δ and sch9Δ instead causes a further but small decrease in cell size ( Figure 3A ) . Since all three mutants are long-lived despite differences in cell size and growth rate , it appears that chronological survival can be uncoupled from the signaling involved in regulating cell growth and size . This is particularly important considering that some of the longest-lived mutants in higher eukaryotes are dwarfs and it is not clear whether life span extension can be separated from dwarfism [4] . The down-regulation of the Sch9 , Tor , or Ras pathways has been implicated in the mediation of the CR effect on longevity [6 , 28 , 34] . We have previously shown that extreme CR/starvation , in which stationary phase cells were switched to water , doubles the mean life span of wild-type yeast [30 , 35] . Furthermore , the life span of already long-lived sch9Δ is further extended by the removal of nutrients , suggesting that either the Sch9 pathway only partially mediates the CR effect or the mechanisms underlying CR are distinct from those triggered by the deletion of SCH9 [30] . To understand the role of Tor , Ras , and Sch9 signaling in CR , we monitored the survival of tor1Δ , ras2Δ , and ras2Δ sch9Δ mutants in water . As observed with sch9Δ , starvation/extreme CR increased mean life span of both TOR1- and RAS2-null mutants ( Figure 3B; Table S2 ) . The mean ( 50% survival ) and maximum ( 10% survival ) life span was markedly increased in CR ras2Δ mutant compared to CR wild-type strain . This was not the case for tor1Δ mutant . Although CR further extended the life span of tor1Δ , there was only 18% increase in mean CLS , and no difference in maximum CLS compared to that of wild-type under extreme CR ( Table S2 ) . Considering that Rim15 is required for chronological survival extension for all three long-lived mutants , these results suggest that the Rim15-controlled Msn2/4 and Gis1 are differentially activated in tor1Δ , sch9Δ , and ras2Δ mutants . The fact that ras2Δ sch9Δ double mutant survive longer than either one of the single mutants ( Figure 3C ) supports this conclusion and suggests that the full beneficial effect of CR may be accounted by the combined effect of down-regulation of both Ras2 and Sch9 signaling . To our surprise , however , extreme CR extended the survival of ras2Δ sch9Δ double knockout mutant , which reached a mean life span of approximately 10-fold of that wild-type grown and incubated in standard glucose/ethanol medium ( Figure 3C; Table S2 ) . This suggests an additive effect between down-regulation of both the Ras/cAMP/PKA and Sch9 pathways and dietary interventions . Alternatively , Ras/cAMP/PKA signaling could be down-regulated further by the inactivation of Ras1 by CR . In fact , Ras1 and Ras2 play redundant roles in the regulation of the cAMP/PKA pathway although their expression profile is different . Unfortunately , the ras1Δ ras2Δ double mutant could not be tested because it is not viable . To elucidate the roles of Rim15 and its downstream transcription factors in CR , we monitored the stress resistance and chronological survival of cells lacking Rim15 , Gis1 , and/or Msn2/4 incubated in water . This extreme CR treatment caused a ∼10-fold increase in oxidative defense in wild-type as well as in mutants lacking Msn2/4 ( Figure 4A ) . On the other hand , the gis1Δ , msn2Δ msn4Δ gis1Δ , and rim15Δ mutations prevented the enhancement in resistance to stress ( Figure 4A ) . The commonly used CR protocol in S . cerevisiae involves a reduction in glucose concentration from 2% to either 0 . 5% or 0 . 05% , which has been shown to extend both the replicative and chronological life span [28 , 29 , 36–38] . In addition to the switch to water , we also tested the effect of the calorie restriction by reducing the glucose concentration in the growth medium from 2% to 0 . 5% . This CR intervention led to an even higher increase in the resistance to both heat shock and oxidative stress ( Figure 4A ) . These effects of calorie restriction were also completely reversed by the lack of Rim15 or all three stress resistance transcription factors MSN2 , MSN4 and GIS1 , but not by the lack of either Msn2/4 or Gis1 alone ( Figure 4A ) . Under the extreme CR condition , mean life span of the msn2Δ msn4Δ and gis1Δ did not differ significantly from that of wild-type , whereas a ∼25% reduction in maximum life span ( measured as the age when 10% of the cells were still alive ) was observed in GIS1-null mutant ( Figure 4B; Table S2 ) . Lack of all three stress response transcription factors led to a 50% reduction of maximum life span compared to wild-type ( Figure 4B ) . By contrast , extreme CR/starvation failed to extend the longevity of Rim15-null mutant ( Figure 4B ) . The results obtained under glucose reduction CR ( 0 . 5% glucose ) were very similar to those under extreme CR with the exception that wild-type cells achieved a mean life span of 31 d instead of 12 d , and the deletion of MSN2/4 had a more marked negative effect on this CR-dependent life span extension ( Figure 4C ) . Taken together , these data suggest that the serine/threonine kinase Rim15 plays a central role in mediating the effect of CR on stress resistance and life span extension by positively regulating the activities of stress resistance transcription factors Msn2/4 and Gis1 . Activation of Msn2/4 and Gis1 leads to the expression of variety of stress response genes with STRE and PDS elements in their promoters . We employed the STRE- and PDS-driven reporter gene assay to examine the gene expression changes under extreme CR condition . One-day-old wild type cells carrying either STRE- or PDS-driven lacZ reporter gene were switched to water . Significant increase in both STRE- and PDS-driven transactivation was observed 2 h after the initiation of CR compared to cells maintained in SDC medium ( Figure 4D and 4E ) . PDS-dependent transactivation increased by 90% , whereas STRE activation increased by 40% , under the extreme CR condition by 8 h . This observation is in agreement with our survival data that Gis1 plays a more important role in extreme CR-induced longevity extension ( Figure 4B ) . The statistical analysis of data derived from genome-wide motif prediction and global expression profiles provides a powerful tool to infer transcriptional regulation in the cell [39] . We have previously reported that there is significant enrichment of STRE and PDS elements in the promoter regions of the genes upregulated in sch9Δ mutant compared to wild-type under normal culture condition ( SDC ) [40] . Here , we analyzed the expression of genes containing STRE ( Msn2/4 ) or PDS ( Gis1 ) elements in their promoter under extreme CR ( switching to water ) . Our data did not indicate an enrichment of either STRE or PDS element in genes upregulated under CR ( either 24 or 48 h ) in wild-type cells ( Table 1 ) . This is probably due to the fact that CR induced a significant but small increase ( 40% to 90% ) in transactivation of Msn2/4 and Gis1 ( Figure 4D ) , which could not be detected in the analysis of array data which was performed at a cutoff of 1 . 7-fold ( CR versus SDC ) . However , CR ( water ) did cause a significant increase in the expression of STRE- and PDS-containing genes in the sch9Δ mutant ( Table 1 ) . These findings are consistent with the fact that CR further extends the life span of sch9Δ mutant , and support the notion that pathways responsible for cellular protection and life span extension in long-lived genetic mutant and in CR-treated cells are overlapping , although their levels of activation are not identical . To determine whether the life span regulatory effects caused by deficiencies in the Ras/cAMP/PKA and Sch9 pathways were additive , we studied the ras2Δ sch9Δ double mutants . Cells lacking both RAS2 and SCH9 showed a mean CLS of 35 d , which is more than 5-fold that of wild-type cells ( Figure 5A; Table S1 ) . Surprisingly , extreme CR/starvation caused an additional doubling of the life span of the ras2Δ sch9Δ ( 10-fold that of wild-type in glucose/ethanol medium ) ( Figure 5A; Table S1 ) . In view of the important role of Rim15 in life span extension in both the long-lived ras2Δ and sch9Δ mutants as well as in the CR-dependent effects , we examined the role of RIM15 in the longevity regulation by ras2Δ sch9Δ . Lack of Rim15 only partially reversed the life span extension associated with deficiencies in both Ras2 and Sch9 ( from more than 5-fold to 2 . 5-fold , Figure 5A; Table S2 ) . The reversion was even less prominent in mutants under extreme CR , where the 10-fold life span extension was reduced to 7 . 5-fold ( Figure 5A; Table S2 ) . These data indicate that Rim15-independent pro-longevity mechanisms are activated in mutants lacking Ras2 and Sch9 signaling and that their beneficial effects are further potentiated by the extreme CR intervention .
Model organisms including yeast , worms , flies , and mice have been studied extensively to understand the mechanisms of aging . Here we present genetic evidence that both CR and evolutionarily conserved signal transduction proteins implicated in life span regulation , including Ras , Tor , and Sch9 , require the serine/threonine protein kinase Rim15 and the downstream stress resistance transcription factors Msn2/4 and Gis1 to extend life span . However , additional factors appear to be involved in the remarkable 10-fold life span extension observed in calorie restricted ras2Δ sch9Δ mutants . We have previously reported that life span extension in SCH9-null and adenylate cyclase deficient mutants depends on Rim15 [11 , 13] . Here we show that deletion of RIM15 also completely abolished the life span extension as well as the stress resistance phenotype caused by the deficiencies in Ras or Tor signaling . The activity of Rim15 has been shown to involve stress response transcription factors Msn2 , Msn4 , and Gis1 [19 , 20 , 21] . Deficiency in Gis1 led to a reversion of life span extension of the sch9Δ and , to a lesser extent , ras2Δ mutants . These data are consistent with the existence of at least two major life span regulatory pathways controlled by Ras/cAMP/PKA and Tor/Sch9 , both of which converge on Rim15 . The present data also point to an important role of stress response transcription factors controlled by Rim15 , Msn2/4 , and Gis1 , in mediating the pro-longevity effect in all long-lived genetic mutants with deficiencies in nutrient sensing pathways ( Figure 5B ) . To study the CR effect on yeast chronological survival , we took two different approaches , starvation and glucose reduction . The first one models the extreme condition that yeast encounter in the wild during complete starvation periods . The extreme CR may be considered as a dietary restriction since all the nutrients in addition to calories are removed from the culture . The reduction of glucose from 2% to 0 . 5% instead is the calorie restriction regimen commonly used in RLS and CLS studies [28 , 29 , 36–38] . Both CR interventions increased cellular protection and extended chronological survival of wild-type cells , with glucose reduction showing a more powerful effect . The difference may be explained , at least in part , by the onset of CR . Unlike the starvation paradigm , in which CR was initiated after cells had entered stationary phase , cells growing in low glucose medium were exposed to CR from the very beginning . In agreement with the hormesis hypothesis of CR [34 , 41] , it is possible that the mild stress imposed by CR early in life leads to an adaptive redirection of energy and resource from growth to survival . Another possibility is that the early reduction of glucose concentration causes changes in gene expression that affect stress resistance and survival at later stages . Others have shown that CR failed to increase the replicative life span of Tor1- or Sch9-deficient mutants [6] . Our results show that the CR effect requires Tor/Sch9-controlled protein kinase Rim15 and its downstream stress response transcription factors . However , extreme CR/starvation further extended the chronological life span of the already long-lived tor1Δ , sch9Δ , and ras2Δ mutants . This difference may be the result of the very different paradigms to study aging: RLS measures the budding potential of a mother cell , whereas the CLS measures the survival of non-dividing cells . It may also be due to the CR paradigms utilized , i . e . , glucose reduction but constant exposure to 0 . 5% glucose and other nutrients ( RLS ) versus starvation in water ( CLS ) . The amino acids or other nutrients still present in the RLS paradigm may block the effect of starvation/CR on the Ras pathway and other stress resistance transcription factors . In fact , RLS extension was also achieved by decreasing the amino acid content of the medium [29] . In our CLS starvation paradigm instead , all nutrients that may contribute to the activation of pro-aging pathways are removed . The CR effect was completely reversed in cells lacking the protein kinase Rim15 but not in the msn2Δ msn4Δ gis1Δ triple mutants , suggesting the presence of additional Rim15-dependent transcriptional factor ( s ) or signaling component ( s ) yet to be identified . Forkhead family transcription factors are evolutionarily conserved from yeast to mammals and have been implicated as mediators of insulin/IGF-I/Akt signaling pathway in the regulation of anti-aging genes in worms , flies , and mammals [42] . PHA-4 , a forkhead transcription factor orthologous to the mammalian Foxa , has been shown to mediate the dietary restriction effect in C . elegans [43] . Results from our preliminary studies on the single deletion mutants of the four known forkhead TFs in S . cerevisiae ( i . e . , Fhl1 , Fkh1 , Fkh2 , and Hcm1 ) are not consistent with a major life span regulatory role of these proteins ( unpublished data ) . Instead , data presented in this study point to zinc finger transcription factors Msn2/4 and Gis1 as key components of the CR-dependent pro-longevity pathway . Based on the database search , the immediate early genes of the Egr-1 family of C2H2-type zinc-finger proteins show the highest score of homology to Msn2/4 [44] . The Egr-1 family TFs have been implicated in a variety of cellular processes including differentiation , mitogenesis , DNA repair , senescence , and apoptosis [45 , 46] . Mammalian Sp1- and Kruppel-like transcription factors are among the candidates homologous to Gis1 . They are involved in insulin- and TGFβ-signaling . Interestingly , Gis1 also contains a jumonji domain , which is first described as a bipartite protein domain present in many eukaryotic transcription factors [47] . Recent evidence from several organisms has shown that a number of jmjC domain-containing proteins are histone demethylases , suggesting a role of Jumonji-domain–containing protein in chromatin remodeling [48] . Interestingly , the DNA binding activities of Egr-1 , Sp1 , and other zinc-finger TFs are sensitive to cellular redox state , and their dysfunction during aging may lead to age-associated pathophysiology [49–52] . While the existence of conserved domains in these yeast proteins is encouraging , it is still premature to speculate about their mammalian counterparts . Although the protein kinase Rim15 is required for life span extension in Ras2 , Tor1 , and Sch9-deficient mutants as well as in yeast under CR , our results indicate that pathways responsible for enhancing stress protection and life span extension in nutrient sensing-impaired genetic mutants and in cells under CR are not identical . On the one hand , the “full” activation of Rim15 and its downstream transcription factors , Msn2/4 and Gis1 , are required for the maximum life span extension , as the pro-longevity effects of ras2Δ , sch9Δ , and CR are additive ( Figures 3B and 3C , and 5A ) . On the other hand , Rim15 only accounts for part of the beneficial effect for ras2Δ sch9Δ mutant under CR , implicating the involvement of additional pro-survival mechanism ( s ) independent of the Rim15-centered nutrient-sensing pathways ( Figure 5A ) . Similar observations were also made in other model systems: CR can further increase the life span of the already long-lived Ames dwarf mice [53]; and it further extends the life span of insulin/IGF-I signaling-impaired chico flies [5] . We and others had shown that the down-regulation of Ras/cAMP/PKA signaling extends the yeast chronological and replicative life span [11 , 13 , 14 , 28] . However , the mammalian cAMP/PKA was only very recently implicated in the regulation of longevity in mice . The type 5 adenylyl cyclase knockout ( AC5-KO ) mice live 30% longer than their wild-type littermates [54] . CA5-KO mice do not show dwarfism , although they weigh slightly less than age-matched controls at 28 months . Similarly to mutants lacking Ras2 or with a reduced adenylate cyclase activity , mouse cells with CA5 disruption show enhanced resistance to oxidative stress , which may be mediated by the upregulation of MnSOD [13 , 54] . Interestingly , the CA5-KO mice have lower growth hormone level [54] , suggesting an attenuated GH/insulin/Akt signaling in these mice . In view of our yeast data showing that CR in combination with the down-regulation of the Ras/cAMP/PKA and Sch9 pathways reached a 10-fold life span increase , it will be interesting to determine the interaction between the insulin/Akt and Ras/cAMP/PKA pathways as well as their combined effect with CR in regulating life span in mammals . Considering the fact that Ras and Sch9 signaling pathways are partially conserved from yeast to mammal ( Figure 5B ) , it will also be important to explore the possibility that potential orthologs of Rim15 and of Msn2/4 and Gis1 may modulate aging in high eukaryotes .
All strains used in this study are derivatives of DBY746 ( MATα leu2–3 , 112 , his3Δ , trp1–289 , ura3–52 , GAL+ ) . Knockout strains were generated by one-step gene replacement as described previously [55] . Strains overexpressing SCH9 or ras2val19 were generated by transforming cells with plasmids pHA3-SCH9 ( a gift from Dr . Morano ) , or pMW101 ( plasmid RS416 carrying ClaI-ras2val19-HindIII fragment form pMF100 , a gift from Dr . Broach ) , respectively . For strains used in STRE- and PDS-lacZ reporter gene assay , the plasmid pCDV454 containing LacZ reporter under the control of a 37 bp SSA3-PDS region ( −206 to −170 ) [23] or the plasmid pMM2 containing four tandem repeats of STRE motif from the HSP12 sequence ( −221 to −241 ) [56] , was integrated into the URA3 locus of wild-type cells . The transcriptional specificity of these reporter genes were confirmed in the msn2Δ msn4Δ and gis1Δ background , respectively ( unpublished data ) . Yeast cells were grown in SDC supplemented with a 4-fold excess of the tryptophan , leucine , uracil , and histidine to avoid possible artifacts due to auxotrophic deficiencies of the strains . Yeast chronological life span was measured as previously described [11 , 57] . Briefly , overnight SDC culture was diluted ( 1:200 ) in to fresh SDC medium to a final volume of 10 ml ( with flask to culture volume of 5:1 ) and were maintained at 30 °C with shaking ( 200 rpm ) to ensure proper aeration . This time point was considered day 0 . Every 2 d , aliquots from the culture were properly diluted and plated on to YPD plates . The YPD plates were incubated at 30 °C for 2 d to 3 d , and viability was accessed by Colony Forming Units ( CFUs ) . Viability at day 3 , when the yeast had reached the stationary phase , was considered to be the initial survival ( 100% ) . Mean and maximum life span ( 10% survival ) was calculated from curve fitting ( one phase exponential decay ) of the survival data ( form pair matched , pooled experiments ) with the statistical software Prism ( GraphPad Software ) . For extreme CR/starvation , cells from 3-d-old SDC culture were washed three times with sterile distilled water , and resuspended in water . Water cultures were maintained at 30 °C with shaking . Every 2 d to 4 d , cells from the water cultures were washed to remove nutrients released from dead cells . For CR modeled by glucose reduction , overnight SDC culture was diluted ( 1:200 ) into fresh SC medium supplemented with 0 . 5% glucose . It is notable that the glucose reduction model employed here is different from that in replicative life span ( RLS ) studies . For RLS analysis , cells are maintained on reduced glucose but otherwise complete ( rich ) medium . In liquid chronological cultures , extracellular glucose was exhausted by day 1 in SC + 0 . 5% glucose as well as standard SDC cultures ( unpublished data ) . Unlike the extreme low glucose cultures ( SC + 0 . 05% glucose ) which reached saturation density of only a quarter of that standard SDC ones , there was no difference in saturation density between cultures with 2% and 0 . 5% glucose , suggesting 0 . 5% glucose is not a limiting factor on cell growth/division ( unpublished data ) . Heat shock resistance was measured by spotting serial dilutions ( 10-fold dilution started at OD600 of 10 ) of cells removed from SDC cultures onto YPD plates and incubating at either 55 °C ( heat-shocked ) or 30 °C ( control ) for 45 min to 150 min . After the heat-shock , plates were transferred to 30 °C and incubated for 2 d to 3 d . For oxidative stress resistance assays , cells were diluted to an OD600 of 1 in K-phosphate buffer , pH6 . 0 , and treated with 100 mM to 200 mM of hydrogen peroxide for 60 min . Serially diluted ( 10-fold ) control or treated cells were spotted onto YPD plates and incubated at 30 °C for 2 d to 3 d . Day1 SDC cultures were mixed with equal volume of 2× calcofluor ( 75 ng/ml in PBS , Molecular Probe ) . After 10 min incubation at room temperature in the dark , cells were washed once with PBS . Images were captured with a Leica fluorescence microscope . Diameter of the cell was measured using ImageJ ( http://rsb . info . nih . gov/ij/ ) . Cells were measured at long and short ( perpendicular ) axes . Diameter was expressed as the average of the long and short axes of the cell . 50 to 80 cells per genotype were measured . Day 1 wild-type cells carrying the STRE- or PDS-lacZ reporter gene ( grown in SDC ) were split into two portions . One was washed three times with sterile water and resuspended in water; the other was maintained in the original SDC medium . Cells were collected at 2 h , 4 h , and 8 h after the initiation of CR . Cell pellet from 1 ml of culture was lysed with Y-PER ( Pierce ) according to manufacturer's protocol . The protein concentration of the lysate was assayed with a BCA kit ( Pierce ) . 55 μl of lysate was mixed with 85 μl of substrate solution ( 1 . 1mg/ml ONPG in 60 mM Na2HPO4 , 40 mM NaH2PO4 , 10 mM KCl , 1 mM MgSO4 , 50 mM 2-mercaptoethanol , pH7 . 0 ) . Absorbance at 420 nm was read every 5 min until 30 min after the initiation of reaction . LacZ activities were determined by fitting the A420/time data to that of serial diluted recombinant β-galactosidase ( Promega ) . LacZ activity was normalized to the total protein in the lysate . A slight modified CR protocol was adopted , where 1 . 5-d-old cells were washed three times and incubated in water . 24 h and 48 h later , cells were collected for RNA extraction . These time points ( to obtain RNA samples at day 2 . 5 and day 3 . 5 ) were selected to avoid the general decrease in metabolism and consequently in gene expression that normally occurs at older ages ( day 4 to day 5 ) [57] . The cRNA generated from these samples was hybridized to Affymetrix GeneChip Yeast 2 . 0 array to obtain the measurement of gene expression . The “Invariant Set” approach was used for normalization at the probe level , and the “Model based” method to summarize and obtain expression for each probe set [58] . A detailed method for motif prediction and motif enrichment test has been described previously [40] . Briefly , for a given gene , if one or more binding sites of a transcription factor ( TF ) binding motif were found within the 800 bp region upstream of the start codon , it was defined as the target gene of that TF . A total of 51 motifs that can be associated with known TFs were used for motif prediction in all known yeast ORFs [39] . The cut-off value of motif matching score was set to 0 . 6 . The hypergometric test was employed to determine whether there was an enrichment of any motif in CR-induced genes ( upregulated more than 1 . 7-fold ) . Finally , we calculated the q-values for each test to correct the multiple testing errors using the “qvalue” package [59] .
Genes examined in this study from the Saccharomyces Genome Database ( http://db . yeastgenome . org/ ) are as follows: SCH9 , ( YHR205W ) ; RAS2 ( YNL098C ) ; TOR1 ( YJR066W ) ; RIM15 ( YFL033C ) ; MSN2 ( YMR037C ) ; MSN4 ( YKL062W ) ; and GIS1 ( YDR096W ) .
|
Reduction in calorie intake is a well-established intervention that extends the life span of a variety of biological model organisms studied . Calorie restriction also delays and attenuates age-related changes in primates , although its longevity-promoting effect has not been demonstrated . Here , we utilized a single cell organism , baker's yeast , to examine the role of evolutionarily conserved genes in life span regulation and their involvement in calorie restriction . The yeast mutants lacking Ras2 , Tor1 , or Sch9 are long-lived . The anti-aging effect observed in these mutants depends on the protein Rim15 and several key regulators of gene expression that are essential in inducing cellular protection under stress . The beneficial effects of calorie restriction are much smaller in yeast that are missing these proteins , indicating their essential role in promoting longevity . Our study also showed that by combining the genetic manipulation and calorie restriction intervention , yeast can reach a life span ten times that of those grown under standard conditions . This extreme longevity requires Rim15 and also depends on other yet-to-be identified mechanisms . Our findings provided new leads that may help to elucidate the mechanisms underlying the anti-aging effect of calorie restriction in mammals .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
] |
[
"genetics",
"and",
"genomics",
"cell",
"biology",
"developmental",
"biology",
"saccharomyces"
] |
2008
|
Life Span Extension by Calorie Restriction Depends on Rim15 and Transcription Factors Downstream of Ras/PKA, Tor, and Sch9
|
In sub-Saharan Africa , systematic surveillance of young children with suspected invasive bacterial disease ( e . g . , septicemia , meningitis ) has revealed non-typhoidal Salmonella ( NTS ) to be a major pathogen exhibiting high case fatality ( ~20% ) . Where infant vaccination against Haemophilus influenzae type b ( Hib ) and Streptococcus pneumoniae has been introduced to prevent invasive disease caused by these pathogens , as in Bamako , Mali , their burden has decreased markedly . In parallel , NTS has become the predominant invasive bacterial pathogen in children aged <5 years . While NTS is believed to be acquired orally via contaminated food/water , epidemiologic studies have failed to identify the reservoir of infection or vehicles of transmission . This has precluded targeting food chain interventions to diminish disease transmission but conversely has fostered the development of vaccines to prevent invasive NTS ( iNTS ) disease . We developed a mathematical model to estimate the potential impact of NTS vaccination programs in Bamako . A Markov chain transmission model was developed utilizing age-specific Bamako demographic data and hospital surveillance data for iNTS disease in children aged <5 years and assuming vaccine coverage and efficacy similar to the existing , successfully implemented , Hib vaccine . Annual iNTS hospitalizations and deaths in children <5 years , with and without a Salmonella Enteritidis/Salmonella Typhimurium vaccine , were the model’s outcomes of interest . Per the model , high coverage/high efficacy iNTS vaccination programs would drastically diminish iNTS disease except among infants age <8 weeks . The public health impact of NTS vaccination shifts as disease burden , vaccine coverage , and serovar distribution vary . Our model shows that implementing an iNTS vaccine through an analogous strategy to the Hib vaccination program in Bamako would markedly reduce cases and deaths due to iNTS among the pediatric population . The model can be adjusted for use elsewhere in Africa where NTS epidemiologic patterns , serovar prevalence , and immunization schedules differ from Bamako .
In industrialized countries , non-typhoidal Salmonella ( NTS ) predominately causes gastroenteritis [1 , 2] . However , in sub-Saharan Africa the NTS serovars S . Typhimurium and S . Enteritidis have become recognized as important causes of severe invasive bacterial disease ( e . g . , septicemia , meningitis , bacteremia ) with high case fatality rates [2 , 3 , 4] . Infants age 6–11 months and toddlers age 12–23 months exhibit the highest incidence of severe invasive NTS ( iNTS ) disease [5] . Whereas host factors such as malnutrition and co-infection with malaria and HIV may contribute to the higher burden of iNTS disease ( i . e . , high case fatality rate and prevalent cause of bacteremia ) in this region compared to industrialized countries [6–7] , fundamental differences in the circulating NTS strains from sub-Saharan Africa are also evident . Available evidence suggests that the vast majority of the S . Typhimurium strains from cases of iNTS disease in sub-Saharan Africa are multi-locus sequence type 313 ( ST313 ) , a genotype unique to Africa that has undergone extensive genomic degradation [8–9] . As the burdens of invasive disease due to Haemophilus influenzae type b ( Hib ) and Streptococcus pneumoniae have plummeted in recent years in sub-Saharan Africa following the introduction of Hib conjugate and multivalent pneumococcal conjugate vaccines [10–11] , recognition of the need to address iNTS disease has increased [12] . Lack of information on the reservoirs and vehicles of transmission of iNTS in sub-Saharan Africa limits opportunities to utilize classic epidemiologic interventions to control iNTS disease . However , successful vaccination programs implemented to control other invasive diseases prevalent among pediatric populations in Mali and other countries of sub-Saharan Africa have stimulated interest in the development of vaccines to control iNTS disease . Several candidate vaccines under development have shown promise in protecting against invasive S . Typhimurium and S . Enteritidis disease in animal models [13 , 14] . These include a bivalent conjugate vaccine based on covalently linking the core and O-antigen polysaccharides of S . Typhimurium ( a Group B [O:4] serovar ) and S . Enteritidis ( a Group D [O:9] serovar ) to the respective Phase 1 flagellin subunits ( FliC ) of each of these serovars [14–16] , a live attenuated oral vaccine [17–18] , and a bivalent Generalized Modules for Membrane Antigens ( GMMA ) vaccine consisting of outer membrane protein blebs from S . Typhimurium and S . Enteritidis that include lipopolysaccharide [12 , 14] . These vaccines also have the potential to provide cross protection against other NTS serovars within Salmonella O Group B ( e . g . , S . Stanleyville ) and O Group D ( e . g . , S . Dublin ) [13 , 16 , 18] . Our research modeled the decrease in the number of cases and deaths attributable to iNTS in children < 5 years of age following the programmatic introduction of a NTS vaccine utilizing the same Expanded Program on Immunization ( EPI ) infrastructure that successfully delivered Hib and pneumococcal conjugate vaccines and that drastically reduced the number of cases of invasive disease caused by those pathogens .
Census data from the National Institute of Statistics ( INSTAT ) of Mali from 2009 [19] provided the pediatric population of different age groups of interest in Bamako as denominators for the model . Crude birth rate and age-specific all-cause mortality data for specific pediatric age groups in Bamako came from Demographic and Health Surveys ( DHS ) of 2001 , 2006 and 2012 [20–22] ( Table 1 ) . The highest and lowest rates reported across the years of DHS reports were used to establish a range of probable values with the intermediate of the three values used as the initial parameter value . The burden of invasive bacterial disease caused by NTS in Mali was first identified during systematic surveillance of the incidence of bacterial pathogens begun in 2002 at l’Hôpital Gabriel Touré ( HGT ) , Bamako , Mali . The surveillance program established by the Center for Vaccine Development , Mali ( CVD-Mali ) and the Center for Vaccine Development ( CVD ) , University of Maryland School of Medicine , was designed to identify bacterial pathogens associated with invasive disease among consented enrolled patients <15 years of age admitted to HGT with fever or clinical signs of invasive bacterial disease [5] . This hospital-based surveillance was conducted under a protocol approved by the Ethics Committee of the Faculté de Médecine , Pharmacie et Odontostomatologie in Bamako , Mali and the University of Maryland Institutional Review Board . Consent was documented on a written form . If the participant's parent or guardian was illiterate , they listened to an audiotaped version of the consent form in their local language and questions were so answered in the presence of a witness . We used anonymized data on 515 pediatric patients under five years of age who were admitted to HGT with laboratory-confirmed iNTS disease between July 1 , 2002 and June 30 , 2014 to develop and validate the model parameters . Numbers of cases within specific age groups , pooled across even years of the HGT surveillance ( i . e . , 2002 , 2004 , 2006 , 2008 , 2010 , 2012 ) , were used with denominators from INSTAT [19] to generate the age group-specific hospitalization rates of severe iNTS disease ( Table 2 ) . Case fatality rates for the model were fatal cases divided by total cases per age group ( Table 3 ) . The model was then validated by comparing the number of cases and deaths due to iNTS per year estimated by the model , without accounting for vaccination effects , against the data from the odd years of HGT surveillance . Data were pooled as means across the years of surveillance because the number of cases per age-group per year was small ( Fig 1 ) . Multiple years of data were included to provide more robust estimates for the Malian pediatric population and to encapsulate some of the variability over time . The proportion of hospitalizations with S . Typhimurium and S . Enteritidis serovars has been seen to change over time . In particular , from 2008 to the present , the incidence of invasive S . Typhimurium infections has decreased , while the incidence of invasive S . Enteritidis infections has increased [5] . Moreover , these serovars exhibit different case fatality rates . Therefore , serovar-specific case fatality rates for hospitalized children were also calculated based on the HGT surveillance data . The expected coverage for an iNTS vaccine was estimated based on data from Hib vaccine implementation in Bamako . Vaccination coverage estimates came from an immunization coverage survey undertaken in 2015 among a sample of infants 6–8 months of age in the population as part of prospective demographic surveillance in the Djikoroni-para quartier of Bamako . The demographic surveillance system allowed population-based estimates to be derived as was done for the Global Enteric Multicenter Study ( GEMS ) [23–24] . Sixty-one mothers or other caretakers of infants 6–8 months of age were asked if they had an immunization card and 60 were able to show the card . The narrow infant age range was selected to document not only evidence of receipt of Hib vaccine but to provide information on the timeliness of immunization which is important to the success of Hib and NTS vaccination as a public health tool . Among these 61 Djikoroni-para infants , 60 had received at least one dose of Hib vaccine 60/61 ( 98 . 4% ) and 55 had received all three doses of Hib vaccine ( 55/61 , 90 . 2% full coverage ) . This Hib vaccine coverage information from Bamako was used as the starting point for our simulations , since it was drawn directly from our modeled population . Hib coverage data from Kenya [26] was utilized to generate wider intervals of coverage values from a larger study sample and broader population data . Coverage data for alternative vaccination programs and booster vaccinations were based on the measles vaccine program implemented in Mali , which targeted children of the same scheduled ages as proposed in the model [33] . Assumptions on the expected efficacy of the vaccines under development to prevent iNTS disease in Mali were based on assessments of the efficacy of Hib conjugate in a randomized clinical trial in The Gambia [24] and from post-licensure impact evaluations on Hib disease in Mali [10] , Kenya [26] and Uganda [29] and a 9-valent pneumococcal conjugate vaccine efficacy trial in The Gambia [30] . Hib conjugate was highly effective in diminishing the disease burden when administered routinely through the Expanded Program on Immunization in Mali [10] , Kenya [26] and other African countries [29] . Efficacy for booster vaccination doses and a catch-up campaign program were based on anticipated results similar to those exhibited by the Hib catch up campaign and booster vaccine interventions performed in the United Kingdom [34–35] . In certain populations , such as those with a high prevalence of HIV cases , immune suppression decreases the amount of protection granted by vaccination against Hib [25] . While the pediatric population of Bamako does not exhibit high levels of HIV , a scenario with low vaccine efficacy due to immune suppression such as seen by Madhi et al . , in South Africa ( a 20% decrease in each vaccine efficacy parameter ) was modeled . The invasive disease such as meningitis , septicemia , bacteremia and septic arthritis caused by invasive non-typhoidal Salmonella is clinically indistinguishable from those types of clinical infections caused by Hib . Each of these pathogens traverses a mucosal barrier leading to a bacteremia during which the bacterial pathogens are cleared by fixed macrophages residing in organs of the reticuloendothelial system . In the case of Hib , it is upper respiratory mucosa that is traversed , while for iNTS it is believed to be intestinal mucosa . Bacteremic organisms that reach the meninges , synovia and pleura can cause meningitis , septic arthritis and empyema , respectively . The NTS conjugate vaccines under development elicit serum antibodies that exhibit both bactericidal and opsonophagocytic functional properties [31–32 , 36] , like the antibodies stimulated by Hib conjugate vaccines [25 , 37–40] . Thus , there exist striking pathogenetic , clinical and epidemiologic similarities between iNTS and Hib pathogens and similar functional activities are exhibited by the antibodies stimulated by the parenteral NTS conjugate vaccines ( in animals ) and by Hib conjugate vaccine in human infants . Therefore , we assumed a similar efficacy and coverage for the NTS vaccine as was observed with Hib conjugate vaccine in the infant and toddler population in Bamako ( and elsewhere in sub-Saharan Africa ) . Relying on Hib vaccination efficacy data allowed us to validate the iNTS vaccine implementation within the model and allowed for reliable comparisons of the protection potentially granted by the iNTS vaccine and various immunization schedules . The incidence and epidemiologic features of iNTS infections among young children sufficiently severe to result in hospitalization were captured using an age-structured Markov chain infectious disease model including Susceptible , Infected , and Recovered status groups . A diagram of the model ( Fig 2A ) with vaccine administered at 6 , 10 , and 14 weeks of life , as was used in the Hib vaccination initiative , was used as the baseline for developing a generalized model capturing all modeled vaccination schedules as illustrated in Fig 2B . Each age group ( a ) included an age-specific number of children susceptible to ( S[a] , 10 groups , where a = 1–10 ) , hospitalized with ( I[a] , 4 groups , where a = 1–3 , 4–8 , 9 , 10 ) , or vaccinated by dose ( d ) against ( V[a , d] , where a varied for different scenarios and d = 1–5 ) severe iNTS disease . Children who recovered ( R ) from severe hospitalization were considered as a single group . Age categories were established by examining statistically significant variation in incidence and case fatality rates of iNTS ( p<0 . 05 ) within the HGT surveillance data , and ages at which the EPI vaccinations were scheduled . Neonates entered the model based on the population birth rate reported for Bamako ( ν , Table 1 ) and were considered to be hospitalized with iNTS disease at the same incidence as other infants < 2 months of age ( Table 2 ) . Children at any age were subject to the age-specific all-cause mortality rates ( μ[a] , Table 1 ) reported in the DHS . Susceptible children ( S[a] ) were moved to a hospitalized status ( I[a] ) at age-specific iNTS hospitalization rates ( β[a] , Table 2 ) . Children hospitalized with iNTS disease experienced mortality at age-specific case fatality rates ( η[a] , Table 2 ) or moved to a recovered status ( R ) . The length of iNTS infection was constrained to a single two-week time step corresponding to the observed duration of iNTS clinical disease in hospitalized Bamako children . Any susceptible children who did not suffer hospitalized iNTS disease , iNTS fatality , or all-cause mortality graduated into the next susceptible age group at a rate appropriate to the two-week time step of the model . All surviving children exited the model upon reaching five years of age , when the rate of hospitalizations due to iNTS rapidly declines [5] . Vaccination against iNTS was initially modeled as a program which occurred at ~6 , ~10 , and ~14 ( a = 2 , 4 , 6 , respectively ) weeks of life to match the same three-dose infant immunization schedule as Hib conjugate . Children potentially received one , two , or three ( dose d = 1–3 , respectively ) doses of the vaccine with age-specific and dose-specific coverage rates ( ϑ[a , d] ) . For vaccinated children , the rate of hospitalization due to iNTS disease was applied only to the proportion of children without an effective vaccination ( 1-ε[d] ) . Vaccine protection was assumed to persist through early childhood , so successfully immunized children remained protected until they aged out of the model . Our major outcome measures were the number of cases and deaths due to iNTS disease . To generate an overall distribution of potential values describing the natural epidemiologic behavior of the disease without implementation of a vaccination program , 1000 simulations were run with key model variables ( i . e . , birth rate , incidence of hospitalization of iNTS cases , case fatality , and background all-cause mortality rates ) randomly drawn each time from a uniform distribution based on the limits of the probable range around each parameter . After being used to simulate the ‘average’ dynamics in absence of vaccine use , a second set of 1000 simulations was used to assess the effect of different assumptions about vaccine efficacy and coverage in the model . Key variables were sampled as before and values for vaccine efficacy and coverage drawn from triangular distributions based on each parameter value and its associated probable range under different scenarios . The influence of each parameter on the number of iNTS cases and fatal iNTS cases generated by the model was assessed by sampling each parameter individually , while holding all other parameter values constant . The ranges of iNTS cases and fatal iNTS cases generated by these simulations were summarized in tornado plots . Next , specific effects of vaccination were examined in a birth cohort of 75 , 978 children , corresponding to the annual births for Bamako . The expected number of cases and number of fatal cases in children < 36 months of age within this cohort was generated under unvaccinated conditions , vaccinated conditions with 100% coverage and efficacy implemented at 6 weeks of life , and vaccinated conditions with coverage and efficacy matching the model parameters described in Table 4 at 6 , 10 , and 14 weeks of life . Additionally , the effects of high , mid , and low vaccine efficacy levels on the overall number of cases and case fatalities were examined by dose , based on the ranges around these parameters . To assess the effects of serovar- specific severity , the vaccination program was modeled using case fatality rates representing only S . Typhimurium or only S . Enteritidis as causal agents ( Table 3 ) . Since the overall incidence of iNTS disease has not changed significantly with the documented serovar shift , we maintained the same overall incidence rates regardless of underlying causal agent . Another 1000 simulations were performed to assess effects of varying the parameter values across the probable ranges for case fatality rates of these two serovars on the number of severe iNTS cases and fatal iNTS cases . The same approach was used to simulate effects of alternative three-dose EPI vaccination regimens . For example , the first two doses were administered at 6 ( a = 2 ) and 10 ( a = 4 ) weeks or at 10 ( a = 4 ) and 14 ( a = 6 ) weeks of life ( concomitant with two doses of pentavalent and pneumococcal conjugate vaccine ) and the third NTS vaccine dose was administered as a booster ( dose d = 4 ) at either age 9 months ( a = 7 ) ( with measles containing vaccine dose 1 [MCV1] ) or at 12 or 15 ( a = 9 ) months of age concomitant with MCV2 . We also modeled a rapid mass immunization catch-up campaign ( dose d = 5 ) targeting all children of age 6–23 ( a = 6–9 ) , 9–23 ( a = 7–9 ) , or 12–23 ( a = 9 ) months of age concomitant with the onset of adding iNTS vaccination to the routine young infant EPI . In each of these schedules , the precise ages at time of vaccination , although ideally targeted at specific weeks of life , in fact vary and are often delayed by several weeks in Bamako . To allow for this , the number of children admitted to the HGT who received one , two , or three doses of vaccine any time within the relevant month of life was used to calculate the sample mean coverage levels among hospital admissions , and not by specific week of implementation . These coverage levels fell within the confidence intervals of the Kenyan coverage data reported by Cowgill et al . [26] that were used to parameterize the model . Model development and analyses were performed using R version 3 . 0 . 1 and utilizing the Markovchain package , version 0 . 5 [27] for the development of the model and the Triangle package , version 0 . 1 [28] for vaccine efficacy and vaccine coverage variable distribution analysis . The code developed for the model and the case data used to develop the model parameters are available on a public GitHub repository [41] .
Based on the parameter values determined from even years , our model predicted a similar number of hospitalized iNTS disease cases for most age groups as was observed during the odd years of HGT surveillance data ( Fig 1 ) , with 37 cases per year , including 7 fatal cases , occurring in a non-vaccinated population . The 1000 model runs generated a range of 14–64 cases per year ( with interquartiles of 29 and 39 ) and a range of 2–14 fatal cases per year ( with interquartiles of 5 and 8 ) , sampling from the probable range of model parameters . Hospital surveillance records did not include neonatal cases who died of iNTS infection before leaving the hospital after birth , so our model assumed a similar level of incidence among this youngest age group as was observed for other children less than one month old . This assumption generated a simulated overestimation of cases in the youngest age group compared to the observed , but attempted to include neonatal cases and iNTS related deaths in our outcome measurements . Varying the parameter estimates used for hospitalized infection rate , all-cause mortality rate , and case fatality rate across the range of each parameter led to a mean of 34 cases per year ( ranging from 14–64 , with interquartiles of 29 and 39 ) and 7 fatal cases per year ( ranging from 2–14 , with interquartiles of 5 and 8 ) , based on 1000 runs of the model . One thousand model runs with varying parameter estimates for birth rate , incidence of hospitalization of iNTS cases , case fatality , background all-cause mortality rates , vaccine efficacy and coverage with three doses of vaccine administered at 6 , 10 , and 14 weeks of life generated a mean of 9 cases per year ( ranging from 5–16 , with interquartiles of 11 and 6 ) and 3 fatal cases per year ( ranging from 2–12 , with interquartiles of 2 and 5 ) . The greatest change in the number of iNTS cases occurred as the incidence rate was sampled across its probable range , while the least amount of change was generated by sampling across vaccine coverage ( Fig 3 ) . The greatest change in the number of fatal iNTS cases was driven by the case fatality rate ( Fig 4 ) . In observing the total number of cases and the number of deaths due to iNTS among a birth cohort , as presented in Fig 5 , the vaccination parameters of the model functioned as expected . If a NTS vaccine was implemented with 100% coverage and 100% efficacy , all cases following an initial dose of vaccine were prevented and all fatal cases were averted . Furthermore , when the model was run with parameters based on the Hib vaccination field trials and post-introduction impact assessments in Africa , the results were very similar to a vaccine with perfect coverage and efficacy . Almost all cases among the birth cohort were prevented even after a single dose , and all cases in the cohort were prevented after two doses . Modeling vaccine coverage and efficacy for an iNTS vaccine equivalent to levels observed with Hib ( with routine young infant immunization only and with no catch-up campaign ) , the number of hospitalized iNTS cases per year decreased by 73% ( from 37 to 10 cases ) and the number of deaths decreased by 43% ( from 7 to 4 deaths ) . These estimates , based on the distribution of S . Enteritidis and S . Typhimurium cases that was observed during the 2002–2014 surveillance , reflect the effect of direct protection alone ( i . e . , with no adjustment for indirect protection from “herd immunity” ) and without a catch-up campaign . Effects of varying the parameter values for vaccine efficacy , number of doses , serovar distributions , and vaccination schedules are shown in Table 5 . Even at the lowest ranges of vaccine efficacy , our model predicted a range of only 4–23 cases per year , based on 1000 simulation runs . This equates to prevention of more than half of the pediatric cases each year if the vaccine exhibits similar efficacy as might be seen with the Hib vaccine in immunodeficient populations . At higher levels of efficacy , as much as 78% of severe iNTS cases ( 29 cases/year ) were averted . The highest level of protection was granted by a 3-dose vaccination schedule targeting infants at 6 , 10 , and 14 weeks of life , which resulted in a 73% decrease in the annual number of severe iNTS cases with moderate vaccine efficacy levels . This varied from 21–29 cases prevented and 2–3 child lives saved per year among the pediatric population of Bamako , based on the vaccine efficacy levels observed in the Hib conjugate field trial in The Gambia [24] and the post-implementation effectiveness assessment in Mali [10] . The increased immunity over time among the population was captured through the number of hospitalized iNTS cases and deaths per 6-month intervals after vaccine implementation , compared to a 6-month pre-vaccination baseline interval , as illustrated in Fig 6 . If a catch-up vaccination campaign was implemented simultaneously along with the introduction of a three-dose young infant EPI vaccination strategy , protection occured sooner among older age children who remained at risk . Fig 7 illustrates the effects of such a catch-up campaign targeting various age groups . The addition of the catch-up campaign prevented at least four and as many as 12 additional cases of severe iNTS during the first three years following the catch-up campaign and start of the vaccine initiative compared to the three-dose schedule without the catch-up campaign . Investigating the effects of different serovar distributions that have been observed over time , our model predicted that if S . Enteritidis , with its higher case fatality rate , was the only iNTS serovar causing disease , 44% of deaths per year would be averted through vaccination , with a range of 5–37 cases per year including 2–18 fatal cases , across 1000 simulation runs . If S . Typhimurium returned as the dominant serovar , up to 80% of iNTS deaths per year would be averted , with a range of 4–31 cases per year , including 1–8 fatal cases .
Implementation of programmatic use of the Hib conjugate vaccine among the pediatric population in Bamako , Mali was extremely successful and reduced the burden of invasive Hib disease hospitalizations by 83% among infants within three years of vaccine introduction [10] . Our results show that if we utilize the same EPI infrastructure to deliver a bivalent conjugate vaccine to prevent invasive disease due to S . Enteritidis and S . Typhimurium , it is reasonable to expect a comparable level of protection and reduction of iNTS cases as was seen with Hib vaccine . Our iNTS model captured the current burden of iNTS in the population and assessed potential effects of various vaccine implementation scenarios . Even alternative vaccination schedules with fewer doses of iNTS vaccine predicted a notable reduction in the burden of iNTS disease . Additionally , if a one-time catch-up campaign is implemented concomitantly with routine vaccination of young infants and if it targets the highest risk age group ( children in the second half of the first year of life ) , our model predicted additional cases of severe iNTS disease deaths would be averted in the first years after introducing the vaccine . If field trials confirm the efficacy of the NTS vaccines currently under development , their future implementation within the EPI could markedly diminish the morbidity and mortality from one of the predominant remaining burdens of invasive bacterial disease among pediatric populations in sub-Saharan Africa . Since neither the reservoir of infection nor the modes of transmission of NTS to young children have heretofore been identified , vaccination currently represents the most plausible interventional strategy for reducing the burden of iNTS disease . Furthermore , the model we have developed could be applied to estimate the effects of implementing an iNTS vaccine in other regions of sub-Saharan Africa , providing the same integrity of information on age-specific case and fatality rates . Two candidate NTS vaccines are progressing towards clinical trials . One candidate developed by the GSK Vaccines Institute of Global Health consists of a bivalent parenteral Salmonella Enteritidis/S . Typhimurium vaccine based on Generalized Modules for Membrane Antigens ( GMMA ) technology [14 , 42–43] . The second NTS candidate , developed by the Center for Vaccine Development of the University of Maryland School of Medicine ( CVD ) and its industrial partner , Bharat Biotech , International ( BBI ) of Hyderabad , India , contains S . Enteritidis and S . Typhimurium conjugate vaccines consisting of the core plus O polysaccharide of those serovars covalently linked to Phase 1 flagellin subunits of the homologous serovar [13–16 , 44] . As each vaccine moves towards clinical trials , a Target Product Profile ( TPP ) must be created that by necessity incorporates multiple assumptions and predictions that guide the development of the project for multiple years before clinical data become available to corroborate or refute the TPP assumptions . The TPP , which must be crafted early in the development of the candidate vaccine , provides a roadmap as it defines the type of vaccine , the route of administration , the target populations and sub-populations to be vaccinated , the number of doses to be administered and the intended immunization schedule for the target populations . The TPP also proposes limits for the expected reactogenicity ( local and systemic ) , the level of efficacy to be achieved , the duration of protection , when a booster dose might be needed , the storage conditions , the vaccine formulation ( s ) , the presentation of the vaccine , whether an adjuvant will be included and what preservative will be present in multi-dose vials . The design of the preclinical toxicology test , the formulations of vaccine to be tested , the design of the Phase 1 and 2 clinical trials , and of the ultimate pivotal Phase 3 efficacy trial all follow guidance provided by the TPP . The mathematical model described herein includes assumptions contained within one TPP . Even at early stages in development of the candidate vaccines to prevent iNTS disease , a mathematical model of what the vaccine might achieve at the future public health level becomes a useful , hopefully predictive , tool . Modifying the parameters of the model offers insights on what the vaccine can achieve . Our model has been used to assess the impact of introducing a bivalent NTS vaccine on decreasing the number of hospitalized iNTS cases and fatalities caused by the two most prevalent iNTS serovars currently found in the Malian pediatric population , S . Enteritidis and S . Typhimurium . However , other serovars have been identified among a minority of hospitalized cases of iNTS that theoretically could also be prevented . Indeed the bivalent vaccines currently in development offer the prospect of cross protection against other serovars . The bivalent conjugate vaccine described by Simon et al . [16] , for example , may offer such cross protection by targeting shared O-polysaccharides . If the effectiveness of the vaccine is reliant on these targeted polysaccharides , which are shared among all serovars within the same serogroup , this vaccine would offer cross protection against all Group B and Group D serovars , including the serovars with the next highest prevalence among hospitalized cases in Bamako , Mali ( S . Stanleyville and S . Dublin , respectively ) [13 , 15–16] . Our findings have some limitations because of the lack of data on segments of the pediatric population where iNTS disease may be occurring but not detected with our surveillance . By focusing only on iNTS cases admitted to hospital , we did not model the overall burden of NTS in the Bamako pediatric population that would include children in the community with iNTS infections who were not ill enough for their caretakers to seek health care or who were brought to traditional healers . It also did not include children with severe iNTS disease who may not have had easy access to the hospital and thus may have died at home . Moreover , blood cultures and cultures of ordinarily sterile body fluids are not 100% sensitive in detecting invasive bacterial infections , particularly if antibiotics were administered prior to reaching the hospital . Thus , some iNTS cases may have been missed by our surveillance techniques , leading to an underestimation of the number of cases and the burden of NTS in the study population . However , cases hospitalized at HGT likely capture a substantial proportion of the more severe clinical forms of iNTS disease which likely have a higher case fatality than milder forms of iNTS disease . The impact of a NTS vaccine in preventing culture-negative severe iNTS cases could be estimated by noting the difference between the decrease in hospitalized iNTS cases following vaccine implementation and a decrease in all-cause hospitalizations [10] . Despite some knowledge gaps about the epidemiologic behavior of iNTS disease in the community , our model provides an informative view that should adequately assess the impact of introducing an effective vaccine . The model presented herein is one of the first attempts to capture mathematically the epidemiologic dynamics of endemic pediatric iNTS disease in Africa and to predict the effects of future implementation of NTS vaccines currently in development on the disease burden . Several key features of the epidemiology of iNTS disease in Bamako , such as the force of infection , reservoirs of infection , and modes of transmission remain unknown and the specific effects of other factors known to modulate host risk and clinical disease severity such as HIV infection , malaria , and malnutrition [6 , 7 , 45] have not been formally measured . By relying on population data to inform the age-specific incidence of severe ( hospitalized ) iNTS disease , we indirectly captured some effects that subclinical infection and chronic carrier states may have without directly including them in the model , since such states have not yet been described for NTS cases [6] . Research efforts are underway in other venues to investigate risk factors , transmission modes , immunology , and natural reservoirs of iNTS . Mali has a lauded EPI and thus immunization coverage in other venues may not be as high . However , the model can be used to predict outcomes with other location-specific estimates of NTS coverage . Our Markov Chain approach was well suited for incorporating the surveillance data representing the currently available numbers of hospitalized iNTS cases , but made the conservative assumption that infection pressure remained constant both in the natural history of the disease and in the face of vaccination . In future iterations we plan to amend and refine our model using data from other field investigations and use ordinary differential equation ( ODE ) and partial differential equation ( PDE ) models that include older children and adults to capture transmission dynamics and emergent properties of vaccination , such as herd immunity . As more literature is published on the epidemiology of iNTS disease , a more sophisticated computational model can be crafted that incorporates new data , adding further emphasis to the importance of implementing such a vaccine to protect young children against this invasive disease .
|
A surveillance program at Gabriel Touré Hospital in Mali observed a high burden of invasive disease caused by non-typhoidal Salmonella ( iNTS ) . This surveillance program was originally instituted to measure the amount of invasive disease ( e . g . , septicemia , meningitis ) caused by two bacteria that invade the respiratory tract: Haemophilus influenzae type b ( Hib ) and Streptococcus pneumoniae ( pneumococcus ) . While documenting the burden of these pathogens , the surveillance program also found that serotypes of iNTS , mainly Salmonella Typhimurium and Salmonella Enteritidis , were common causes of severe invasive disease . As the number of cases of Hib and pneumococcus markedly decreased following the introduction of relevant vaccines , the relative threat of iNTS increased . Little is known about the reservoir of iNTS , whether it resides in humans , animals , or the environment , or how it is spread to susceptible children . Without this knowledge , it is not possible to employ certain disease control methods useful in interrupting the transmission of other pathogens . Therefore , vaccination remains the one promising control strategy for this disease . Our research modeled the potential effects of introducing an iNTS vaccine . The findings are of great importance to Mali and other developing countries where young children are at a high risk of developing iNTS disease .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"mali",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"immunology",
"geographical",
"locations",
"microbiology",
"pediatrics",
"vaccines",
"preventive",
"medicine",
"bacterial",
"diseases",
"enterobacteriaceae",
"vaccination",
"and",
"immunization",
"bacteria",
"bacterial",
"pathogens",
"africa",
"salmonella",
"typhimurium",
"public",
"and",
"occupational",
"health",
"infectious",
"diseases",
"booster",
"doses",
"medical",
"microbiology",
"microbial",
"pathogens",
"salmonella",
"people",
"and",
"places",
"biology",
"and",
"life",
"sciences",
"conjugate",
"vaccines",
"vaccine",
"development",
"organisms"
] |
2017
|
Modeling the Potential for Vaccination to Diminish the Burden of Invasive Non-typhoidal Salmonella Disease in Young Children in Mali, West Africa
|
Variation in the branching of plant inflorescences determines flower number and , consequently , reproductive success and crop yield . Nightshade ( Solanaceae ) species are models for a widespread , yet poorly understood , program of eudicot growth , where short side branches are initiated upon floral termination . This “sympodial” program produces the few-flowered tomato inflorescence , but the classical mutants compound inflorescence ( s ) and anantha ( an ) are highly branched , and s bears hundreds of flowers . Here we show that S and AN , which encode a homeobox transcription factor and an F-box protein , respectively , control inflorescence architecture by promoting successive stages in the progression of an inflorescence meristem to floral specification . S and AN are sequentially expressed during this gradual phase transition , and the loss of either gene delays flower formation , resulting in additional branching . Independently arisen alleles of s account for inflorescence variation among domesticated tomatoes , and an stimulates branching in pepper plants that normally have solitary flowers . Our results suggest that variation of Solanaceae inflorescences is modulated through temporal changes in the acquisition of floral fate , providing a flexible evolutionary mechanism to elaborate sympodial inflorescence shoots .
A striking manifestation of plant evolution is observed in the diverse branching and patterning of inflorescences , which are the shoots that bear flowers [1 , 2] Inflorescences are derived from the growth of dome-shaped groups of pluripotent cells called apical meristems . Apical meristems first produce leaves , and upon flowering induction , they produce inflorescence meristems that transition to floral meristems , which produce flowers . Extensive variation in inflorescence complexity is found in the nightshade ( Solanaceae ) family , where flowering marks the end of main shoot growth , and vegetative aerial growth is renewed from axillary meristems in a perennial growth system known as “sympodial” [3–5] . The simplest Solanaceae inflorescence is a solitary flower , represented by pepper ( Capsicum annum ) in Figure 1A . Tomato ( Solanum lycopersicum ) , on the other hand , generates a few-flowered inflorescence organized in a zigzag branch ( Figure 1B ) , but there are three classical mutants called compound inflorescence ( s ) ( Figure 1D and 1E ) , anantha ( an ) ( Figure 1F and 1G ) , and falsiflora ( fa ) ( Figure 1H ) that bear highly branched inflorescences resembling wild Solanaceae species like S . crispum ( Figure 1C ) [6–8] These similarities suggest that branching complexity may arise from tuning a common underlying developmental program . We set out to begin to unravel the basis of Solanaceae inflorescence diversity using these mutants whose variation ranges from branched inflorescences that produce hundreds of fertile flowers as seen in s [6] , to the branching shoots of an that terminate in cauliflower-like tissue [7] , to the leafy inflorescences of fa , which is defective in the tomato ortholog of LEAFY ( LFY ) [9] .
The tomato plant is a compound shoot formed from reiterated sympodial shoot units ( SYM ) that arise from vegetative meristems that produce three leaves before terminating with an inflorescence [10] . The tomato inflorescence is also a compound shoot , which is condensed , consisting of sequential one-nodal inflorescence sympodial units ( ISUs ) each terminated by a single flower [11] . During early inflorescence development , individual ISUs developed in a progression of two phases . In the first phase , a sympodial inflorescence meristem ( SIM ) , which was distinct from a SYM because it formed within the inflorescence itself , arose and produced a new SIM on its side before differentiating into a floral meristem ( FM ) in a second phase . These events created the first ISU and the SIM of the second ISU ( Figure 2 and Figure S1 ) . This pattern reiterated as subsequent SIMs developed perpendicular to one another , producing a zigzag pattern of flower initiation ( Figure 2A ) . In an and fa mutants , the primary meristems failed to become flowers , remained indeterminate , and repeatedly initiated secondary SIMs that , themselves , repeatedly produced SIMs ( Figure 2 and Figure S1 ) . s was more asynchronous , as SIMs eventually transitioned to flowers after producing 2–4 axillary SIMs in a variable , environment-dependent manner ( Figure 2 and Figure S1 ) . Although the branching effects were similar between the three mutants , floral phenotypes were not . Mutants of fa were primarily vegetative , producing numerous leaves that developed early as primordia coming off the flanks of meristems ( Figure 1H and Figure S1K ) . Mutants of an , on the other hand , produced leaf primordia mixed with other tissue that at maturity resembled modified sepals or bracts ( Figure 1F and 1G ) . It is interesting that s mutants maintained the capacity to produce normal flowers , indicating a reduced role in the flower relative to the SIM , although on occasion we observed some leaf-like primordia ( Figure S1F ) . Thus , beyond distinctions in controlling floral organ identity ( Figure 1 ) , s , an , and fa mutants exhibit delayed ISU maturation , resulting in additional branches through the ongoing initiation of lateral SIMs . Notably , SIM branching in diverse Solanaceae is based on an s-like program , as seen in early inflorescence development of S . crispum ( Figure 2D and Figure S2 ) . This suggests that delays in floral termination ( perhaps mediated by S , or the genetic pathway that S defines ) provide a developmental framework for the modulation of sympodial branching in the Solanaceae . To identify the genes responsible for these phenotypes , s and an were localized to linked regions of chromosome 2 , and s was positionally cloned using a remarkable level of multi-genome synteny between the eudicot species poplar ( Populus trichocarpa ) , Barrel Medic ( Medicago truncatula ) , and grape ( Vitis vinifera ) ( Figure 3 ) . Several genes were shared in a short chromosomal segment ranging from 105–140 kb , and aligning these regions revealed three transcription factors: two AP2-like genes and a WUSCHEL-homeobox ( WOX ) that each co-segregated with s . Sequencing of all three genes revealed independent point mutations in the WOX gene from two alleles of s ( s-classic and s-multiflora ) , and Southern blot analysis showed chromosomal changes in an additional allele ( s-n5568 ) , demonstrating that s is mutated in this gene ( Figure 3 and Figure S3B and S4 ) ( GenBank ( http://www . ncbi . nlm . nih . gov/Genbank/ ) accessions FJ190663 and FJ190664 ) . To our knowledge , this is the first example of gene identification using multi-genome synteny among four eudicot species . Our data suggest that an even greater level of synteny remains to be discovered , and that non-model species will realize similar benefits as more genomes are sequenced . WOX proteins share homology with the meristem maintenance gene WUSCHEL and are plant-specific transcription factors [12] . Among 14 WOX genes in Arabidopsis , S is most similar to WOX9/STIMPY ( STIP ) that functions with WOX8/STIPL to regulate embryonic patterning [13] . In contrast , we have found that S is a major determinant of inflorescence architecture in tomato . In the context of a European Solanaceae project ( Eu-Sol ) , we established and phenotyped a collection of more than 6000 domesticated tomato varieties for various traits ( Materials and Methods; https://www . eu-sol . wur . nl ) . The power of such a large germplasm resource resides in the fact that extensive natural allelic variation with both qualitative and quantitative effects has been selected and maintained since tomato was first domesticated [14] . Thus , these varieties provide a complement to the stronger , mostly deleterious effects , of alleles derived from artificial mutagenesis [8] . Among the 6 , 000 tomato lines , we identified 23 accessions with highly branched inflorescences and all were allelic to s Surprisingly , 22 of these lines carried the s-classic allele of the original mutant described 100 years ago , indicating that early breeders were positively selecting this mutant , probably for aesthetic value and fruit production ( Figure 4 ) [6] . However , it was also possible that one or more of these lines arose independently , generating a mutation in the same nucleotide as s-classic . To address this question , we sequenced the coding region of all 22 lines and un-branched controls and found that all were identical except CC5721 . Interestingly , this line carried four single-nucleotide polymorphisms ( SNPs ) that were shared with at least one un-branched variety , indicating that the s mutation in CC5721 may have arisen independently from a genetically distinct progenitor line ( GenBank accessions FJ190665 , FJ190666 , and FJ190667 ) . Two pieces of evidence lend support to this claim . Firstly , the four SNPs were distributed close ( all within 1 , 000 bp ) to the s lesion . Secondly , we sequenced a short segment of DNA from the tightly linked bacterial artificial chromosome ( BAC ) 298N3 and found that CC5721 had six SNPs and a 9–bp insertion-deletion ( indel ) that distinguished it from the other 21 domesticated types carrying s-classic ( Figure 3b ) ( GenBank accessions FJ215691 and FJ215692 ) . Still , in the absence of a geographic distribution of haplotypes , we cannot exclude a remote possibility that CC5721 arose as a result of an intra-genic recombination between s-classic and an unbranched variety . Regardless , at least three independently arisen alleles of s ( s-classic , s-multiflora , and Rose Quartz Multiflora ) are responsible for a major portion of the diversity in tomato inflorescence architecture . The similarity between the phenotypically strong allele s-multiflora and strong an mutants suggested a functional link in regulating an underlying inflorescence branching program ( Figure 1 ) . Furthermore , we created double mutant plants of weak an alleles and s and found they were phenotypically enhanced to resemble strong an ( Figure S5 ) . Interestingly , stronger phenotypes were observed for both inflorescence branching and floral identity . Specifically , we found that the sepal and carpelloid tissue of weak an mutants became much more meristematic with less organ identity ( Figure S5B ) . In some double mutants , additional leaves formed in the inflorescence , resembling fa mutants ( unpublished data ) . This suggests that S and AN have overlapping roles in inflorescence architecture as well as floral identity . We noted that an resembled a Lotus japonicus mutant called proliferating floral organs ( pfo ) ( Figure 1e ) [15] . PFO encodes an F-box protein orthologous to Antirrhinum FIMBRIATA ( FIM ) and Arabidopsis UNUSUAL FLORAL ORGANS ( UFO ) [16] , and the tomato ortholog of this gene co-segregated with an . Six alleles had mutations in the coding region , revealing that an is mutated in the tomato ortholog of FIM/UFO ( Figures 2C , 3E , and Figure S3A and S6 ) ( GenBank accession FJ190668 ) . The similar inflorescence and floral phenotypes found in an and fa mutants [17] may , therefore , stem from conserved functional associations of their gene products as described in Arabidopsis [18] . However , the relationship between S and AN was less clear , and their expression patterns were therefore explored . S was expressed to varying degrees in all tissues except roots , whereas most AN expression was restricted to floral buds , indicating a primary function in inflorescence and flower development . FA accumulated predominantly in shoot apices ( Figure 5A ) [9] . We explored further the expression of S and AN using in situ hybridization , which revealed temporally distinct patterns during inflorescence development . S was expressed in a wedge shape radiating outward from 2–3 cells from the center of immature SIMs ( Figure 5B and 5C ) . This expression initiated shortly after lateral bulging of the SIM and was transient , because it disappeared before floral termination . AN expression initiated in incipient FMs shortly after down-regulation of S . AN expression was less intense than S , and was limited to the upper layers of the rapidly maturing SIM ( Figure 5D and 5E ) . Both genes were reactivated in flower primordia in a ring of cells that marked a boundary domain , first between sepal and petal primordia and later between petals and stamens . These floral expression patterns are consistent with the failure of an mutants to initiate normal flowers , suggest a role for S in the flower , and likely explain the enhanced developmental and molecular phenotypes that s imposes on floral organ identity in weak alleles of an ( Figure S5 ) . Indeed , double mutants show little or no an expression similar to strong an mutants alone ( Figure S5D ) . The expression pattern of S suggested that it functioned early in SIM maturation to promote the transition to FM , whereas AN operated soon after to provide early FM identity . To test these hypotheses , we examined the expression of S and AN in s , an , and fa mutants . S was expressed in all mutant backgrounds , and , as in wild-type , was detected in younger lateral SIMs of an inflorescences ( Figure 5F and 5G ) . This indicates that an meristems still reach a pre-floral SIM state . AN expression , on the other hand , was undetectable by RT-PCR ( reverse-transcriptase PCR ) in fa mutants , consistent with the proposal that FA functions upstream of AN [17] ( Figure 5F ) . Initial expression of AN in s mutants was delayed , and subsequently detected in only a small subset of SIMs compared to wild-type . In those meristems expressing AN , the signal was deeper and more intense than normal ( Figure 5H ) . In situ hybridization from older inflorescences revealed some meristems lacking S and AN activity altogether , which we verified by whole-mount in situ hybridization ( Figure 5H and unpublished data ) This indicates that different meristems are at different phases of ISU maturation , and may also reflect the frequent observation of modified leaves or bracts in older an inflorescences if some meristems retain a more vegetative state . Taken together , these expression patterns support a mechanism where S and AN promote successive stages in the progression of an inflorescence meristem to floral specification through sequential transient activities that gradually promote maturation of SIMs ( expressing S ) to early FMs ( expressing AN ) ( Figure 5I ) . Loss of either gene provides SIMs with an extended period of indeterminacy that facilitates ISU elaboration according to an underlying program of sympodial growth ( Figure 5J ) . Furthermore , the observation that S expression is maintained in an mutants and vice-versa , and that their expression is restricted to temporally distinct domains , supports the notion that these genes have separate but overlapping functions in the maturation of individual ISUs , consistent with the enhancement of weak an alleles by s ( Figure S5 ) . The expression patterns of S and AN along with their mutant phenotypes lead to a model in which temporal differences in the maturation of a SIM to an FM can regulate the duration of sympodial inflorescence branching . In other words , a slower transition enables more inflorescence branching and vice-versa . This suggests that the SIM phase and the early FM phase of a single flower can each provide a developmental window in which a compound inflorescence can form . We tested this hypothesis genetically by taking advantage of mutants of single flower truss ( the tomato ortholog of FT , which is a major component of florigen ) , whose inflorescences are indeterminate vegetative shoots with single flowers separated in space by leaves [19] ( Figure 6A ) . In sft:an double mutants , we observed that individual flowers became branched inflorescences , though less so than in an mutants alone ( 2–4 versus 20–25 branches at the same age , Figures 1F , 6B , and 6C ) . By contrast , branching in sft:s double mutants resulted in elaboration of the vegetative inflorescence , but normal flowers still formed ( Figure S7 ) . Taken together , these results support the proposal that S acts earlier within a single inflorescence meristem to regulate sympodial branching , whereas AN acts later as FM identity is reached . Our model suggests that Solanaceae inflorescences with only single flowers may result from rapid termination of the FM and hence elimination of the SIM stage , but that single flower species can still produce branched inflorescences . We addressed this by mutagenizing pepper ( C anuum ) , which identified one mutant ( called Ca-an ) that produced an indeterminate shoot instead of a flower . This structure lacked petals and stamens and branched more extensively in a mixed genetic background , resembling tomato an mutants ( Figure 6E and Figure S8 ) . We sequenced pepper AN from Ca-an and found a missense mutation from the wild-type progenitor sequence causing a nucleotide change just prior to one of our tomato an alleles ( an-e1444 ) that co-segregated with the mutant phenotype ( Figure S6 ) , indicating that Ca-an is mutated in the pepper ortholog of FIM/UFO ( GenBank accession FJ190669 ) . Like tomato , Ca-AN was expressed in a ring of cells flanking developing petals and stamens ( Figure S8 ) . Interestingly , Ca-AN could not be detected in an earlier inflorescence meristem , which lends support to the idea that pepper has a short SIM phase and progresses rapidly to floral termination . Yet , Ca-an mutants revealed a latent potential to branch , indicating that Solanaceae AN shares a conserved role in promoting FM determinacy with its orthologs in other species [15 , 16 , 20 , 21] . Of all other known UFO mutants , the pfo mutant from L . japonicus is most similar to Ca-an , with a compact branched structure described as a reiteration of sepals and FMs . Normal L . japonicus produces pairs of flowers in the axils of leaves , and so loss of UFO function provides an extended period of indeterminacy to each pair of inflorescence meristems . By contrast , the stp mutant of pea generates similar organ defects but produces secondary FMs within the primary flower . Thus , UFO has a highly conserved role in floral identity , but its control of inflorescence branching is more species-specific and likely reflects differences in mechanisms of inflorescence meristem initiation . Notably , branching of tomato an mutants was more extreme than in Ca-an mutants ( Figures 1 and 6 ) . This indicates that underlying the tomato SIM phase is a program promoting branching and that the foundation for more complex branching is an inflorescence composed of reiterated SIMs . These data suggest that highly branched species like S . crispum evolved from an ancestral form that resembled tomato , as opposed to pepper ( Figure S2 ) .
Our results reveal a genetic foundation for the Solanaceae inflorescence and provide evidence for a possible mechanism that modulates simple and complex inflorescence structures known as “cymes” [1 , 2] While the generation of a cymose inflorescence through sympodial growth is likely a complex process involving many unknown genetic and environmental factors , we provide a major advance in understanding how cymes may be modified into more complex structures based on elaboration of the ubiquitous ISU shoot system ( Figure 2 ) [5] . This mechanism uses conserved machinery ( AN/UFO and FA/LFY ) that regulates inflorescence and flower development in other species [15 , 16 , 20–26] . Interestingly , the effects of UFO on inflorescence architecture vary considerably , ranging from infrequent replacement of single flowers with secondary inflorescence shoots in Arabidopsis ufo mutants [27] , to the production of ectopic flowers in the inflorescences of pea stp mutants [20] , to the large mass of inflorescence/floral tissue in pfo mutants of L japonicus [15] , and as shown here , the an mutant of tomato and pepper . Furthermore , we describe S/WOX9 as a novel component in the control of inflorescence architecture—a role that was not detected for its Arabidopsis ortholog . We also find that the tomato ortholog of TERMINAL FLOWER1 ( TFL1 ) called SELF PRUNING ( SP ) , which has a major effect on Arabidopsis inflorescences [2] , is neutral on sympodial inflorescence branching in normal tomato inflorescences ( unpublished data ) , and exhibits indirect effects on s inflorescence branching ( Table S1 ) . These differences may originate from the evolution of distinct growth habits . Branching complexity in sympodial species relies on termination of inflorescence meristems through the transition of a SIM to an FM . We suggest that a transient expression of S followed by AN was co-opted in Solanaceae sympodial development to boost two phases of sympodial meristem growth in this specialized shoot , both of which can potentiate branching ( Figure 3I ) . In monopodial dicot species such as Arabidopsis or Antirrhinum , the inflorescence meristem produces no comparable SIMs , being indeterminate and generating lateral single flowers . This indeterminacy may explain why WOX9 , by itself , is dispensable for inflorescence development [28 , 29] . Indeed , inflorescence ramification in monopodial dicot plants is more often stimulated through identity change [16 , 30 , 31] , which could also explain some of the branching effects observed in Ca-an ( Figure 6 ) Thus , while the evolutionary diversification of plant inflorescence architecture is united under a common developmental theme [2] , plants with different growth habits use related as well as distinct developmental modules to regulate branching [32] . We propose that Solanaceae inflorescence variation is based on controlling sympodial branching through temporal changes in the acquisition of floral fate , which is most flexible within the SIM phase . Short delays in the activation of genes like S ( or as-yet-undiscovered other genes in the S pathway ) followed by an abrupt switch to floral termination may explain the evolution and quantitative variation of compound inflorescences in the genus Solanum ( Figure S9 ) , as well as in other sympodial species , like trees [5] . Such a mechanism would provide a flexible way to guarantee the production and simultaneous maturation of large numbers of flowers , thereby ensuring a crucial aspect of reproductive success and perhaps providing a new tool for the manipulation of crop yields .
Classic alleles of s ( s-classic LA3094 ) , an ( LA0536 ) , and fa ( LA0854 ) , and those of representative wild tomato species were gifts from the C . M . Rick Center ( Davis , California; http://tgrc . ucdavis . edu ) . An additional allele of s ( LA0560; s-multiflora , C . M Rick Center ) was verified by complementation test . A third s allele and six additional an alleles were identified as inflorescence mutants in a screen of a tomato mutant library [8] . Wild tomato species were gifts from the C . M . Rick Center . Wild tomato species , such as S . lycopersicoides , can be difficult to grow and maintain until flowering and only two representative plants were available for phenotypic analyses , but inflorescence complexity within each plant was uniform throughout . More distantly related Solanum species were gifts from the Botanical and Experimental Garden at Nijmegen , The Netherlands . Up to three representative plants were used for phenotypic analyses . The ∼6 , 000 domesticated tomato varieties were collected from various public and private germplasm sources . All plants were grown in greenhouses under natural light or in agricultural field conditions in Israel using standard irrigation and fertilization regimes . The s mutant was originally mapped on the long arm of chromosome 2 , and verified using 20 mutants selected from an F2 population derived from a cross with the wild species S . pimpinellifolium ( LA1589 ) . This positioned s in the region overlapping introgression lines IL2–3/2–4 on the tomato introgression line map [33] . A larger mapping population was generated by crossing s-n5568 with the wild tomato species S . pennellii ( LA0716 ) . F1 hybrid plants were self-fertilized to produce a mapping population of 5 , 000 F2 plants . Five hundred individual s mutant plants were scored with CAPS-PCR markers from the most current tomato genetic map ( Solanaceae Genomics Network at http://www . sgn . cornell . edu ) , focusing on the region of IL2–3/2–4 . Additional markers surrounding the tightly linked CNR locus were provided by K . Manning [34] . Marker density was improved using conserved synteny identified between seven markers in a 15-cM window in tomato and a 500-kb segment of Arabidopsis chromosome 1 ( marker information available upon request ) . Two co-segregating markers ( 0 recombinant chromosomes out of 1 , 000 gametes ) were used to isolate a BAC from a S . lycopersicum HindIII library kindly provided by J . J . Giovannoni and J . VanEck at Cornell University ( Ithaca , New York ) . DNA fragments from three independent restriction enzyme digestions of a single BAC clone were sub-cloned into TOPO TA cloning vectors ( Invitrogen ) for shotgun sequencing . Fragments containing genes were annotated using BLASTX against the Arabidopsis protein database and used to search other genomes for additional synteny . Sequences from four tightly linked markers ( Figure 3 ) were used in a BLASTN or TBLASTX search against genomes of P trichocarpa , M truncatula , and V vinifera . Genes in syntenic regions ranging from 110–140 kb were aligned manually and searched for candidate genes , which identified the Apetala-2 ( AP2 ) and Wuschel-homoebox ( WOX ) transcription factors . A tomato-specific WOX marker was produced by degenerate PCR based on conserved regions in the WOX from these three species and an EST from Petunia hybrida ( accession number EB174485 ) . Transcript ends were determined by rapid amplification of cDNA ends ( RACE ) ( Sambrook ) using total RNA isolated from young inflorescences with TriReagent ( Sigma-Genosys ) . DNA from s-like varieties with compound inflorescences from the Core Collection was PCR amplified with gene-specific primers and used in a CAPS-PCR assay diagnostic of s-classic . The expressivity of the s phenotype is affected by genetic background , which became evident when phenotyping 22 domesticated varieties each carrying the s-classic allele , but varying in many phenotypic characters , including branching . Modifiers are responsible for these differences , which may or may not have a functional relationship to S . Furthermore , it is well-documented that sympodial shoot growth in tomato is highly sensitive to light intensity , which could also contribute to quantitative variation between accessions . On occasion , modestly branched accessions were observed that produced only 2–4 additional branches compared to normal , which , if not allelic to s , could potentially modify ( enhance ) the s phenotype . Yet , the majority of extreme branching variation was due solely to changes in S function , indicated by normal segregation of families segregating for each s allele in a common genetic background ( cv . M82 ) . Thus , differences in phenotypic strength , as seen in s-multiflora , result from modifier loci , but these are much weaker in their effects compared to s mutations . The an mutant was originally mapped to the long arm of chromosome 2 , and subsequently positioned in the region overlapping IL2–3/2–4/2–5 . The phenotype of the pfo mutant in L japonicus resembled weak alleles of an and led us to search for the tomato ortholog of FIM/UFO . A single EST ( SGN-U341425 ) with homology to FIM/UFO was used to generate a CAPS-PCR marker that mapped to the same region as an ( http://www . sgn . cornell . edu ) . DNA from six EMS alleles was amplified using gene-specific primers and sequenced directly , which identified five independent mutations . The central portion of coding sequence of the an-classic allele could not be amplified by PCR , suggesting a structural change or large insertion ( unpublished data ) . This rearrangement in the an-classic allele was verified using DNA Southern blot hybridizations ( Figure S3 ) according to established protocols . An EMS mutagenesis of the pepper variety Maor was performed according to a protocol for tomato seeds [8] . Among 1 , 500 M2 families of pepper , one inflorescence mutant was identified based on phenotypic similarity to weak alleles of tomato an This new mutant was first mapped by restriction fragment length polymorphism ( RFLP ) analysis to a region of chromosome 2 in pepper that is syntenic with tomato chromosome 2 where anantha was positioned previously . DNA from the mutant was isolated and sequenced using primers designed from the tomato gene . Co-segregation of the mutation with the pepper an phenotype was verified in an F2 population of approximately 100 plants , and the mutation was found to be derived from the Maor variety progenitor sequence . Our allele changes a nearly invariant glycine among F-box proteins into a charged amino acid , glutamic acid ( see Figure S6 ) . This glycine is the second amino acid in a short stretch of ∼10 highly conserved amino acids in UFO orthologs for which at least one mutant allele is available in Arabidopsis , pea , tomato , and now pepper . Thus , this region is a hot spot for mutations that give very similar floral phenotypes in multiple species . Developmental and morphological analyses on single and double mutants were performed on alleles originating from the tomato cultivar M82 . M82 lines were either mutant or wild type for the gene SELF PRUNING ( SP ) , which had only modest effects on inflorescence phenotypes in s or an that could be attributed to changes in the length of sympodial units—a phenotype regulated by SP . Single and double mutants of sft were the allele sft-7187 [19] . Mutants of fa were in the background of Rheinlands Ruhm . Branching events were counted on two independent inflorescences from Core Collection varieties containing s-classic and non-s controls . For SEM , immature inflorescences from sympodial shoots were dissected and processed through an EtOH series , critical-point dried , and coated with gold particles for microscope analysis on a Philips XL30 ESEM FEG . RT-PCR was performed using a One-Step RT-PCR kit ( Qiagen ) on total RNA isolated by TriReagent ( Sigma-Genosys ) according to the manufacturers' protocols . Primer sequences are available upon request . Tissues for in situ hybridization were dissected and fixed according to standard protocols [35] . In vitro transcribed RNA probes were generated from 5′ partial ( S ) or full-length ( AN ) cDNA clones and transcripts were detected using standard in situ hybridization techniques . Whole-mount in situ hybridization was performed as described [36] , using the same probes .
|
Among the most distinguishing features of plants are the flower-bearing shoots , called inflorescences . Despite a solid understanding of flower development , the molecular mechanisms that control inflorescence architecture remain obscure . We have explored this question in tomato , where mutations in two genes , ANANTHA ( AN ) and COMPOUND INFLORESCENCE ( S ) , transform the well-known tomato “vine” into a highly branched structure with hundreds of flowers . We find that AN encodes an F-box protein ortholog of a gene called UNUSUAL FLORAL ORGANS that controls the identity of floral organs ( petals , sepals , and so on ) , whereas S encodes a transcription factor related to a gene called WUSCHEL HOMEOBOX 9 that is involved in patterning the embryo within the plant seed . ( F-box proteins are known for marking other proteins for degradation , but they can also function in hormone regulation and transcriptional activation ) Interestingly , these genes have little or no effect on branching in inflorescences that grow continuously ( so-called “indeterminate” shoots ) , as in Arabidopsis . However , we find that transient sequential expression of S followed by AN promotes branch termination and flower formation in plants where meristem growth ends with inflorescence and flower production ( “determinate” shoots ) . We show that mutant alleles of s dramatically increase branch and flower number and have probably been selected for by breeders during modern cultivation . Moreover , the single-flower inflorescence of pepper ( a species related to tomato , within the same Solanaceae family ) can be converted to a compound inflorescence upon mutating its AN ortholog . Our results suggest a new developmental mechanism whereby inflorescence elaboration can be controlled through temporal regulation of floral fate .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"developmental",
"biology",
"plant",
"biology"
] |
2008
|
The Making of a Compound Inflorescence in Tomato and Related Nightshades
|
The differentiation of discrete and continuous movement is one of the pillars of motor behavior classification . Discrete movements have a definite beginning and end , whereas continuous movements do not have such discriminable end points . In the past decade there has been vigorous debate whether this classification implies different control processes . This debate up until the present has been empirically based . Here , we present an unambiguous non-empirical classification based on theorems in dynamical system theory that sets discrete and continuous movements apart . Through computational simulations of representative modes of each class and topological analysis of the flow in state space , we show that distinct control mechanisms underwrite discrete and fast rhythmic movements . In particular , we demonstrate that discrete movements require a time keeper while fast rhythmic movements do not . We validate our computational findings experimentally using a behavioral paradigm in which human participants performed finger flexion-extension movements at various movement paces and under different instructions . Our results demonstrate that the human motor system employs different timing control mechanisms ( presumably via differential recruitment of neural subsystems ) to accomplish varying behavioral functions such as speed constraints .
Discrete movements constitute singularly occurring events preceded and followed by a period without motion ( i . e . , with zero velocity ) for a reasonable amount of time , such as a single finger flexion or flexion-extension cycle [1] , [2] . Continuous movements lack such recognizable endpoints , and normally are considered rhythmic if they constitute repetitions of particular events , in which case they often look quite sinusoidal . While it is trivial that discrete movements can be repeated periodically , the question whether motor behavior is fundamentally discrete or rhythmic is not . Is motor behavior fundamentally discrete , reducing rhythmic movement to mere concatenations of discrete movements [3] , [4] ? Or is motor control fundamentally rhythmic , in which case discrete movements are merely ‘aborted’ cycles of rhythmic movements [5]–[7] ? Alternatively , both types of movements may belong to distinct classes that are irreducible to each other [8]–[10] , hence implying the utilization of different movement generating mechanisms . Proponents of the ‘discrete perspective’ have sought evidence for discrete movement control through the identification of movement segments in movement trajectories . However , segmented motion need not imply segmented control [11] . In fact , the possibility to settle the dispute ( solely ) on the basis of kinematic features of movement ( movement time , peak velocity , symmetry of velocity profiles , etc . ) has recently been questioned [12] . Other researchers have aimed to identify the neural structures associated with discrete and rhythmic movements . For instance , Schaal and colleagues [9] showed that the brain areas that were associated with rhythmic movements were approximately a subset of those that were active during discrete movement execution . Differential involvement of neural subsystems does not provide a classification principle , however . Unambiguous classification requires the identification of invariance that is unique to each class so that the intersection of these two sets of characteristics is empty . Such a result will provide unambiguous evidence that two classes indeed are distinct . Dynamic systems theory offers such a classification principle based on phase flow topologies , which identify all behavioral possibilities within a class . Its significance lies in the fact that the classification is model-independent; every behavior within a class can be mapped upon others , whereas maps between classes do not exist . We use this principled approach to address the controversy whether discrete and rhythmic movements are fundamentally different . To that aim , we introduce the notion of phase flow topologies , identify the invariance separating two movement classes , and present an experimental study testifying to the existence of ( at least ) two different movement classes . Deterministic , time-continuous and autonomous systems can be unambiguously described through their flow in state ( or phase ) space , defined as the space spanned by the system's position x and velocity ( under the commonly adopted assumption that the deterministic component of movement trajectories can be fully described by two state variables ) . Whereas the phase flow quantitatively describes the system's evolution as a function of its current state ( x , ) ; the system's qualitative behavior is solely determined by its phase flow topology . From the Poincaré-Bendixson theorem [13] , [14] it follows that the only possible topologies in two dimensional systems are composed of elements referred to as fixed points , limit cycles , and separatrices . A fixed point of the system identifies a rest state ( i . e . , rate of change is zero , ) , and , if stable , all trajectories in phase space eventually converge to it ( Figure 1A ) . A system located at a fixed point can only depart from it in the presence of an external stimulation . A separatrix is a subset of points in the phase space that divides locally distinct phase flows ( Figure 1A and 1B ) . In most cases for two-dimensional phase spaces , a separatrix is a line from which the flow points away in approximately opposite directions . Even simpler , for one-dimensional phase spaces any unstable fixed point is a separatrix . Limit cycles ( Figure 1C ) are closed loops in a two-dimensional phase space . If a limit cycle is stable , then all trajectories converge to it . A system on a limit cycle will repetitively traverse the same trajectory in phase space and sustain a periodic motion . Since these elements , fixed points and limit cycles , compose all phase flows in two dimensions , we associate discrete and rhythmic movements with these . The Hartman-Grobman theorem [13] , [14] states that the flow in the local neighborhood of a fixed point is topologically equivalent to that of its linearization , which implies that a continuous invertible mapping ( a homeomorphism ) between both local phase spaces exists . From these theorems it follows that dynamical systems belong to the same class if , and only if , they are topologically equivalent . Therefore , movements that can be shown to be governed by fixed point dynamics versus movements governed by limit cycle dynamics are not reducible to each other , and as such we can make the strong claim that they are from different equivalence classes . In consideration of the notion of topological equivalence , Jirsa and Kelso [15] recently formulated a generic model construct that allows for a stable fixed point and a separatrix ( referred to as the mono-stable regime ) or a stable limit cycle regime ( Figure 1 ) in its corresponding phase space ( see Text S1 ) . These topologies correspond to single ( i . e . , discrete ) flexion-extension movements and rhythmic movement , respectively . This perspective has three crucial features . First , the qualitative behavior in each regime is model independent . Second , each single movement execution in the mono-stable regime depends on an external triggering ( mathematically speaking , the system is non-autonomous ) . In contrast , in the ( autonomous ) limit cycle regime no external stimulation is required and movement is self-sustaining . Third , the phase flow underlying movement is invariant on the time scale of the movement in both cases . Here , we examine this perspective by directly investigating numerically generated phase flows as well as those generated by humans and show that discrete and continuous movements belong to distinct dynamical classes .
We computationally examined the generic model under a large parameter and frequency range in order to examine the robustness and limits of its behavior in both dynamical regimes ( see Materials and Methods ) . In the limit cycle regime , the timing requirement ( i . e . , the computationally implemented movement frequency ) was met under all movement paces ( i . e . , frequencies ) . In contrast , in the mono-stable regime the actual timing deviated from the required timing due to a period-doubling when the movement pace exceeded approximately 2 . 0 Hz . ( Figure 2A ) , which occurs due to the arrival of stimulus n before movement n−1 has finished . These observations were robust under all parameter settings within each dynamical regime ( see Text S1 and Figures S1 , S2 , and S3 ) , although the frequency at which the period doubling occurred showed a small variation as a function of one of the model parameters . In fact , while the exact frequency at which stimulus – movement interference occurs will show little variation as a function of the specific model realization ( i . e . , through function g1 and g2; see Equation 1 in Text S1 ) , its occurrence with increasing frequency of stimulation is unavoidable . By implication , every discrete movement system has an upper ( frequency ) limit in generating sequential movements . In the behavioral experiment human participants ( n = 8 ) performed an auditory-paced unimanual finger flexion-extension timing task under similar movement paces ( from 0 . 5 Hz to 3 . 5 Hz; step size 0 . 5 Hz ) that were presented in ascending or descending order ( see Materials and Methods ) . The participants were instructed to synchronize their full flexion with the metronome under three instruction conditions: to move as fast as possible ( with staccato like movements being initiated to end/start a cycle ) , as smooth as possible ( move so that the finger is continuously moving during the movement period interval ) or without any specific instruction . We refer to these conditions as ‘discrete’ , ‘smooth’ , and ‘natural’ , respectively ( Figure 2B ) . Please note that , notwithstanding the repetitiveness of the movements , these instructions may elicit movements generated by distinct control mechanisms but do not prescribe the latter . We reconstruct the vector fields underlying the phase flow ( see Figure 3 and Materials and Methods ) using a novel technique [16] , [17] that has been successfully tested on simulated data from dynamical systems [18] , [19] and applied in fields like ( among others ) physics [16] , [17] , engineering [20] , economics [21] , and which was recently introduced in the study of human movement [19] , [22] , [23] . In addition , we investigate the phase spaces in terms of two-dimensional probability distributions and performed more ‘traditional’ kinematic analysis commonly utilized in the ( human ) movement sciences ( see Text S1 and Figures S4 , S5 , S6 , S7 , S8 , and S9 ) . Figure 3 represents the vector fields ( Figure 3A , 3B , 3D , 3E ) from five trials of a single participant and the corresponding angle diagrams ( Figure 3C and 3F , respectively ) , and clearly indicates the existence of a fixed point ( Figure 3A–3C ) and a limit cycle ( Figure 3D–3E ) . Figure 4A–4C ( upper row for each subfigure ) shows the angle diagrams averaged across all participants for each frequency and instruction condition . Obviously , the averaging across participants , to some extent , smears out the representation of the topological structures , as indicated by the standard deviations across participants of the angle reconstructions in the lower rows of Figure 4A to 4C . Regardless , the existence of a single fixed point at slow movement paces in the discrete condition , indicating the utilization of the mono-stable regime dynamics , can be appreciated from Figure 4A ( upper row ) . In the natural and smooth condition the vector fields are less structured at slow paces , especially at 0 . 5 Hz ( Figure 4A–4C ) . Scattered ( to some degree ) vector fields and the existence of either one or two fixed points appear at 0 . 5 Hz in the smooth and the natural condition . The fixed point ( s ) appears clearer at 1 . 0 Hz to 2 . 0 Hz in both conditions . Under all instruction conditions , however , the fixed point ( s ) vanishes at high movement paces and invariantly gives way to limit cycle dynamics ( Figure 4A–4C ) . These results indicate that humans utilize distinct timing mechanisms in a movement pace-dependent manner .
What are the implications of these finding ? First and foremost , our results lay the foundation of a motor behavior classification scheme based on mathematical theorems . We demonstrated that discrete and fast rhythmic movements constitute distinct classes; their genesis is , by implication , underwritten by different mechanisms . Fast rhythmic movements are autonomous and their timing emerges from the movement dynamics . In contrast , discrete movements are non-autonomous: Their timed execution cannot originate from their dynamics and hence requires external time keeping , most likely arising from a neural structure or network that is not implicated in the implementation of the dynamics . In that regard , the discrete movements studied here constituted full , repetitive ( flexion-extension ) cycles . Similar movements are sometimes referred to as continuous movements in the presence of temporal events [24] , [25] . We refer to them as ‘discrete’ as they are governed by fixed point dynamics . Regardless , please note that even though in many cases the exact timing of a discrete movement is hardly of importance , every discrete movement initiation ( be it embedded in a regular or irregular sequence of movements or not ) requires ‘external’ stimulation , which is ultimately timed . This also holds for an additional class of discrete movements , namely , point-to-point movements ( cf . [9] ) , in which two stable fixed points exist simultaneously ( see Supporting Information , and [15] ) . While our findings are by and large in line with the more ‘traditional’ and purely behaviorally-defined classification [2] as well as recent versions thereof in terms of movement continuity [24] , [25] , they also identify their limitation; continuous movements do not constitute a single class . This limitation indeed strengthens our call for a classification of movement rooted in mathematical theory that bears directly on the mechanisms underlying movement genesis . The movements at a slow pace , in particular at 0 . 5 Hz , under the ‘smooth’ instruction ( and for some participants under the ‘natural’ instruction ) were invariantly characterized by ( relatively ) irregular phase flows ( see Figure 4C ) . The Poincaré-Bendixson theorem [13] , [14] rules out topological structures other than fixed points ( and separatrices ) and limit cycles in two-dimensional phase space . The ( relatively ) irregular phase flows ( with indices of multiple fixed points ) may ( by hypothesis ) represent movements whose phase flow changes on a similar time scale as the movement . Such flows can be predicted for equilibrium point models [4]–[6] that , from a dynamical perspective , can be interpreted in terms of ( the relocation of ) a fixed point [26] . In fact , phase flow changes on the time scale of the movement also underwrite an alternative dynamical model [7] . Accordingly , discrete movements are accounted for by the destabilization and subsequent stabilization of fixed points interspersed by a time interval in which a limit cycle exists that effectively generates the ( discrete ) movement . The destabilization is accounted for by an external impact relative to the dynamics ( ‘behavioral information’ ) . In other words , discrete movement generation is non-autonomous according to this account also . The notion of time keepers versus timing resulting from movement dynamics are not new . On the contrary , these notions are central to two distinct theoretical camps ( the information processing perspective and dynamical system approach , respectively ) that have little interaction ( [27]; and see e . g . , the special issue of Brain & Cognition 48 , 2002 ) . The notion of a time keeper ( or central timer ) became firmly established by the well-known two-level timing model [28] , [29] . Accordingly , the behavioral expression in tapping movements – the often observed negative correlation between consecutive tapping intervals – is the resultant of the repetitive movement initiation by a central time keeper and the impact of the motor delays preceding and following each particular tap ( which are all random variables ) . Notwithstanding the various elaborations of ( ‘cognitive’ ) timing models ever since [30]–[33] , the notion of time keeping is inherently connected with abstract mental representations . In contrast , eschewing representational concepts , dynamicists view timing and coordination as properties arising from ( self-organized ) pattern formation processes [34]–[37] . Here , we elaborated on two distinct dynamical organizations and report evidence that humans ‘implement’ either of these depending on movement rate . In the non-autonomous scenario movement initiation ( and thus timing ) depends on a mechanism external to the dynamics . While we framed this in terms of time keeping , this should not be taken to imply that we adhere to a representational account thereof ( cf . [36] ) . In other words , the non-autonomous case should not be simply equated with a dynamical version of a two-level model ( notwithstanding the – to some extent superficial – similarity in terms of a distinction between ‘clock’ and ‘motor’ components ) . The implication of external timekeeper during discrete movements begs the question what neural structure ( s ) could fulfill this function ? Spencer and colleagues [25] showed that patients with cerebellar lesions have deficits in producing discontinuous but not continuous movements , which supports the idea that the cerebellum is implicated in timing in the non-autonomous but not autonomous case ( see also [38]–[40] ) . However , Schaal and colleagues [9] , using fMRI , reported contralateral activity in several non-primary motor areas and the cerebellum during discrete wrist movements that was absent during their rhythmic counterparts . This result favors the suggestion that timing is a property originating from a distributed neural network [41] , [42] . Indeed , the neural basis underlying timing remains yet to be elucidated . Implementing the present paradigm in the context of brain imaging may help establishing that aim . Finally , it has been repeatedly suggested that motor control is simplified through the use of ‘motor primitives’ , the motor system's elements thought of as its ‘building blocks’ . The modular organization of the vertebrae spinal motor system and the reproducibility of specifically coordinated muscle activity upon stimulation of certain modules ( neural circuits ) instigated the idea that motor behavior is organized along such hard-wired structures [43]–[45] . On a more abstract level , the two timing architectures we identified here qualify as candidate building blocks in human motor control .
We numerically investigate the equationin which a and b , and γ , represent parameters , ω represents the system's eigenfrequency , τ represent a time constant , and I the external stimulation . For all simulations we use τ = 1 , and if applicable , a stimulus duration corresponding to 80 ms and magnitude of 3 . 5 . For the mono-stable regime , the following parameter settings are implemented: γ = 1; ω = 1; a = [1 . 01 , 1 . 09] with steps of 0 . 02; b = [−0 . 1 , 0 . 8] with steps of 0 . 1; and I = [0 . 25 Hz , 4 . 00 Hz] with steps of 0 . 25 Hz . For the limit cycle regime , the implemented parameters are: a = 0; b = [−0 . 2 , 0 . 3] with steps of 0 . 1; and ω = [0 . 25 Hz , 4 . 00 Hz] with steps of 0 . 25 Hz . For each frequency ω , γ is chosen to as to ensure that the system oscillates with the appropriate frequency . All simulations are performed using a fourth-order Runge-Kutta method . Gaussian white noise ξ ( t ) is added to the evolution equations of the y-variable , where 〈ξ ( t ) 〉 = 0 , 〈ξ ( t ) ξ ( t ) 〉 = Q2δ ( t−τ ) , Q = 0 . 01 . The triangular brackets 〈·〉 denote time averages . Eight participants ( mean age = 27 . 9 years ) took part in the experiment . Seven participants were ( self-reported ) right-handed , one participant was left-handed . Participants reported an average of 2 . 75 years of musical experience with a minimum of 0 years and a maximum of 8 years . The protocol was approved by the Purdue University Committee on the Usage of Human Research Participants and was in agreement with the Declaration of Helsinki . Informed consent was obtained from all participants . Data were collected using a Polhemus Liberty-8 receiver ( 23×13×11 mm , 4 gm ) that was affixed to the participant's index finger with adhesive tape . This receiver was controlled by Matlab using an AuSIM-AuTrakMatlab USB driver and collection interface via library C++ calls . Three dimensional position data were collected at 240 Hz . The motion in the medio-lateral direction was used for further analysis . The flexion-extension movements were performed in the transverse plane involving no physical contact with any object . During the performance , the participants were seated at a 77-cm high table , and each participant rested the medial portion of his or her hand on a padded wooden block and Velcro held their hand in place . Ten trials were performed under three instruction conditions . Under each instruction condition , the participant was instructed to time the full finger flexion with the occurrence of the metronome tone . The instruction for the ‘natural’ condition was to do so in a manner that felt most natural . The instruction for the ‘smooth’ condition was to execute the movements as smooth ( sinusoidal ) as possible so as to be moving always ‘at an even pace’ . For the ‘discrete’ condition the instruction was to execute each complete flexion and extension movement as quickly as possible . In each condition five trials were performed with increasing metronome pace ( from 0 . 5 Hz to 3 . 5 Hz; step size 0 . 5 Hz ) and five trials with decreasing pace . Every frequency plateau lasted for 15 tones . Participants were instructed to quickly and smoothly adjust to changes in pace . A 30 second rest interval was provided between trials . Feedback was given after a trial if the participant's average cycle duration for any of the seven metronome paces had deviated more than 15 percent of the goal interval duration . The order of increasing or decreasing set of trials was performed in a blocked design . All participants performed the first condition ( ‘natural’ ) on day one . The order of the other two conditions was balanced for all participants . Each session lasted approximately one and a half hour . Human movement is inherently stochastic; its dynamics constitutes a deterministic and a stochastic ( i . e . , random ) component [19] , [34] , [35] . The future state of a stochastic process is conditional upon the probability for its state to be at a given time instant at a specific point in phase space , which can be described by probability distributions [34] , [46] . The computation of probability distributions allows one to disentangle the deterministic and stochastic dynamical components underlying stochastic processes [16]–[19] . Here , we extract these components to focus on the deterministic dynamics . Thereto , for each trial , we computed the movement velocity and normalized all position ( x ) and velocity ( y ) time-series to the interval [−1 , 1] . Next , using a grid size of 31 , we computed for all trials the conditional probability matrix , P ( x , y , t|x0 , y0 , t0 ) , that is , the probability to find the systems at state ( x , y ) at a time t given its state ( x0 , y0 ) an earlier time step t0 . Subsequently , we computed the Kramers-Moyal coefficients [16]–[20] representing the drift coefficient according to The coefficients Dx and Dy were averaged across the five trial repetitions for each participant , instruction condition and movement frequency . From the first two coefficients ( that represent the x- , and y-component of the corresponding velocity vector ) , we computed for each bin the angle θ between its corresponding velocity vector and that of each of its neighbors ( provided their existence ) according toin which u and v represent two neighboring vectors defined by Dx ( x , y ) and Dy ( x , y ) at position x and y in phase space . Next , we extracted the maximal value of θ in phase space . The existence of locally opposing vectors ( i . e . , with an angle of approximately 180° ) indicate the existence of a fixed point . We then computed for each instruction condition×movement frequency condition the mean and standard deviation of the maximal angle across participants and frequency order .
|
A fundamental question in motor control research is whether distinct movement classes exist . Candidate classes are discrete and continuous movement . Discrete movements have a definite beginning and end , whereas continuous movements do not have such discriminable end points . In the past decade there has been vigorous , predominantly empirically based debate whether this classification implies different control processes . We present a non-empirical classification based on mathematical theorems that unambiguously sets discrete and continuous rhythmic movements apart through their topological structure in phase space . By computational simulations of representative modes of each class we show that discrete movements can only be executed repetitively at paces lower than approximately 2 . 0 Hz . In addition , we performed an experiment in which human participants performed finger flexion-extension movements at various movement paces and under different instructions . Through a topological analysis of the flow in state space , we show that distinct control mechanisms underwrite human discrete and fast rhythmic movements: discrete movements require a time keeper , while fast rhythmic movements do not . Our results demonstrate that the human motor system employs different timing control mechanisms ( presumably via differential recruitment of neural subsystems ) to accomplish varying behavioral functions such as speed constraints .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"neuroscience/behavioral",
"neuroscience",
"neuroscience/motor",
"systems",
"neuroscience/theoretical",
"neuroscience",
"computational",
"biology/computational",
"neuroscience"
] |
2008
|
Distinct Timing Mechanisms Produce Discrete and Continuous Movements
|
Field-applicable tests detecting asymptomatic Mycobacterium leprae ( M . leprae ) infection or predicting progression to leprosy , are urgently required . Since the outcome of M . leprae infection is determined by cellular- and humoral immunity , we aim to develop diagnostic tests detecting pro-/anti-inflammatory and regulatory cytokines as well as antibodies against M . leprae . Previously , we developed lateral flow assays ( LFA ) for detection of cytokines and anti-PGL-I antibodies . Here we evaluate progress of newly developed LFAs for applications in resource-poor settings . The combined diagnostic value of IP-10 , IL-10 and anti-PGL-I antibodies was tested using M . leprae-stimulated blood of leprosy patients and endemic controls ( EC ) . For reduction of the overall test-to-result time the minimal whole blood assay time required to detect distinctive responses was investigated . To accommodate LFAs for field settings , dry-format LFAs for IP-10 and anti-PGL-I antibodies were developed allowing storage and shipment at ambient temperatures . Additionally , a multiplex LFA-format was applied for simultaneous detection of anti-PGL-I antibodies and IP-10 . For improved sensitivity and quantitation upconverting phosphor ( UCP ) reporter technology was applied in all LFAs . Single and multiplex UCP-LFAs correlated well with ELISAs . The performance of dry reagent assays and portable , lightweight UCP-LF strip readers indicated excellent field-robustness . Notably , detection of IP-10 levels in stimulated samples allowed a reduction of the whole blood assay time from 24 h to 6 h . Moreover , IP-10/IL-10 ratios in unstimulated plasma differed significantly between patients and EC , indicating the feasibility to identify M . leprae infection in endemic areas . Dry-format UCP-LFAs are low-tech , robust assays allowing detection of relevant cytokines and antibodies in response to M . leprae in the field . The high levels of IP-10 and the required shorter whole blood assay time , render this cytokine useful to discriminate between leprosy patients and EC .
Leprosy , a curable infectious disease caused by Mycobacterium leprae ( M . leprae ) that affects the skin and peripheral nerves , is one of the six diseases considered by the WHO as a major threat in developing countries [1] . Despite being treatable , leprosy often results in severe , life-long disabilities and deformities [2] due to delayed- or misdiagnosis . Transmission of leprosy is clearly unabated as evidenced by the number of new cases , 10% of whom are children , that plateaued at nearly 250 , 000 each year since 2005 [1] . Continued transmission in endemic areas likely occurs from the large reservoir of individuals who are infected subclinically . Thus , early detection of M . leprae infection , followed by effective interventions , is considered vital to interrupt transmission as highlighted by the WHO 2011–2015 global strategy [3] . Despite this pressing need , field-friendly tests that detect asymptomatic M . leprae infection are lacking , nor are there any biomarkers known that predict progression to disease in infected individuals . Lateral flow assays ( LFAs ) , are simple immunochromatographic assays detecting the presence of target analytes in samples without the need for specialized and costly equipment . Combinations of LFAs with up-converting phosphor ( UCP ) reporter technology are useful for detection of a variety of analytes , e . g . , drugs of abuse [4] , protein and polysaccharide antigens from pathogens like Schistosoma and Brucella [5] , [6] , bacterial and viral nucleic acids [7] , [8] and antibodies against M . tuberculosis , HIV , hepatitis virus and Yersinia pestis [9]–[11] . The phosphorescent reporter utilized in UCP-LFAs is excited with infrared light to generate visible light , a process called up-conversion . UCP-based assays are highly sensitive since up-conversion does not occur in nature , avoiding interference by autofluorescence of other assay components . Importantly , UCP-LF test strips can be stored as permanent records allowing re-analysis in a reference laboratory . In leprosy , the innate and adaptive immune response to M . leprae matches the clinical manifestations as substantiated by the characteristic spectrum ranging from strong Th1 immunity in tuberculoid leprosy to high antibody titers to M . leprae with Th2 cytokine responses in lepromatous leprosy [12] . In view of this spectral character , field-applicable tests for leprosy should allow simultaneously detection of biomarkers for humoral- as well as cellular immunity . Tests used in leprosy diagnostics include the broadly investigated serological assay detecting IgM against PGL-I [13] , [14] . Although this test is useful for detection of most multibacillary ( MB ) patients [15] , [16] , as the antibody levels correlate well with the bacillary load , detection of anti-PGL-I Ab has limited value in identifying paucibacillary ( PB ) leprosy patients [17] . In areas hyperendemic for leprosy more than 50% of young schoolchildren surveyed had positive anti-PGL-I responses [18] . Still , the vast majority of individuals with a positive antibody titer will never develop leprosy [13] . With respect to cellular responses in leprosy diagnosis , studies have focussed on M . leprae-unique antigens which can probe T-cell M . leprae-specific responses resulting in the identification of M . leprae ( -unique ) antigens that specifically induced IFN-γ production in M . leprae infected individuals [19] , [20] . Combined with serology , the use of these IFN-γ release assays ( IGRAs ) provided significant added value since they identified the majority ( 71% ) of PGL-I negative healthy household contacts in Brazil [21] while control individuals not exposed to M . leprae were IGRA-negative . Similar IGRAs allowed detection of the extent of M . leprae exposure along a proximity gradient in EC in one city in Brazil and in Ethiopia [22]–[24] . Although ELISA techniques , as used in IGRAs , are more widely applied than before , they still require laboratory facilities which are not available at all health centres in leprosy endemic areas . To accommodate ELISAs to field-applicable tests for leprosy diagnosis , we previously developed UCP-LFAs for detection of IFN-γ and IL-10 as well as antibodies against the M . leprae-specific phenolic glycolipid-I ( PGL-I ) for high-tech , laboratory-based microtiter-plate readers [25] , [26] . Since IFN-γ , the hallmark cytokine of Th1 cells , has generally been assessed as a biomarker to detect anti-mycobacterial immunity , we first developed a IFN-γ-UCP-LFA [25] . Recently , IFN-γ induced protein 10 ( IP-10 ) was found useful for detection of M . tuberculosis infection [27] and can also be used to indicate levels of M . leprae exposure and thereby the risk of infection and subsequent transmission [22] , [23] . Furthermore , since IP-10 is produced in large quantities , facilitating the use of simplified test platforms such as LFA [28] , we investigated its potential as an alternative to IFN-γ for leprosy diagnosis . Accordingly , we developed quantitative , dry reagent UCP-LFAs for field-detection of IP-10 and anti-PGL-I antibodies and evaluated these in a leprosy endemic area in Ethiopia .
This study was performed according to ethical standards in the Helsinki Declaration of 1975 , as revised in 1983 . Ethical approval of the study protocol was obtained from the National Health Research Ethical Review committee , Ethiopia ( NERC # RDHE/127-83/08 ) and The Netherlands ( MEC-2012-589 ) . Participants were informed about the study objectives , the required amount and kind of samples and their right to refuse to take part or withdraw from the study at any time without consequences for their treatment . Written informed consent was obtained from all study participants before venipuncture . HIV-negative , newly diagnosed untreated leprosy patients and healthy endemic controls ( EC ) were recruited at the Armauer Hansen Research Institute ( AHRI ) in Addis Ababa , Ethiopia , The Leiden University Medical Center ( LUMC ) and the Erasmus Medical Center ( EMC ) , The Netherlands from October 2011 until November 2012 . Leprosy was diagnosed based on clinical , bacteriological and histological observations and classified by a skin biopsy evaluated according to the Ridley and Jopling classification [2] by qualified personnel . EC were assessed for the absence of signs and symptoms of tuberculosis and leprosy . Staff members working in the leprosy centers or TB clinics were excluded as EC . Mantoux-negative , healthy Dutch donors recruited at the Blood Bank Sanquin , Leiden , The Netherlands were used as nonendemic controls ( NEC ) . None of these NEC had lived in or travelled to leprosy- or TB endemic areas , and , to their knowledge , had not experienced any prior contact with TB or leprosy patients . M . leprae candidate genes were amplified by PCR from genomic M . leprae DNA and cloned using Gateway technology ( Invitrogen , Carlsbad , CA ) with pDEST17 expression vector containing an N-terminal histidine tag ( Invitrogen ) [29] . Purified recombinant proteins were produced as described [22] , [29] and contained endotoxin levels below 50 IU per mg recombinant protein as tested using a Limulus Amebocyte Lysate ( LAL ) assay ( Cambrex , East Rutherford , NJ ) . Recombinant proteins were tested to exclude protein non-specific T cell stimulation and cellular toxicity in IFN-γ release assays using PBMC of in vitro PPD-negative , healthy Dutch donors recruited at the Blood Bank Sanquin , Leiden , The Netherlands . None of these controls had experienced any known prior contact with leprosy or TB patients . Within 3 hours of collection , venous heparinized blood ( 450 µl per well ) was incubated in 48-well plates at 37°C at 5% CO2 , 90% relative humidity with 50 µl of antigen solution ( 100 µg/ml ) . After incubation periods of 1 h , 4 h , 6 h or 24 h ( as indicated ) , 150 µl of supernatants were removed from each well and frozen in aliquots at −20°C until further analysis . Synthetic PGL-I ( ND-O-HSA ) and M . leprae whole cell sonicate were generated with support from the NIH/NIAID Leprosy Contract N01-AI-25469 ( available through the Biodefense and Emerging Infections Research Resources Repository listed at http://www . beiresources . org/TBVTRMResearchMaterials/tabid/1431/Default . aspx ) . Disaccharide epitope ( 3 , 6-di-O-methyl-β-D-glucopyranosyl ( 1→4 ) 2 , 3-di-O-methylrhamnopyranoside ) of M . leprae specific native PGL-I glycolipid was synthesized and coupled to human serum albumin ( ND-O-HSA ) as previously described by Cho et al . [30] . Inactivated ( irradiated ) armadillo-derived M . leprae whole cells were probe sonicated with a Sanyo sonicator to >95% breakage . IgM antibodies against M . leprae PGL-I were detected with natural disaccharide of PGL-I linked to human serum albumin ( ND-O-HSA ( 500 ng/well in 50 µl ) provided through the NIH/NIAID Leprosy Contract N01-AI-25469 ) as previously described [31] . Serum dilutions ( 50 µl/well; 1∶800 ) were incubated at RT for 120 min in flat-bottomed microtiter plates ( Nunc ) coated with NDO-HSA . After washing diluted enzyme linked secondary antibody solution ( anti-human IgG/IgM/IgA – HRP; Dako , Heverlee , Belgium; 50 µl/well ) was added to all wells and incubated at RT for 120 min . After washing diluted TMB solution ( 50 µl/well ) was added to all wells and incubated in the dark for 15 min at RT . The reaction was stopped by adding 50 µl/well 0 . 5 N H2SO4 . Absorbance was determined at wavelength of 450 nm . Samples with a net optical density at 450 nm ( OD ) above 0 . 149 were considered positive . The ELISA performance was monitored using a positive and negative control serum samples on each plate . For ELISAs 96 well Nunc MaxiSorp microtitre-plates were used and the presence of biotinylated antibody was detected enzymatically using streptavidin-HRP ( horse-radish peroxidase ) : IFN-γ was determined using anti-IFN-γ coating Ab mAb mO-13-32-22 ( U-CyTech Biosciences , Utrecht , the Netherlands ) and biotinylated anti-IFN-γ pAb pB-15-43-13 ( U-CyTech Biosciences ) as detection Ab . Culture supernatants were diluted 1∶2 in buffer ( 1% BSA/PBS ) and the cut-off value to define positive responses was set beforehand at 100 pg/ml . The assay sensitivity level was 40 pg/ml . Values for unstimulated cell cultures were typically <20 pg/ml . IP-10 was determined using anti-IP-10 capture Ab ( clone B-C50 ) and biotinylated anti-IP-10 detection Ab ( clone B-C55; Diaclone , France ) in culture supernatants diluted 1∶100 with dilution buffer . The cut-off value to define positive responses was set beforehand at 2 , 000 pg/ml . The assay sensitivity level was 40 pg/ml . Values for unstimulated cell cultures of NEC were typically <2 , 000 pg/ml . IL-10 was determined using anti-IL-10 mAb mO-13-10-12 ( U-CyTech Biosciences ) as coating Ab and biotinylated anti-IL-10 pAb mB-15-10-26 ( U-CyTech Biosciences ) as detection Ab in culture supernatants diluted 1∶2 . The cut-off value to define positive responses was set beforehand at 100 pg/ml . The assay sensitivity level was 10 pg/ml . Concentration values for unstimulated whole blood were typically ≤10 pg/ml . UCP conjugates specific for cytokines IP-10 , IL-10 , IFN-γ were prepared following earlier described protocols [26] , by conjugating 5 µg anti-IP-10 ( BC-50; Diaclone ) , 20 µg anti-IL-10 mAb ( coating mAb in ELISA , mO-13-10-12; U-CyTech ) or 25 µg anti-IFN-γ ( BB-1; Diaclone ) per 1 mg carboxylated UCP particles , respectively . Wet UCP conjugates were stored at a concentration of 1 mg/ml at 4°C . An UCP-IP-10 dry conjugate was made by drying 100 ng in a 5% sucrose matrix overnight at 37°C in 0 . 65 ml U-shape polypropylene tubes ( Ratiolab tubes for 96-well micro test plate , VWR International , Amsterdam , The Netherlands ) ; dried materials were stored in aluminum foil bags ( Lamigrip pouches Overtoom International , Den Dolder , The Netherlands ) with silica dry pellets at ambient temperature [6] , [32] . Reporter conjugates for detection of humoral immune response , an IgM- and Ig-specific UCP conjugates , were prepared as described earlier [9] , [26] by conjugation of 25 µg goat anti-human IgM ( I0759; Sigma-Aldrich , Saint Louis , MO , USA ) , protein-A ( Repligen Corp . ) or IgG/IgM/IgA/Kappa/Lambda–HRP ( Dako ) , respectively . Wet conjugates were stored at a concentration of 1 mg/mL at 4°C . Freeze dried pellets , so-called lyospheres , containing 100 ng UCPprotein A conjugate were produced ( Biolyph LLC , Hopkins , MN , USA ) and stored in vacuum-sealed glass vials as described earlier [33] . LF strips ( 4 mm width ) for IP-10 , IL-10 and IFN-γ were prepared with a test ( T ) line at 2 . 0 cm comprised of 50 ng anti-IP-10 BC-55 ( Diaclone ) , 700 ng anti-IL-10 mAb mO-10-10-28 ( U-CyTech Biosciences ) or 200 ng anti-IFN-γ BG-1 ( Diaclone ) respectively . The antibody pairs were identical to those used for ELISA but not containing a biotin hapten . LF strips for cytokine detection were further provided with a goat anti-mouse pAb ( M8642; Sigma-Aldrich ) flow-control ( FC ) line of respectively 100 ng and 200 ng at 2 . 5 cm . LF strips for detection of antibodies against PGL-I were provided with 50 ng synthetic PGL-I ( ND-O-HAS ) on the test ( T ) line and 100 ng rabbit anti-goat IgG ( G4018; Sigma-Aldrich ) on the flow-control ( FC ) line . LF strips for IP-10 and PGL-I multiplex detection were prepared using the same compositions as the strips for the individual targets , but now were provided with two T- and two FC-capture lines . Capture lines were separated by 4 mm located at 1 . 5 ( T1 , IP-10 ) , 1 . 9 ( T2 , PGL-I ) , 2 . 7 ( FC1 , goat anti-mouse ) , and 2 . 3 cm ( FC2 , rabbit anti-goat ) . The UCP-LFAs for cytokine detection ( IFN-γ , IL-10 , IP-10 ) comprise two phases , designated solution phase and immunochromatography phase [26] . Solution phase: 10 µl of 100-fold diluted sample ( translating to 0 . 1 µl undiluted sample ) for IP-10 and 10 µl undiluted sample for IL-10 and IFN-γ is mixed with 90 µl High Salt Lateral Flow ( HSLF ) buffer ( 100 mM Hepes pH 7 . 2 , 270 mM NaCl , 1% BSA ( w/v ) , 0 . 5% Tween-20 ( v/v ) ) containing 100 ng specific UCP reporter conjugate and incubated for 60 min on a thermoshaker at 37°C and 900 rpm . The immunochromatography phase: the above mixture is applied to cytokine specific LF strip and allowed to flow for at least 30 min . After immunochromatography , LF strips are scanned in a Packard FluoroCount microtiterplate reader adapted with an infrared laser . Upon IR excitation ( 980 nm ) , UCP reporter particles emit green light detectable using a 550 nm band pass filter . Results are displayed in histograms in relative fluorescence units ( RFUs ) measured at Test and Flow-Control lines , or as the ratio value between Test ( T ) and Flow-Control ( FC ) RFUs using Lateral Flow Studio software V 3 . 3 . 5 ( QIAGEN Lake Constance GmbH ) . For strip analysis in Ethiopia a lightweight portable LF strip reader with UCP capability was used ( UCP-Quant , an ESEQuant LFR reader custom adapted with IR diode; QIAGEN Lake Constance GmbH , Stockach , Germany ) [6] . Best reproducibility is obtained when analyzing completely dry LF strips , whereas wet LF strips generate lower T and FC signals . Ratio values between wet- and dry-scanned strips are not significantly different when scanned with readers with sufficient sensitivity that contain a high power IR laser and an adjustable photo multiplier [34] . Since wet-format assays require a sonication step , not suitable for field applications [6] , the IP-10-UCP-LFA was adapted to allow implementation of dry reagents ( dry conjugate and lyophilized buffer ) similar as described for Schistosomiasis [6] and RSV [33] . Next , the dry-format IP-10-UCP-LFA was transported to Ethiopia at ambient temperature and used by local staff after short instruction . In order to evaluate the field performance of these dry-format UCP-LFAs at the Ethiopian site , a lightweight dedicated UCP-LF strip analyzer was provided . For detection of anti-PGL-I IgM antibodies two protocols were used: a rapid sequential flow protocol without incubation using the UCPprotein-A or UCPαIgG/IgM/IgA/Kappa/Lambda conjugate , or a two phase protocol similar to the above described protocol for cytokine detection only using UCPαIgM instead of cytokine-specific UCP conjugates . The sequential flow protocol using the UCPprotein-A conjugate is referred to as consecutive flow ( CF ) as described [8] , [9] , [35] . The CF protocol comprised three sequential flow steps: first 40 µl of a diluted clinical sample ( 2 . 5% ( v/v ) in HSLF assay buffer ) , after 2 min followed by a wash step with 20 µl HSLF and a final flow after 5 min with 70 µl UCP-conjugate ( 100 ng in HSLF ) . Multiple strips can be handled simultaneously by prefilling 96 well ELISA microtitre-plates ( Nunc MaxiSorp ) with the appropriate three solutions and transferring LF strips from one well to the other . Immunochromatography is allowed to continue for at least 30 min before LF strips are analyzed ( see above ) . For the dry-format UCP-LFA to detect anti-PGL-I antibodies , dry UCPprot-A reagent in the form of lyospheres [2] was used . Simultaneous detection of IP-10 and anti-PGL-I IgM was performed following the two phase protocol described above for cytokine detection . The solution phase comprised the incubation ( 60 min; 37°C; 900 rpm ) of 10 µL 100-fold diluted sample ( translating to 0 . 1 µL of the original undiluted clinical sample ) with 90 µl HSLF buffer containing 100 ng of the UCPαIP-10conjugate ( wet ) and 100 ng of the UCPαIgM conjugate . The immunochromatography phase was identical to that described for the cytokine-only testing protocol and allowed to continue for at least 30 min before analysis of LF strips ( see above ) . Note that the above protocol may not be applicable when performing antibody detection with the UCPprotein-A conjugate due to unwanted interaction of protein-A with the UCPαIP-10 conjugate [26] . Differences in cytokine concentrations between test groups were analysed with the two-tailed Mann-Whitney U test for non-parametric distribution using GraphPad Prism version 5 . 01 for Windows ( GraphPad Software , San Diego California USA;www . graphpad . com ) . For correlations R2 was calculated with the Pearson correlation using GraphPad Prism version 5 . 01 . The statistical significance level used was p≤0 . 05 .
M . leprae unique antigens can be used to indicate M . leprae exposure using IFN-γ and IP-10 as read-outs [22] , [23] , [36] . Also , IFN-γ and IP-10 are associated with Th1-mediated protection against mycobacteria , whereas the anti-inflammatory cytokine IL-10 dampens Th1 cells' responses [37]–[39] . In view of the high levels of IP-10 produced compared to IFN-γ [22] , [28] and since , in contrast to IFN-γ , IP-10 is not affected by low CD4 counts in TB patients with HIV [28] , we investigated whether IP-10 , as an alternative to IFN-γ , can be applied as a pro-inflammatory biomarker . To evaluate the combined diagnostic value of IL-10 , IP-10 and IFN-γ , we first determined their concentrations by ELISAs in 24 h WBA of 11 Ethiopian leprosy patients ( 9 BL , 2 BT ) and 12 EC . In addition , anti-PGL-I antibodies were determined for each individual as well ( Figure 1 ) . The IP-10 production measured in WBA displayed a pattern similar to that of IFN-γ , although the overall IP-10 concentrations were much higher: median levels of both cytokines in response to M . leprae and ML2478 in patients' WBA were not significantly different from those for EC in this leprosy endemic area . These data are consistent with our previous findings , leading to the use of IFN-γ/IP-10 production in response to ML2478 to determine the level of exposure to M . leprae irrespective of infection [22] . In contrast , IL-10 concentrations in response to ML2478 , were significantly lower for EC ( Figure 1C ) . Since the balance of pro- and anti-inflammatory cytokines in response to M . leprae regulates the clinical outcome after infection , diagnostic tests for leprosy measuring both type of responses will be helpful in the decision on which individuals need ( preventive ) treatment . IP-10/IL-10 ratios for stimulated and unstimulated WBA samples demonstrated significantly different values between patients and EC , in particular for unstimulated samples ( Figure 1D ) . Finally , detection of a biomarker for humoral immunity , anti-PGL-I antibody levels , demonstrated significantly higher titers for leprosy patients , further contributing to a discriminating profile between leprosy patients and EC in leprosy endemic areas ( Figure 1E ) . Since short overall test-to-result times are preferred for diagnostic assays , the supernatants of WBA of Ethiopian leprosy patients and EC were analyzed for the presence of IFN-γ , IL-10 and IP-10 after 1 h , 4 h , 6 h and 24 h stimulation . For IFN-γ and IL-10 , levels that varied significantly from unstimulated samples were only detected after 24 h ( data not shown ) . For IP-10 , however , already after 6 h significant production was observed in antigen stimulated samples ( Figure 2 ) . Important to note is that after 6 h , IP-10 levels in ML2478-stimulated samples were significantly higher ( p = 0 . 02 ) in patients compared to EC ( Figure 2B ) , whereas no distinctive responses were observed for IFN-γ at that time point . PHA-induced IP-10 levels were high for all individuals after 6 h and substantial IP-10 levels were only detectable in M . leprae-stimulated samples after 24 h . Thus , besides the higher levels of IP-10 , also the shorter whole blood assay time required render IP-10 combined with ML2478 or as ratio with IL-10 directly in serum , a preferred pro-inflammatory biomarker to discriminate between leprosy patients and EC . For detection of IFN-γ , IL-10 as well as antibodies against M . leprae PGL-I , we previously developed up-converting phosphor lateral flow assays ( UCP-LFAs ) [25] , [26] . Because of the potential of IP-10 to identify M . leprae infection in a shorter test-to-result time as well as the value of IP-10/IL-10 ratios , we now selected IP-10 for UCP-LFA development , using the wet-format for IL-10 described previously [26] . Validation of these IL-10 and IP-10 UCP-LFA by comparison to ELISAs utilizing the same antibody pairs and antigen-stimulated WBA samples of non-endemic controls ( NEC ) , demonstrated good correlations between UCP-LFAs and ELISAs for IP-10 and IL-10 ( R2 0 , 854 and R2 0 , 816 , respectively; Figure 3 ) . In view of the greater stability in the field , dry assay format IP-10-UCP-LFA were produced and evaluated in Ethiopia as well: IP-10 values obtained in both wet and dry assays showed a good correlation ( R2 0 , 790; Figure 4A ) indicating the value for field application of the dry-format IP-10-UCP-LFA . Similarly , the unstimulated WBA samples were locally ( in Ethiopia ) tested for the presence of antibodies against PGL-I as well . Quantitive analysis of the UCPprot-A ratios and ELISA OD values correlated well ( R2 0 . 689; Figure 4B ) indicating 100% agreement in respect to serological status of the samples ( qualitative analysis ) . To further evaluate UCP-LF applications with this Ethiopian sample set , IL-10 levels of 84 samples ( 21 patients , 3 stimuli and medium ) were also tested , using the available wet-format IL-10-UCP-LFA in parallel with ELISA . Since the IL-10-UCP-LFA was used with 100-fold larger sample input than the IP-10 assay , some of the discrepancies observed for IL-10 between ELISA and UCP-LF assay were probably due to particulate material present in WBA samples . Despite these differences , IL-10-UCP-LFA and ELISA correlated well ( R2 0 , 735; Figure 4C ) . For direct comparison of single UCP-LFAs performance in a field- versus laboratory setting , the UCP-LF strips for IP-10 and anti-PGL-I antibodies analyzed in Ethiopia were sent to The Netherlands and re-analysed using a dedicated , high-tech UCP scanner , a Packard FluoroCount microtiter-plate reader adapted with an infrared laser ( 980 nm ) capable to scan 20 strips simultaneously . Comparison of ratios obtained in both tests showed an excellent correlations between both scanners ( IP-10: R2 0 , 960 and PGL-I: R2 0 , 901; Figure 5 ) , demonstrating that the UCP-LF strips can be stored as permanent record allowing re-analysis in a reference laboratory . Since leprosy endemic areas are often short of sophisticated laboratories , these results indicate that UCP-LFAs represent robust test suitable for resource-poor settings . IP-10 levels as well as anti-PGL-I antibody concentrations were present in high concentrations allowing reliable detection even with small amounts of serum thereby improving the robustness in field assays . To further simplify the use of the UCP-LFA for leprosy diagnostics in a field setting , we next developed a multiplex UCP-LFA for simultaneous detection of anti-PGL-I antibodies and IP-10 in whole blood samples , analogous to the earlier described anti-PGL-I/IL-10 multiplex UCP-LFA [26] . The advantage of this specific chemokine/antibody combination is that similarly diluted serum samples can be used , facilitating multiplex analysis of cellular and humoral immunity . For extensive comparison of single and multiplex UCP-LFAs Dutch leprosy patients' WBA samples were used as well to accommodate for more samples . Multiplex UCP-LFA and the single UCP-LFA for IP-10 and anti-PGL-I antibodies showed good correlations ( R2 0 , 961 and 0 , 897; Figure 6 ) demonstrating the applicability of this multiplex UCP-LFA .
Effective diagnostics are essential tools for the control , elimination and eradication of neglected diseases such as leprosy . Since leprosy endemic areas are often short of sophisticated laboratories , it is imperative to develop diagnostic tests for early detection of M . leprae infection that are suitable for field settings . The main requisite for such diagnostic tests is the selection of suitable biomarkers . WBA using M . leprae ( -specific ) antigens induce a ‘fingerprint’ of ( the ratio of ) pro- and anti-inflammatory cytokines that , combined with detection of anti-PGL-I antibodies , can be used as a biomarker profile for M . leprae infection . Notwithstanding the frequent use of IFN-γ , IP-10 represents an equally valid biomarker for pro-inflammatory responses to mycobacteria [22] , [23] , [27] , [36] , [40] , [41] . This chemokine is produced by various cell types , including monocytes/macrophages , and is involved in recruitment of lymphocytes and neutrophils to sites of inflammation . IP-10 can be used to differentiate between high and low M . leprae exposure levels [22] and it also provides a biomarker associated with type 1 reactions ( T1R ) in leprosy patients [42] , [43] . Moreover , IP-10 , is much less influenced by CD4 cell count and , in contrast to IFN-γ , can be used in HIV+ individuals [28] . Considering the similarities in IP-10 responses of M . leprae- and M . tuberculosis infected individuals , and the high concentrations in which it is produced , we developed a UCP-LFA for IP-10 and investigated its diagnostic potential for leprosy ( this study ) and TB in Africa ( Corstjens et al . , in preparation ) . Although most IGRAs require an antigen stimulation time of at least 24 h , we here demonstrate that IP-10 , in contrast to IFN-γ , already showed a significant divergence between Ethiopian leprosy patients and EC after 6 h stimulation with the M . leprae-unique protein ML2478 . This considerably reduces the overall assay time and could conveniently provide a sample-to-result on the same day . Since host immunity and immuno-pathogenicity in response to M . leprae comprises multifaceted interactions between a diversity of cells secreting different molecules , it is rather unlikely that only a single compound is linearly correlated to protection or to disease progression [44] . Diagnostic tests that determine ratios of different types of cytokines will therefore be informative regarding disease development after M . leprae infection [19] , [45] as was previously illustrated by IFN-γ/IL-10 and IFN-γ/IL-17 ratios in Mtb infected individuals [46] , [47] , but also for the development of T1R [42] . Relatedly , another valuable observation made here was the significant difference in IP-10/IL-10 ratios in sera of leprosy patients and EC , even without antigen stimulation . These data illustrate that the proportion of pro- to anti-inflammatory cytokines is consistent with clinical outcome after infection . Consequently , over time changes in the IP-10/IL-10 ratio for one individual will provide relevant clinical information with respect to the outcome of infection . Dry-format UCP-LFAs are ideally suited for performance in the field and can be shipped and stored conveniently at ambient temperature and have prolonged shelf life of more than two years in African settings [6] . In this study we selected IP-10 and anti-PGL-I antibodies for field-evaluation of the dry-format UCP-LFA , and development of dry-format UCP-LFA for more analytes is in progress . This evaluation showed that both dry-format UCP-LFAs were equally sensitive as ELISAs and could be applied in the concentration range of 100 to >100 , 000 pg/ml . Also , the availability of affordable and portable UCP-LF strip readers showed suitability of the assay in field settings where ELISA equipment is not available or is more challenging to use . The LF strips were read with an easy to operate , portable reader that allows full instrument-assisted assay analyses avoiding operator bias . Due to the chemical stability of the assay components , the strips can be kept in patients' files and read again after long periods of time . Besides the speed and ease of performance , another advantage of the UCP-LFA is that multiple analytes can be detected on the same LF-strip . Feasibility of multiplexed analysis was demonstrated previously for IL-10 and anti-PGL-I antibodies in spiked sera [26] . In this study multiplexing was successfully shown for IP-10 and anti-PGL-I antibodies in whole blood samples . Although the current UCP-LFA conditions for IL-10 quantitation demand a 100-fold larger sample input than the IP-10 assay , a single strip allowing quantitative detection of IP-10 , IL-10 as well as anti-PGL-I antibody detection is feasible . Revision of the position ( distance from the sample pad ) and antibody load of the test lines , would allow the use of 1 µL samples instead of the currently applied 0 . 1 and 10 µL for IP-10 and IL-10 respectively . Moreover , multiplexing can be achieved by running two or more LF strips from a single sample in parallel as was for instance described for a simple multiple channel device running ten UCP-LF strips from a single sample [11] . This study describes the first steps towards development of a UCP-LFA as a field test measuring pro- and anti-inflammatory cellular- as well as humoral immunity to M . leprae , thereby including read-outs for multiple classifications of the leprosy spectrum . Such tests can be useful tools in leprosy control programs for classification of leprosy and allow early diagnosis of leprosy or leprosy reactions , leading to timely treatment and reduced transmission .
|
Leprosy is one of the six diseases considered by WHO as a major threat in developing countries and often results in severe , life-long disabilities and deformities due to delayed diagnosis . Early detection of Mycobacterium leprae ( M . leprae ) infection , followed by effective interventions , is considered vital to interrupt transmission . Thus , field-friendly tests that detect asymptomatic M . leprae infection are urgently required . The clinical outcome after M . leprae infection is determined by the balance of pro- and anti-inflammatory cytokines and antibodies in response to M . leprae . In this study , we developed lateral flow assays ( LFA ) for detection of pro-inflammatory ( IP-10 ) vs . anti-inflammatory/regulatory ( IL-10 ) cellular immunity as well as antibodies against M . leprae and evaluated these in a field setting in Ethiopia using lightweight , portable readers . We show that detection of IP-10 allowed a significant reduction of the overall test-to-result time from 24 h to 6 h . Moreover , IP-10/IL-10 ratios in unstimulated plasma differed significantly between patients and EC , which can provide means to identify M . leprae infection . Thus , the LFAs are low-tech , robust assays that can be applied in resource-poor settings measuring immunity to M . leprae and can be used as tools for early diagnosis of leprosy leading to timely treatment and reduced transmission .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"humoral",
"immunity",
"infectious",
"diseases",
"medicine",
"and",
"health",
"sciences",
"clinical",
"laboratory",
"sciences",
"immune",
"cells",
"diagnostic",
"medicine",
"cell",
"biology",
"animal",
"cells",
"immunity",
"clinical",
"immunology",
"immunity",
"to",
"infections",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"immunology",
"acquired",
"immune",
"system",
"immune",
"response",
"immune",
"system"
] |
2014
|
Field-Evaluation of a New Lateral Flow Assay for Detection of Cellular and Humoral Immunity against Mycobacterium leprae
|
Vanillyl alcohol oxidase ( VAO ) is a homo-octameric flavoenzyme belonging to the VAO/PCMH family . Each VAO subunit consists of two domains , the FAD-binding and the cap domain . VAO catalyses , among other reactions , the two-step conversion of p-creosol ( 2-methoxy-4-methylphenol ) to vanillin ( 4-hydroxy-3-methoxybenzaldehyde ) . To elucidate how different ligands enter and exit the secluded active site , Monte Carlo based simulations have been performed . One entry/exit path via the subunit interface and two additional exit paths have been identified for phenolic ligands , all leading to the si side of FAD . We argue that the entry/exit path is the most probable route for these ligands . A fourth path leading to the re side of FAD has been found for the co-ligands dioxygen and hydrogen peroxide . Based on binding energies and on the behaviour of ligands in these four paths , we propose a sequence of events for ligand and co-ligand migration during catalysis . We have also identified two residues , His466 and Tyr503 , which could act as concierges of the active site for phenolic ligands , as well as two other residues , Tyr51 and Tyr408 , which could act as a gateway to the re side of FAD for dioxygen . Most of the residues in the four paths are also present in VAO’s closest relatives , eugenol oxidase and p-cresol methylhydroxylase . Key path residues show movements in our simulations that correspond well to conformations observed in crystal structures of these enzymes . Preservation of other path residues can be linked to the electron acceptor specificity and oligomerisation state of the three enzymes . This study is the first comprehensive overview of ligand and co-ligand migration in a member of the VAO/PCMH family , and provides a proof of concept for the use of an unbiased method to sample this process .
Vanillyl alcohol oxidase ( VAO ) is a flavoenzyme involved in the mineralisation of aromatic compounds in the soil fungus Penicillium simplicissimum . VAO is active with a wide range of para-substituted phenolic compounds , but received its name because the enzyme initially was found to catalyse the oxidation of vanillyl alcohol ( 4-hydroxy-3-methoxybenzyl alcohol ) to vanillin ( 4-hydroxy-3-methoxybenzaldehyde ) [1] . The natural substrate of the enzyme could be 4- ( methoxymethyl ) phenol , as it is the only substrate known to induce VAO expression [2] . VAO is one of the designating members of the VAO/PCMH family of oxidoreductases that share a conserved FAD-binding domain [3–5] . Members of this family often bind their flavin cofactor in a covalent mode . The closest known and characterised relatives of VAO in the VAO/PCMH family are the bacterial enzymes eugenol oxidase ( EUGO ) and p-cresol methylhydroxylase ( PCMH ) . While substrate specificities of these three enzymes differ , they all require their substrates to be para-substituted phenols [6–10] . The amino acid sequence of VAO is most similar to EUGO , sharing 47% sequence identity , while VAO and PCMH share 40% sequence identity . Interestingly , the sequence identity between the two bacterial enzymes is only 33% . All three enzymes bind their FAD cofactor covalently . In VAO , this bond is formed via an autocatalytic process , resulting in a covalent link between His422 and FAD [11] . In EUGO , the flavin is linked to His390 [12] , while in PCMH , the flavin is covalently bound to Tyr384 [13] . VAO is a homo-octamer composed of a tetramer of dimers , where each subunit consists of 560 amino acids . Each subunit is organised into two domains , the FAD-binding domain and the cap domain [9] , finding the active site at their interface ( Fig 1 ) . In vitro , VAO forms a mixture of dimers and octamers , with the octamers prevailing at physiological ionic strength [14 , 15] . EUGO and PCMH have different quaternary structures . EUGO is exclusively homodimeric , while PCMH consists of a heterotetramer . This heterotetramer consists of two FAD-binding subunits forming a homodimer and two cytochrome c subunits . The homodimeric structure of the FAD-binding subunits is the same in all three enzymes . See also supplementary S1 Fig for an overview of the quaternary structures of these three enzymes . Important active site residues of VAO are Tyr108 , Asp170 , His422 , Tyr503 and Arg504 ( see the zoom into the active site in Fig 1 ) . Tyr108 , Tyr503 and Arg504 are proposed to be crucially involved in the deprotonation of substrates upon their arrival in the active site [9 , 16] . Asp170 has a multifunctional role . It is involved in the autocatalytic incorporation of the flavin cofactor and assists in increasing the redox potential of VAO ( making it less negative ) [17] . Asp170 can also act as active site base [17] , and is involved in the enantioselective hydroxylation of 4-alkylphenols [18] . His422 increases the oxidation power of the FAD cofactor by increasing its redox potential ( making it less negative ) through covalent flavin binding [11 , 19] . The catalytic mechanism of VAO involves two half-reactions . In the reductive half-reaction , the enzyme-bound flavin is reduced by the phenolic substrate to generate a quinone methide intermediate [6 , 20] . In the oxidative half-reaction , the reduced flavin is oxidised by dioxygen , generating hydrogen peroxide . Depending on the substrate , there are differences in the reaction mechanism . With 4- ( methoxymethyl ) phenol , water addition to the quinone methide intermediate gives an unstable hemiacetal , which decomposes to 4-hydroxybenzaldehyde . VAO can produce vanillin in a one-step reaction from vanillyl alcohol as well as in a two-step reaction from p-creosol , where vanillyl alcohol is the decomposition product of the initially formed air-stable flavin-creosol adduct [21] . The reaction with p-creosol ( Fig 2 , plates 1 to 5 ) is rate-limited by the extremely slow decomposition of the flavin-N5 substrate adduct ( Fig 2 , plate 4 ) [9] . In the reaction of vanillyl alcohol to vanillin , there is no addition of water to the vanillyl alcohol quinone methide ( Fig 2 , plates 6 to 8 ) [20] . Although many details of the catalytic mechanism of VAO have been uncovered , it is unknown how the reaction participants enter and exit the active site . No path for solvent or ligand access to the active site is visible from the crystal structure of VAO . It is of fundamental interest to understand how reaction participants enter an enzyme’s active site to be able to identify catalytic bottlenecks . Such bottlenecks can then be used to guide substrate or enzyme redesign to eliminate them . This work provides a proof of concept for the use of an unbiased method to sample entry and exit pathways of a complex enzyme system . We aim to discover the migration path ( s ) of reaction partners of VAO for the reaction from p-creosol to vanillin . Recent studies , with other enzymes , have pointed to the importance of remote residues for catalysis or the presence of multiple paths for ligands [23–27] . It is also worthy to note that during enzyme catalysis , the entry and exit of reaction partners may result in traffic jams in migration paths . We try to address these issues by modelling migration paths of reaction partners of VAO . Given the large size of VAO and the involvement of multiple ligands , it is conceivable that substrates and products use different paths . In this study , we used Protein Energy Landscape Exploration ( PELE ) , a Monte Carlo based program [28] , to investigate the entry and exit paths of three phenolic ligands in VAO . Protonated and deprotonated phenolic ligands were used to be able to observe possible differences in binding energies during the simulations and thus establish when substrate deprotonation and product re-protonation happen in the reaction pathway of VAO . An entry and exit path for phenolic ligands at the subunit interface , along with two additional exits paths , leading through the cap domain or the FAD-binding domain , were identified . It was found that protonated p-creosol freely migrated to the active site , while deprotonated p-creosol did not . In addition , the behaviour of dioxygen and hydrogen peroxide was modelled . Migration via the flavin re side turned out to be the shortest path for these molecules .
The initial coordinates for the VAO protein were taken from the 1VAO entry in the protein data bank ( PDB ) [9] . Although the protein is assembled as an octamer , due to computational limitations only the dimer , the smallest functional unit of VAO , was modelled . No allosteric regulation has been observed for VAO in its dimeric or octameric oligomerisation state . Loop deletion mutagenesis has revealed that the VAO dimer has similar catalytic properties as the octamer [29] . The protein structure was prepared assisted by the protein preparation wizard available in the Schrödinger software package [30] . All crystal water molecules were removed and missing hydrogen atoms were added to protein residues . Residues such as histidines , glutamates and aspartates were inspected for appropriate protonation states at pH 7 . 5 to match experimental conditions . PROPKA [31] and H++ web server ( http://biophysics . cs . vt . edu/H++ ) [32] were employed to determine the pKa of each residue . His56 , His61 , His313 , His506 and His555 were modelled as ε protonated while all other His residues were δ protonated . Furthermore , the NE2 atom in the imidazole ring of His422 has no hydrogen atom since it is covalently bound to the C8M atom of the FAD cofactor and considered to be neutral . Missing residues from position 41 to 47 , forming a loop at the surface of the protein , were added and minimised using PRIME [33] . The final system was composed of a total of 17674 atoms . Six phenolic and two small ligands were used including: COP , COD ( p-creosol , where the P and D refer to the protonated and deprotonated state of the ligand; this notation is identical for all molecules ) , VAP , VAD ( vanillyl alcohol ) VNP , VND ( vanillin ) and dioxygen and hydrogen peroxide . All charges for the FAD cofactors were obtained through quantum mechanics/molecular mechanics ( QM/MM ) calculations inside the protein environment with an 8 Å layer of explicit waters . The QM region included all the isoalloxazine ring atoms and the QM/MM cut was made between the C1 and C2 atoms of the FAD molecule . All the ligands were optimized through QM calculations in an implicit solvent . QM/MM calculations used the all atom OPLS2005 force field [34] , M06 functional [35] with the 6-31G** basis set [36] and Poisson Boltzmann Finite element ( PBF ) implicit solvent [37] ( same functional and basis set used in QM ) . The parameter files for all ligands and the FAD cofactor are supplied as supplementary information S1 . PELE [28] was used to map the migration ( exit and entry ) path of different ligands between the VAO active site and the solvent . PELE is a Monte Carlo based algorithm that produces putative new configurations through a sequential ligand and protein perturbation scheme , side chain prediction and minimisation steps . A detailed description of the PELE methodology can be found elsewhere [28] . Ligand perturbation involves a random translation and rotation , while protein perturbation involves a displacement on the alpha carbon following one go the six lowest Anisotropic Normal Modes ( ANM ) [38] . These steps compose a move that is accepted ( defining a new minimum ) or rejected based on a Metropolis criterion for a given temperature . The combination of ligand and protein backbone perturbations results in an effective exploration of the protein energy landscape . This approach is capable of reproducing large conformational changes associated with ligand migration and has already been shown to produce reliable results [39–41] . For the entry simulations , the ligand was placed in six different positions at the bulk solvent . Similar parameters to the exit simulations were used , but with larger translations ( from 1 . 25 and 2 . 5 Å ) during substrate perturbations . The direction of such perturbations was kept for three consecutive steps . Moreover , for the entry simulations an adaptive scheme was used ( using the new C++ version of PELE [42] plus an OBC solvent [43] ) . After a short simulation of 12 Monte Carlo steps , the adaptive scheme clustered the ligand position and new initial conditions were chosen , prioritizing those clusters with less population . Due to the use of different versions of PELE , the binding energies differ by a level of magnitude for ligands migrating in and out of VAO ( see Results ) . In this context , it is important to note that the binding energies calculated by PELE are indicative of favourable or unfavourable positions for the ligand . While we refer to them as binding energies , they are interaction internal energies between the protein and the ligand , with a significant larger value than experimental ones , and should be analysed in a qualitative manner . The exit simulation protocol was identical for all substrates and began by manually docking the phenolic molecule ( COP , COD , VAP , VAD , VND , VNP ) to the si side of the isoalloxazine ring of the flavin in the active site based on crystallographic evidence from the VAO structure with isoeugenol ( PDB ID: 2VAO ) [9] . Simulations were performed in equal number starting from both protein subunits . The phenolic ligand was then perturbed with random translations ( from 0 . 75 and 1 . 75 Å ) and rotations ( 0 . 05 and 0 . 25 rad ) and requested to move away from the active site using PELE’s spawning procedure . This procedure selects a reaction coordinate ( e . g . an atom-atom distance , ligand RMSD ) and abandons the trajectories with worst values in the reaction coordinate to restart the simulation at the best value ( which is constantly updated ) . Trajectories are abandoned only if they fall behind a user-defined range ( from the best value ) . In our case , the spawning coordinate was the distance of the ligand center of mass to a point in the void volume of the active site ( X , Y , Z coordinates 92 , 25 , 41 or 108 , 47 , 70 for simulations starting in the active site of chain A or B , respectively ) , with the distance range from the spawning coordinate being 4 Å . Protein perturbation was based on the six lowest normal modes [44] and side chain sampling including all residues within 8 Å from the ligand . Finally , after global minimisation had optimised the new configuration , this configuration was filtered with a Metropolis acceptance test . In this test , the energy is described with an all atom OPLS2005 force field [34] with a surface generalized Born solvent model [45] . Simulations were stopped when the wall clock limit time ( 48h ) was reached or the phenolic ligand exited the protein completely . A total of 800 independent trajectories were produced for each ligand . Simulations exploring exit paths from the active site of the small ligands dioxygen and hydrogen peroxide were performed using a non-biased approach as described previously [40] , using random translations and rotations appropriate to achieve continuous paths ( between 0 . 5 and 1 . 5 Å and 0 . 1 to 0 . 3 rad , respectively ) . Data analysis was performed using VMD ( Version 1 . 9 . 1 , [46] ) , PyMOL ( The PyMOL Molecular Graphics System , Version 1 . 7 Schrödinger , LLC . ) , and R ( Version 3 . 3 . 1 [47] ) . Figures were created using the aforementioned programs , with the addition of GnuPlot ( Version 4 . 4 [48] ) and ChemDraw ( Version 15 . 1 . 0 . 144 , PerkinElmer Informatics ) . Atom contacts between the ligands and the protein were calculated using a python script . A contact was defined as a distance smaller than 2 . 7 Å between an atom from any given residue and an atom belonging to the ligand . Root mean square fluctuation ( RMSF ) values of residues in the simulations were calculated using VMD ( Version 1 . 9 . 1 , [46] ) . Only RMSF values from residues in the chain of the dimer where the ligand was placed were used for the analysis . To compare the structural conservation of path residues , the crystal structures of VAO , EUGO and PCMH ( 1VAO , 5FXE and 1DII respectively ) were aligned using the PROMALS3D multiple sequence and structure alignment server [49] . The crystal structures were also compared using PyMOL ( The PyMOL Molecular Graphics System , Version 1 . 7 Schrödinger , LLC . ) and the built in alignment function . To compare the sequence conservation of path residues , a set of VAO- , EUGO- and PCMH-like sequences was identified . The protein sequences from crystal structures 1VAO , 1DII and 5FXE were used as queries for similarity searches with Consurf [50] , searching the uniref90 database using default settings for the search of homologs . This approach resulted in three alignments ( one each for VAO- , EUGO- and PCMH-like sequences ) . The obtained alignments were manually checked and VAO- and EUGO-like sequences that did not contain a histidine residue corresponding to His422 and His390 respectively were removed . PCMH-like sequences not containing a tyrosine residue corresponding to Tyr384 were removed as well . This selection was made because the named residues covalently bind the FAD cofactor , which has a significant impact on enzyme function [10–12] . The homology search resulted in 121 , 117 and 30 VAO- , EUGO- and PCMH-like sequences remaining in the respective alignments ( named MSAVAO , MSAEUGO and MSAPCMH , respectively ) . These alignments where then merged using the MAFFT webserver ( mafft . cbrc . jp/alignment/software/merge . html ) to match the residue positions ( resulting in MSAall ) . The alignments were then uploaded to Consurf [49] to calculate the residue variety in percentages , always using the pdb structure 1VAO as query . This approach ensures that the same alignment positions are compared when the preservation of different residues in the three alignments is compared as described below; e . g . that residue number 422 in VAO is always compared to the residue present in the other sequences at the corresponding position in the three alignments . All alignments ( MSAVAO , MSAEUGO , MSAPCMH and MSAall ) can be found as supplementary information S1 Alignment . fa , S2 Alignment . fa , S3 Alignment . fa and S4 Alignment . fa , respectively . An R script was used to analyse the preservation of different groups of residues in these alignments , amongst which residues involved in the paths identified in this study . We defined nine different groups of residues: the entire protein ( 560 amino acids ) , residues forming the cap domain ( from residue 270 to 500 ) , the re path residues ( 22 residues ) , the FAD path residues ( 36 residues ) , the cap path residues ( 25 residues ) , the subunit interface path residues ( 53 residues ) , a selection of random residues ( 58 ) , surface residues of VAO ( 90 residues ) and residues at the subunit interface ( 86 residues ) . Surface residues were defined using the built in tool of the SwissPDBviewer , with default settings ( 30% accessibility of residues by solvent in dimeric VAO ) [51] . Interface residues , meaning residues at the subunit interface , were defined as residues within 4 Å of either subunit . Preservation of these residues was analysed based on similar or dissimilar residues . Similar residues were defined according to the BLOSUM62 similarity parameters [52] , resulting in six sets of residues ( Trp , Tyr and Phe; Met , Ile , Leu and Val; His , Arg and Lys; Asn , Asp , Gln and Glu; Ser , Thr , Pro Ala and Gly; and Cys . ) The preservation of residues was analysed per position in the alignments and a cutoff of 50% preservation was used to categorise them as similar in the alignment . If preservation of the similar residues was above 90% , they were categorised as conserved in the alignment . This analysis was performed for MSAVAO , MSAEUGO and MSAPCMH , for each of the nine groups of residues . This resulted in a percentage of similarity and a percentage of conservation for each group of residues for each sequence alignment of VAO- , EUGO- and PCMH-like sequences . This analysis did not yet allow us to compare the conservation or similarity of these three alignments to each other . Different residues can be preserved in each of the three alignments , but we were interested in finding out how many residues are preserved in all three alignments ( MSAall ) . We therefore performed the following analysis: we used the same definitions for similarity and conservation as above , but this time analysed the preservation of the nine groups of residues in MSAall . We added additional categories to this analysis: dissimilar and PCMH-unlike . Dissimilar positions were defined as positions less than 50% similar . PCMH-unlike positions are dissimilar positions for which VAO-like and EUGO-like sequences share a similar residue , but differ from PCMH . This resulted in percentages of similarity and conservation as well as percentages of dissimilarity and PCMH-unlikeness .
Using the adaptive PELE simulations , the surface of the VAO dimer was explored for COP entry , see Fig 4A . As stated , the ligand was placed in the bulk solvent and allowed to freely explore the surface and any possible entry path ( s ) . It was found that using this approach , the ligand moved into the active site exclusively via the subunit interface . In order to refine the energy landscape along the entry path , we started additional simulations from the surface of the subunit interface with a reduced translation range ( maximum ligand translation of 1 . 5 Å ) . These refinement simulations confirmed the entry for COP , but indicated that COD was unable to enter the protein via this path . Binding energies calculated showed that it was energetically favourable for COP to reach the active site via the subunit interface , but energetically unfavourable for COD to do the same ( Fig 4C ) . The COP entry path , which is identical to the third exit path described below , involves mainly residues Arg183 , Val185 , Asp192 , His193 , Met195 , Met303 , Ile428 , Met462 , His466 Ile468 , Tyr503 and Arg504 . Residues His466 and Tyr503 are easily changing conformation to grant the ligand access to the active site . This path is symmetrical , leading from the starting point at the subunit interface to either of the two active sites of the dimer , as can be seen in Fig 4C . The most contacted residues were His193 , Met195 and Ile428 and the most flexible residues were Asp192 , Gly462 , His466 and Tyr503 , as can be seen in Fig 4B . To identify exit paths , simulations were performed for the dimeric protein and the ligands were placed into one of the two subunits and both subunits were used as starting points in equal amounts . We observed symmetry in our data ( see Fig 3 ) : some paths were fully symmetrical in simulations starting from either subunit and other paths were partly symmetrical , most likely due to incomplete sampling as can be seen from Figs 3 and 4A . The fact that we observed identical results independent of the starting point is a good indication that all potential paths have been sampled . Extrapolating data from each of the two subunits to the other subunit gave fully symmetrical paths , which were continuously connected . Three independent exit paths were observed for simulations with phenolic ligands , as shown in Fig 3 and parts A of Figs 4 , 5 , S2 and S3 . The paths shown in different colours correspond to i ) the cap path ( in orange ) where ligands exit the protein through the cap-domain , ii ) the FAD path ( in blue ) where ligands traverse the protein through the FAD-binding domain , and iii ) the subunit interface path ( in magenta ) , which corresponds to a path connecting the subunits ( and agreeing with the entry path described above ) . Not all simulations led to the ligand leaving VAO . Some simulations showed that the ligand also moved from one path to the other , resulting in mixed paths . Only one of the three paths , the cap path , is exclusively connecting the active site to the solvent . The other two paths identified , the FAD path and subunit interface path , lead to interfaces between subunits or dimers of VAO . Paths leading to the interface of subunits or dimers allow the enzyme on one hand to connect the different active sites of the subunits ( Fig 3 and parts A of Figs 4 , 5 , S2 and S3 ) . On the other hand , these paths also allowed the ligands to exit via the interface of the dimer or the subunit , as there is sufficient space for the ligands available there . Because these paths can be expected to work into both directions , in and out , they could also act as another access point for the ligands to the active site . However , analysis of residue flexibility represented by RMSF data as well as atom contact counts has to be taken into account to discover possible bottlenecks in these paths . The RMSF values and atom contacts calculated per simulation were combined and plotted in boxplots ( parts B and C of Figs 4 , 5 , S2 and S3 ) . The more flexible residues , mostly located at the surface , were often also the most contacted ones . Those most flexible surface residues were Arg274 and Arg325 , while the buried ones showing larger flexibility were Arg398 , Tyr503 and Arg504 . Residue Lys545 was the most contacted residue for all ligands except VNP . Other flexible and/or highly contacted residues in simulations with most ligands were Trp98 ( not with VNP ) , Ile100 and Asn105 ( not with VAD and VNP ) Val185 ( not with VAP ) , Arg312 ( not with VAD and VNP ) , surface residue Ser329 ( not with COD ) and Leu507 ( not with VNP ) . The binding energies of all the phenolic ligands were on comparable scales , but the range of the calculated binding energies varied between ligands , as can be seen in supplementary S2C Fig . The binding energies plotted against the distance of the ligand from the FAD spread least towards higher energy values for COP and VAP compared to COD and VAD . When comparing the data from the entry path and the exit paths , COP showed very little spreading of the data and presents lower binding energies than COD in both cases . Combining this information suggests that it is energetically more favourable for protonated substrates to migrate into and out of VAO than for deprotonated substrates or products . In the following sections , each path and the behaviour of path residues will be described individually . In this path , ligands migrated to the “upper-part” of the active site ( orange in Figs 3 , 5 , S2 and S3 ) , passing Glu410 , Trp413 , Arg312 and Arg398 to exit the protein in the cap domain . Arg398 and Trp413 have been proposed to be involved in a size-exclusion mechanism that could limit the size of the substrate-binding pocket [9] . The cap path requires the ligands to pass Arg312 and Arg398 , which are among the most frequently contacted and most flexible residues for most ligands . Arg398 interacts with Glu410 and Trp413 and it is this interaction that the ligands need to disrupt to exit the active site via this route . Simulations often showed ligands to be trapped between the two arginines as they tried to pass this portal . This path was the shortest of the three identified for the phenolic ligands , but contained a bottleneck composed of residues Arg312 , Arg398 , Glu410 and Trp413 ( as can be seen in Figs 3 , 4 , S2 and S3 ) . Combining the data for all the ligands , residues belonging to the cap path that are among the most flexible or contacted residues were Leu171 , Arg312 , Leu316 , Tyr354 , Pro390 , Glu391 , Asn392 , Val394 , Arg398 and Thr401 . In this path ( blue in Figs 3 , 5 , S2 and S3 ) , ligands migrated through Tyr503 and Arg504 to move closely along the FAD phosphate-ribityl chain towards the adenosine monophosphate part of the cofactor and to exit the protein between Trp98 , Leu507 and Lys545 . Trp98 is required to move aside , which is reflected in the RMSF and contact data ( as can be seen in Fig 3B and 3C in the middle panel and supplementary S2A and S2B Fig ) . This path led to the dimer-dimer interface in octameric VAO , where we find sufficient space for the ligand to exit VAO or migrate to the neighbouring dimer in the octameric symmetry unit . However , Trp98 , Leu507 and Lys545 constituted a bottleneck in this path . Combining the data for all the ligands , residues belonging to the FAD path that were among the most flexible or contacted residues were Trp98 , Pro99 , Asn105 , Arg114 , Val115 , Ser118 , Phe424 , Thr505 , Leu507 , Met510 , Lys545 and Ser546 . In this path , ligands moved towards the subunit interface of the VAO dimer ( magenta in Figs 3 , 5 , S2 and S3 ) . The behaviour of different ligands modelled varied slightly in this path and they followed different minor variations of it , however all these minor paths led to the subunit interface . Simulations that were started with a ligand in one subunit sometimes showed the ligand completely entering the active site of the other subunit . Ligands were frequently found in the upper part of the subunit interface , where they were already surrounded with solvent . As stated , simulations mapping the entry path also showed the ligand passing through the same area . Three residues clearly involved in the subunit interface path , but not showing up as frequently contacted or highly flexible residues in the analysis , were Asp192 , Met195 and Glu464 . Ligands passed quickly through a portal formed by Glu464 , Asp192 and Met195 , and from there directly into the subunit interface ( for an illustration see supplementary S4 Fig ) . The portal formed by these three residues appeared to be the easiest for the ligands to pass through , not requiring significant movements of large residues and not involving a high frequency of contacts . The portal in this path connects the two active sites in the VAO dimer to each other . Additional residues involved in this path are His466 and Tyr503 , which were moving aside to grant access to the active site in the entry simulations and were also allowing ligands to exit . See Fig 6A for details on the movement of His466 and Ty503 as well as Met195 in our simulations . Residues that were in contact with ligands and that line the subunit interface were Arg463 , Arg300 and Met303 . Combining the data for all the ligands , residues belonging to the subunit interface path that were among the most flexible or contacted residues were Arg183 , Val185 , Tyr187 , Trp194 , His197 , Tyr244 , Tyr249 , Ser255 , Arg274 , Arg300 , Met303 , Asn307 , Thr310 , Ala356 , Met402 , Met462 , Ile468 ( see also parts B and C of Figs 5 , S2 and S3 ) . The co-substrate dioxygen and co-product hydrogen peroxide were placed on the si or re side of the isoalloxazine ring of the FAD cofactor and left to explore the VAO dimer . On the re side of FAD , a possible oxygen binding pocket is located as indicated by a bound chloride ion in PDB ID: 1VAO and by a water molecule in the same position in PDB ID: 2VAO [9 , 53] . When placed on the si side of FAD , dioxygen and hydrogen peroxide behaved similarly to the phenolic ligands , but did not leave the enzyme after 200–250 simulation steps , probably due to under sampling . There is a difference in the behaviour of the co-ligands though: they both rarely used the FAD path and hydrogen peroxide had a clear preference for the cap path , almost exiting the protein when migrating through that path . When placing the co-ligands on the re side of FAD , they were able to leave the enzyme after 50 to 250 simulation steps via a path through the interface of the cap and FAD domain . This path will be referred to as re path for the remainder of this text and is illustrated in cyan in Figs 3 and 7 . Contrary to the paths identified for phenolic ligands , this path is visible as an access channel to the active site from the crystal structure of VAO ( see Fig 6B ) . Binding energies for dioxygen and hydrogen peroxide were found to be on a different scale from those of the phenolic ligands , which is to be expected due to the difference in size of these molecules ( computed binding energies are extensive properties ) , for comparison see parts D of Figs 4 , 5 and 7 . However , binding energies were comparable for these co-ligands when placed at the si or re side of FAD . In both cases , variations in binding energies for hydrogen peroxide were larger than for dioxygen , but overall binding energies were more negative for hydrogen peroxide ( Fig 7D ) . This might suggest that hydrogen peroxide more likely gets stuck in local energy minima , which could mean that dioxygen reaches the active site in a faster and more directed manner than hydrogen peroxide leaves it . Surface residues found in the analysis of RMSF and atom contacts were residues Lys43 , Asp44 , Ile46 and Arg482 . Residues of the re path among the most flexible or contacted residues were Tyr51 , Met52 , Thr55 , Pro60 , His61 , Gly103 , Arg104 , Ser106 , Tyr148 , Leu171 Gly174 , Leu411 and Lys475 ( as illustrated by Fig 7 parts B and C ) . Tyr51 and Tyr408 needed to move aside to allow dioxygen and hydrogen peroxide to leave the active site , a behaviour likely determining access to the flavin re side for these molecules . Tyr51 is also positioned close enough to the proposed oxygen binding pocket that it could be involved in FAD re-oxidation . See also Fig 6A for details of the movement of Tyr51 and Tyr408 in our simulations . A movie illustrating all the paths is available in the supplementary data ( S1 Movie ) . We analysed the structural conservation of path residues in three crystal structures of VAO , EUGO and PCMH . Fig 8 shows the obtained structure-based sequence alignment and illustrates the conservation of residues in the different paths through coloured symbols . We also compared the conformations of the most conserved residues involved in the paths in the three aligned structures ( indicated by star symbols in Fig 8 ) . With the exception of Met195 , Gly420 , His466 and Tyr503 the conformations of all conserved residues in the crystal structures were identical . The different conformations of these four residues will be described below . Met195 was found to have different conformations in the crystal structures of VAO , EUGO and PCMH . These conformations correspond well to the movement by Met195 observed in our simulations ( see Fig 6A for an illustration ) . Gly420 is located in a surface loop , which is in close proximity to the cytochrome c subunit in PCMH . This loop adopts a different conformation in PCMH than in VAO and EUGO . This could be in part due to the presence of the cytochrome c subunit in PCMH , which is absent in EUGO and VAO , and in part due to the covalent histidyl linkage to the FAD cofactor in VAO and EUGO but not PCMH . His466 and Tyr503 have already been observed to have two conformations in the high resolution crystal structure of PCMH ( PDB ID: 1DII [8] ) . The conformational freedom of these residues is not observed in the other crystal structures but is in agreement with the movements of His466 and Tyr503 observed in our simulations ( see Fig 6A for an illustration ) . It is noteworthy that the re path identified in VAO leads to where the cytochrome c subunit is located in PCMH . We compared this channel in VAO , EUGO and PCMH and found that it is non-existent in PCMH and narrower in EUGO compared to VAO ( see Fig 6B for an illustration ) . Several residues involved in the paths we identified in VAO are structurally conserved in EUGO and PCMH . If the paths we identified are also present in so far uncharacterised VAO- , EUGO- and PCMH-like enzymes , conservation of path residues should also be observed . This would further support the existence of the paths we have identified . We therefore performed an analysis of sequence conservation . We analysed sequence conservation in MSAVAO , MSAEUGO and MSAPCMH to establish if the paths identified in this study are also conserved in so far uncharacterised VAO- , EUGO- and PCMH-like enzymes . To this purpose , we used MSAVAO , MSAEUGO and MSAPCMH , as described in Materials and Methods . We also defined nine groups of residues to be analysed in detail: the residues in the four paths ( cap , FAD , subunit interface and re path ) , surface or interface residues , a random selection of residues , residues in the cap domain and the entire protein . We found that within these three alignments , all sequences were more similar than dissimilar ( see Fig 9A for all data ) . Residues in the cap path and at the surface were least similar and conserved in all three alignments . Residues in the FAD path and at the subunit interface were most similar and conserved in all three alignments . Note that if similarity or conservation is higher than in the entire protein or the random selection , this indicates higher selective pressure on this group of residues , and lower selective pressure if the opposite is the case . Preservation of the groups of residues analysed in this manner should not be used as an indication of similarity or conservation within all three sets of sequences ( Fig 9A ) . At one position in MSAVAO , a residue can be conserved or similar , but a different residue can be conserved or similar in the MSAEUGO or MSAPCMH . To be able to determine overall preservation of residues in all three sets of sequences ( MSAall ) , we continued our analysis as described in Materials and Methods . The results of this analysis are summarised in Fig 9B . We found that when analysing the full length MSAall , 29% of all positions were more than 90% conserved , and 25% of all positions were similar in VAO and EUGO , but varied in PCMH ( were PCMH-unlike ) . Residues in the cap path were most dissimilar , and also the most PCMH-unlike . Residues in the subunit interface path and at the subunit interface were most similar and residues in the FAD path and at the subunit interface were most conserved . It is noteworthy that all dissimilar residues in the re path are only dissimilar due to PCMH-like sequences . It is conspicuous that the percentages of similarity and conservation in the two analyses differ significantly due to the preservation of different residues in VAO- , EUGO- and PCMH-like sequences . This is reflected in the percentages of dissimilarity and PCMH-unlikeness . Note the difference in preservation between residues in the cap domain and the cap path , which indicates that there is less selective pressure ( or selective pressure for different residues ) on residues in the cap path and the entire cap domain .
We have identified three migration paths in VAO for phenolic ligands ( see Fig 10 for an overview ) . We do not observe any preference for a specific path for any of the three phenolic ligands analysed ( p-creosol , vanillyl alcohol and vanillin ) . Two of these migration paths , the cap and FAD paths , are less likely paths for these ligands . They require large movements of amino acid side chains ( namely Arg312 and Arg398 or Trp89 and Lys545 , respectively ) to allow the phenolic ligand to pass , which make them more difficult passages to the active site . This can be seen in the simulations , where the ligands often got stuck next to these residues . It is also evident from the high number of atom contacts ligands experienced with these residues . VAO also accepts phenolic substrates with larger hydrophobic side chains than the ligands studied here [9] . These larger ligands would be prone to have even more difficulties passing the bottlenecks in the cap and FAD path . Therefore , and because it represents the only entry and exit path identified , we argue that the VAO subunit interface path is the most probable route for phenolic ligands to enter or leave the active site . In the subunit interface path , two residues are observed to change conformation to allow ligands to leave the active site . These two residues , His466 and Tyr503 , do not present obstacles for the migration of the ligands but are in a closed conformation once a ligand is bound . Our results suggest a dual role for Tyr503 , which is known to be involved in substrate deprotonation [16] , and which we can show here to be also involved in substrate migration . His466 and Tyr503 can be regarded as concierges , limiting access to the active site . Interestingly , in the crystal structure of PCMH ( PDB ID: 1DII [8] ) , two possible conformations are observed for His436 and Tyr473 , which correspond to His466 and Tyr503 in VAO ( see Fig 6A for an illustration ) . We would also like to highlight Met195 as the most flexible of the three portal residues to complete the list of flexible and probably functionally important residues in this path . It is of interest to keep in mind that this path shows that the two active sites of the VAO dimer are connected via this portal . We were also able to demonstrate that protonated phenolic substrates enter the enzyme but deprotonated ones do not . Energy profiles of exit simulations also indicate more favourable binding of the protonated substrates and indifference towards the protonation state of the product . We have also identified an additional migration path , the re path for the co-ligands dioxygen and hydrogen peroxide . This path leads co-ligands to and from the re side of FAD . The co-ligands were also able to migrate through the paths identified for the phenolic ligands . The re path is the shortest connection to the solvent , compared to the paths on the si side of FAD . On the re side of FAD , a chloride ion present in the VAO crystal structure ( PDB ID: 1VAO ) is indicative of a possible oxygen binding site [53 , 54] . There is no evidence from binding energies calculated in our simulations that the si path is preferred . Our simulations were performed with oxidised FAD , and while dioxygen reacts with reduced FAD , there are no indications of significant conformational changes in VAO crystal structures linked to the changed oxidation state of FAD ( PDB ID: 1VAO vs PDB ID: 1AHU ) . All these observations make the re path the most probable migration path for dioxygen and hydrogen peroxide . In this path , it is possible that Tyr51 is involved in directing dioxygen to the FAD , as it is one of the most contacted and flexible residues in this path . It is also located in the probable oxygen-binding pocket on the re side of FAD , making it a candidate for assisting in reduced flavin oxidation . From our results of the co-ligand simulations on the si side of FAD , we can see that co-ligands tend to migrate into the cap path . Crystallographic data ( PDB ID: 1VAO ) show that within the cap path , water molecules are located close to Arg312 and Arg398 as well as at the paths exit/entry on the surface of VAO . Water is able to access the active site and to participate in the reaction on the si side of FAD , reacting with the quinone methide of certain substrates [20] . The cap path is the only path at the si side of FAD directly connected to the solvent . In the case of binding of alkylphenols with hydrophobic aliphatic side chains , already bound water molecules would likely be expulsed from the active site via the cap path . This would explain why these substrates are dehydrogenated and not hydroxylated by VAO [55] . We will now discuss our findings in VAO in the context of the close bacterial relatives EUGO and PCMH . All three enzymes require their substrates to be para-substituted phenols [6–10] . In PCMH and EUGO , a possible route for substrate access to the active site via the subunit interface has been suggested purely on crystallographic evidence [8 , 12] . In this route , three residues ( Glu177 , Met180 and Asp434 in PCMH ) have been proposed to be involved . These residues are Asp192 , Met195 and Glu464 in VAO , with Asp192 being in the position of Glu177 and Glu464 in the position of Asp434 . EUGO has recently had its crystal structure solved [12] . In EUGO , all the above-mentioned residues involved in the subunit interface path of VAO are also present ( Fig 6 ) . We could show that residues at the subunit interface and in the subunit interface path are the most similar in all three of these closely related members of the VAO/PCMH family , an indication that increased selective pressure is in place for these residues . Additionally , the combination of this increased selective pressure and the presence of the entrance/exit migration path at the subunit interface ( and absence of evidence of allosteric regulation in VAO ) could indicate that dimerisation of VAO is functionally linked to ligand migration . Residues in the cap path are most diverse in all three enzymes . This lends support to the proposal that all three enzymes prefer the subunit interface path for ligand migration . The small amount of PCMH-like sequences remaining in the analysis after applying our selection criteria makes interpretation of the data with respect to PCMH less reliable . VAO and EUGO re-oxidise their FAD cofactor using molecular oxygen , while PCMH requires a cytochrome c subunit for this . This functional difference is reflected in the preservation of the re path in these enzymes ( see specifically the PCMH-unlike category in Fig 9B ) . Interestingly , dissimilarity in the residues of this path is only due to PCMH-like sequences and not due to differences between the two oxidases . It is noteworthy that the re path we identify here is located in a visible channel in VAO and EUGO . In PCMH , the channel is absent and the cytochrome c subunit in PCMH is located where dioxygen would exit the protein ( see Fig 6B ) . Evolution of the re path in these oxidases may thus have coincided with loss of the cytochrome c subunit . The strong conservation of residues in the FAD path could indicate that these residues are essential to the function of all three enzymes . It has been established that maturation of the VAO holoenzyme involves the initial non-covalent binding of the FAD cofactor to the already folded dimeric apoenzyme [11 , 56] . It could be that FAD binds to the apoenzyme via this path . A possible scenario is that FAD enters the enzyme with the isoalloxazine moiety first and then binds to the enzyme in the correct orientation already . The alternative scenario , in which FAD binds with its ADP moiety first , would require reorientation of the FAD inside the enzyme to fit the isoalloxazine moiety into the active site . From the crystal structure , it is not evident how this would happen as the free space around bound FAD is rather limited . Initial binding of FAD to the enzyme surface and migration of FAD into its binding pocket is likely guided by pi stacking interactions with the three tryptophans located in the FAD path ( Trp98 , Trp549 and Trp558 ) . Very little is known about substrate access to the active site for the other members of this flavoprotein family . Dioxygen migration has been studied in alditol oxidase , berberine bridge enzyme ( also called ( S ) -reticuline oxidase ) and L-galactono-1 , 4-lactone dehydrogenase . Ala113 in L-galactono-1 , 4-lactone dehydrogenase was identified as a gatekeeper residue that prevents the enzyme from functioning as an oxidase [57] . Berberine bridge enzyme also contains such a gatekeeper residue . The variant G164A showed 800-fold decreased oxygen reactivity [58] . In alditol oxidase funnel shaped paths have been identified through molecular dynamics simulations , leading dioxygen to the “gatekeeper” Ala105 in alditol oxidase and from there to N1/N3 of the flavin ring [23] . This path leads to the re side of FAD , as does the re path we identified in this study . Our path , however , leads to its destination via the dimethylbenzene part of the flavin ring . We have therefore identified for the first time a substrate migration channel , the subunit interface path , and a novel co-substrate migration path . In summary , we have carried out an exhaustive computational analysis of ligand migration pathways in VAO ( see Fig 10 for an overview ) . We have identified a gated entry and exit path at the subunit interface of VAO for small phenolic ligands . The residues forming a portal in this path ( Met192 , Met195 and Glu464 in VAO ) are conserved in PCMH and EUGO . Two residues , His466 and Tyr503 in VAO , act as concierges , obstructing the path to the active site for phenolic substrates once substrate is bound , as has been postulated before for PCMH [8] . An additional , different entry and exit path for dioxygen and hydrogen peroxide at the interface of the two domains of VAO has been identified ( re path ) . Tyr51 in this path could assist reduced flavin oxidation in VAO . The only path identified directly connecting the si side of FAD to the solvent ( the cap path ) could be facilitating water access to the active site . Overall , our study illustrates the ability of ligand simulation techniques to advance the mechanistic understanding of enzyme function .
|
Enzymes are bionanomachines , which speed up chemical reactions in organisms . To understand how they achieve that , we need to study their mechanisms . Computational enzymology can show us what happens in the enzyme’s active site during a reaction . But molecules need first to reach the active site before a reaction can start . The process of substrate entry and product exit to the active site is often neglected when studying enzymes . However , these two events are of fundamental importance to the proper functioning of any enzyme . We are interested in these dynamic processes to complete our understanding of the mode of action of enzymes . In our work , we have studied substrate and product migration in vanillyl alcohol oxidase . This enzyme can produce the flavour vanillin and enantiopure alcohols , but also catalyses other reactions . The named products are of interest to the flavour- and fine-chemical industries .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"crystal",
"structure",
"chemical",
"compounds",
"oxides",
"split-decomposition",
"method",
"condensed",
"matter",
"physics",
"enzymology",
"organic",
"compounds",
"multiple",
"alignment",
"calculation",
"flavin",
"crystallography",
"hydrogen",
"peroxide",
"alcohols",
"enzyme",
"chemistry",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"solid",
"state",
"physics",
"sequence",
"alignment",
"bioinformatics",
"proteins",
"chemistry",
"physics",
"biochemistry",
"organic",
"chemistry",
"computational",
"techniques",
"post-translational",
"modification",
"database",
"and",
"informatics",
"methods",
"biology",
"and",
"life",
"sciences",
"phenols",
"physical",
"sciences",
"cofactors",
"(biochemistry)",
"peroxides"
] |
2017
|
The ins and outs of vanillyl alcohol oxidase: Identification of ligand migration paths
|
Mutations causing replication stress can lead to genomic instability ( GIN ) . In vitro studies have shown that drastic depletion of the MCM2-7 DNA replication licensing factors , which form the replicative helicase , can cause GIN and cell proliferation defects that are exacerbated under conditions of replication stress . To explore the effects of incrementally attenuated replication licensing in whole animals , we generated and analyzed the phenotypes of mice that were hemizygous for Mcm2 , 3 , 4 , 6 , and 7 null alleles , combinations thereof , and also in conjunction with the hypomorphic Mcm4Chaos3 cancer susceptibility allele . Mcm4Chaos3/Chaos3 embryonic fibroblasts have ∼40% reduction in all MCM proteins , coincident with reduced Mcm2-7 mRNA . Further genetic reductions of Mcm2 , 6 , or 7 in this background caused various phenotypes including synthetic lethality , growth retardation , decreased cellular proliferation , GIN , and early onset cancer . Remarkably , heterozygosity for Mcm3 rescued many of these defects . Consistent with a role in MCM nuclear export possessed by the yeast Mcm3 ortholog , the phenotypic rescues correlated with increased chromatin-bound MCMs , and also higher levels of nuclear MCM2 during S phase . The genetic , molecular and phenotypic data demonstrate that relatively minor quantitative alterations of MCM expression , homeostasis or subcellular distribution can have diverse and serious consequences upon development and confer cancer susceptibility . The results support the notion that the normally high levels of MCMs in cells are needed not only for activating the basal set of replication origins , but also “backup” origins that are recruited in times of replication stress to ensure complete replication of the genome .
In late mitosis to early G1 phase of the cell cycle , DNA replication origins are selected and bound by the hexameric origin recognition complex ( ORC; [1] ) . ORC then recruits the initiation factors CDC6 and CDT1 , which are required for loading MCM2-7 , thereby forming the “pre-replicative complex” ( pre-RC ) . The formation of pre-RCs is termed origin “licensing” and this gives origins competency to initiate a single round of DNA synthesis before entering S phase . MCM2-7 is a hexamer of six distinct but structurally-related minichromosome maintenance ( MCM ) proteins ( reviewed in [2]–[5] ) . In vivo and in vitro evidence indicates that the MCM2-7 complex is the replicative helicase [6]–[8] . MCM2-7 proteins are abundant in proliferating cells [9] , and are bound to chromatin in amounts exceeding that which is present at active replication origins or required for complete DNA replication [10]–[14] . Although these and other studies showed that drastic decreases in MCMs are tolerated by dividing cells , there are certain deleterious consequences . In Xenopus extracts and mammalian cells , excess chromatin-bound MCM2-7 complexes occupy dormant or “backup” origins that are activated under conditions of replication stress , compensating for stalled or disrupted primary replication forks [11] , [15]–[16] . The depletion of these backup licensed origins was associated with elevated chromosomal instability and susceptibility to replication stress , factors that might predispose to cancer . In previous work , Shima et al found that a hypomorphic allele of mouse Mcm4 ( Mcm4Chaos3 ) caused high levels of GIN and extreme mammary cancer susceptibility in the C3HeB/FeJ background [17] . This provided the first concrete evidence that endogenous mutations in replication licensing machinery may have a causative role in cancer development . The ethylnitrosourea ( ENU ) -induced Mcm4Chaos3 point mutation changed PHE to ILE at residue 345 ( Phe345Ile ) . This amino acid is conserved across diverse eukaryotes and is important for interaction with other MCMs [18] . Budding yeast engineered to bear the orthologous mutation exhibit DNA replication defects and GIN [17] , [19] . Surprisingly , MEFs from Mcm4Chaos3 mice not only had reduced levels of MCM4 , but also MCM7 [17] , suggesting that the point mutation might destabilize the MCM2-7 complex . Subsequently , it was reported that mice containing 1/3 the normal level of MCM2 succumbed to lymphomas at a very young age , and had diverse stem cell proliferation defects [20] . These mice also had 27% reduced levels of MCM7 protein , and their cells exhibited decreased replication origin usage when under replication stress ( treatment with hydroxyurea ) conditions [21] . These studies imply that relatively modest decreases in any of the MCMs may be sufficient to cause cancer susceptibility , developmental defects , and GIN [20] . Here , we report that genetically-induced reductions of MCM levels in mice , achieved by breeding combinations of MCM2-7 alleles , caused several health-related defects including increased embryonic lethality , GIN , cancer susceptibility , growth retardation , defective cell proliferation , and hematopoiesis defects . Remarkably , genetic reduction of MCM3 , which mediates nuclear export of excess MCM2-7 complexes in yeast [22] , rescued many of these defects , presumably attributable to observed increases in chromatin-bound MCM levels . These data suggest that relatively minor misregulation or destabilization of MCM homeostasis can have serious consequences for health , viability and cancer susceptibility of animals .
To extend previous findings that Mcm4Chaos3Chaos3 cells exhibited decreases in MCM4 and MCM7 protein , and to determine if the decreased levels were differentially compartmentalized in the cell , we quantified soluble and chromatin-bound MCM2-7 levels in mouse embryonic fibroblasts ( MEFs ) by Western blot analysis . As shown in Figure 1A , all MCMs were decreased in both compartments by at least 40% compared to WT cells . Because Mcm4Chaos3/Chaos3 MEF cultures have slightly decreased proliferation and G2/M delay ( Figure 1A and [17] ) , it is possible that the lower MCM levels in mutant MEFs are entirely attributable to growth defects . To test this , we assessed the levels of nuclear MCM2 in S-phase cells by flow cytometry ( Figure 1B ) . Although MCM2 levels in WT and Mcm4Chaos3/Chaos3 G1 nuclei were essentially the same ( P = . 65; t-test ) , mutant cells transitioned from G1 to S with 40% less nuclear MCM2 content than in WT ( P< . 02; t-test ) . The levels of nuclear MCM2 in WT decreased through S phase more sharply than in mutants , which transitioned to G2 with only ∼23% less than controls ( Figure 1B ) . This differential decline is apparent in the flow plots , where WT cells exhibit a greater downward slope in the S compartment ( Figure 1B ) . The decreases in MCM2 from early to late S were 51% in WT and 38% in mutants . The MCM2 intra-S modulation phenomenon is also addressed in subsequent experiments . The marked differences in nuclear MCM2 concentration between actively proliferating ( S-phase ) WT and mutant cells indicates that a biochemical or regulatory basis , rather than a population skewing , underlies the differences in protein levels . Another possible explanation for the coordinated decrease in MCMs is that the mutant MCM4Chaos3 protein destabilizes the MCM2-7 hexamer and causes subsequent degradation of uncomplexed MCMs . Other groups reported that knockdown of Mcm2 , Mcm3 , or Mcm5 in human cells decreased the amount of other chromatin-bound MCMs [15]–[16] , leading to a similar proposition that the cause was MCM2-7 hexamer destabilization [16] . If true , then we would expect mRNA levels to be unchanged in mutant cells . To test this , we performed quantitative RT-PCR ( qRT-PCR ) analysis of Mcm2-7 , and several control housekeeping genes in Mcm4Chaos3/Chaos3 MEFs . Analysis of 5 littermate pairs of primary MEF cultures revealed that transcript levels for each of these genes in mutant cells was 51–65% of WT , similar to the protein decreases ( Figure 1C ) . Levels of mRNA in the 7 housekeeping genes analyzed were not altered significantly ( Figure 1C , right panel ) . This data suggest that either reduced MCM4 levels per se , or defects resulting from the Mcm4Chaos3 allele , cause a decrease in the levels of all Mcm mRNAs . Interestingly , the mRNA reduction appears to occur post-transcriptionally , a phenomenon that is currently under investigation ( Chuang and Schimenti , unpublished observations ) . The Mcm4Chaos3 allele was identified in a forward genetic screen for mutations causing elevated micronuclei ( MN ) in red blood cells , an indicator of GIN [17] . While the altered MCM4Chaos3 protein may cause DNA replication errors as does a yeast allele engineered to contain the same amino acid change [19] , it is also possible that the decrease in overall MCM levels in Mcm4Chaos3 mutants contributes to , or is primarily responsible for , elevated S-phase DNA damage and GIN as is seen in various cell culture models ( see Introduction ) . To test this possibility , we generated mice from ES cells bearing gene trap insertions in Mcm2 , Mcm3 , Mcm6 , and Mcm7 ( Figure 2A; alleles are designated as Mcm#Gt ) . These gene traps are designed to disrupt gene expression by fusing the 5′ end of the endogenous mRNA ( via use of a splice acceptor ) to a vector-encoded reporter , resulting in a fusion protein lacking the C-terminal portion of the endogenous ( MCM ) protein . As with a previously-reported Mcm4 gene trap [17] , each of these alleles proved to be recessive embryonic lethal ( Figure S1 ) . Furthermore , each allele appeared to be a null , since mRNA levels in heterozygous MEF cultures were ∼50% lower than WT controls ( Figure 2B ) . To determine if heterozygosity for various Mcms caused pan-decreases in Mcm mRNA levels as does homozygosity for Mcm4Chaos3 , mRNA levels for each of the Mcm2-7 genes were also quantified . Whereas Mcm2Gt/+ cells did show ∼20% decreases in the other Mcms , the Mcm3 , Mcm4 , Mcm6 and Mcm7 gene trap alleles did not ( Figure 2B ) . Thus , it appears that the marked Mcm pan-decreases in Mcm4Chaos3/Chaos3 cells are not due to decreased Mcm4 RNA per se , but rather a response to replication defects cause by the mutant protein . Notably , the pan Mcm2-7 downregulation in Mcm2Gt/+ cells is consistent with the observation that MCM7 is decreased in Mcm2IRES-CreERT2/IRES-CreERT2 mice , although mRNA levels were not evaluated in that study [20] . After breeding the gene trap alleles into the C3HeB/FeJ genetic background for at least 2 generations ( Mcm4Chaos3/Chaos3 females get mammary tumors in this background ) , blood MN levels were measured . Heterozygosity for each allele caused an increase in the fraction of cells with MN ( Figure 2C ) . Compound heterozygosity further increased MN on average , as did heterozygosity for 3 or more gene traps ( Figure 2C ) , indicating that genetically-based decreases in any of the MCMs precipitate GIN . As outlined above , previous studies showed that reductions of particular MCMs in cells or mice reduces the levels of other MCMs , causing GIN , cancer , and developmental defects . However , the reduction in MCM levels required to precipitate these consequences , and whether there is a threshold effect , is unclear . To explore the consequences of incremental MCM reductions on viability and cancer in mice , we crossed the Mcm4Chaos3 and gene trap alleles into the same genome . In the case of Mcm2 , there was a striking and highly significant shortfall of Mcm4Chaos3/Chaos3 Mcm2Gt/+ offspring at birth ( Figure 3A; Figure S2 ) . Heterozygosity for Mcm2Gt itself was not haploinsufficient , as indicated by Mendelian transmission of Mcm2Gt in crosses of heterozygotes to WT ( 119/250; χ2 = 0 . 448 ) . These results demonstrate that there is a synthetic lethal interaction between Mcm4Chaos3 and Mcm2Gt that is related to MCM2 levels . Additionally , the surviving Mcm4Chaos3/Chaos3 Mcm2Gt/+ offspring were severely growth retarded; males weighed ∼50% less than Mcm4Chaos3/Chaos3 siblings ( Figure 3B; this genotype causes disproportionate female lethality ) . Another indication of a quantitative MCM threshold effect is that C3H-Mcm4Chaos3/Chaos3 mice are developmentally normal , but Mcm4Chaos3/Gt animals die in utero or neonatally ( Figure 3A ) [23] . The synthetic interaction between Mcm4Chaos3 and Mcm2Gt might be specific , or it may reflect a general consequence of reduced replication licensing ( and consequent elevated replication stress ) . We therefore tested whether hemizygosity for Mcm3 , Mcm6 or Mcm7 would also cause synthetic phenotypes in the Mcm4Chaos3/Chaos3 background . The Mcm4Chaos3/Chaos3 Mcm6Gt/+ genotype caused highly penetrant embryonic lethality; only 10% of the expected number of such animals survived to birth ( Figure 3A; Figure S2 ) . The Mcm4Chaos3/Chaos3 Mcm7Gt/+ genotype caused both embryonic and postnatal lethality . The number of liveborns was ∼50% of the expected value , and only 8% of those ( 5/62 ) survived to weaning ( Figure 3A; Figure S2 ) . Additionally , as with Mcm2 , hemizygosity for Mcm6Gt and Mcm7Gt in the Mcm4Chaos3/Chaos3 background caused growth retardation ( Figure 3B ) . The decrease in male weight was ∼20% and ∼80% respectively , compared to Mcm4Chaos3/Chaos3 siblings at the oldest age measured ( Mcm4Chaos3/Chaos3 Mcm7Gt/+ animals died before wean , so the oldest weights were taken at 10 dpp ) . In contrast to the synthetic phenotypes with Mcm2 , 4 , 6 and 7 , there was no significant decrease in viability ( Figure 3A ) or weight ( not shown ) in Mcm4Chaos3/Chaos3 Mcm3Gt/+ mice . This seeming inconsistency is addressed in the following section . As mentioned earlier , mice with ∼35% of WT MCM2 protein , but not 62% , showed early latency ( 10–12 week ) lymphoma susceptibility [20] . To identify if there is a critical MCM threshold for cancer susceptibility , we aged a cohort of Mcm2Gt/+ mice , representing approximately intermediate MCM2 levels . As shown in Figure 4A , these animals did not show a dramatic cancer-related mortality in the first 12 months of life . However , we did find that ∼3/4 of these animals had tumors at death or necropsy by 18 months of age ( data not shown ) . These combined data are suggestive of a potential gradient of susceptibility , but that there is a critical minimum threshold of MCM levels , between ∼35 and 50% in the case of MCM2 , required to avoid early cancer and other developmental defects . To further resolve this phenomenon , surviving Mcm4Chaos3/Chaos3 Mcm2Gt/+ mice were aged and monitored . They began dying at 2 months of age , and all were dead ( or sacrificed when they appeared moribund ) by 7 months ( Figure 4A ) . Gross necropsy and histopathological analyses revealed or suggested lymphomas/leukemias in 20 of these animals ( summarized in Table S1 with histological examples in Figure S3; detailed histopathology analysis of a T cell leukemic lymphoma is presented in Figure 4B ) . Six of these had chest tumors that were likely thymic lymphomas . The cause of death for the remaining 7 animals was undetermined . Consistent with previous studies [17] , most Mcm4Chaos3/Chaos3 mice hadn't yet succumbed from tumors or other causes by 12 months of age . Additional animals of these genotypes are incorporated in Figure 6 , but histopathological analyses weren't conducted . These data show clearly that removing a half dose of MCM2 from Mcm4Chaos3/Chaos3 cells is sufficient to produce greatly elevated cancer predisposition to the already-underrepresented survivors at wean . Mcm4Chaos3/Chaos3 Mcm2Gt/+ MEFs had 45% the amount of Mcm2 mRNA as Mcm4Chaos3/Chaos3 cells ( Figure 7C ) , which already had a 38% reduction compared to WT ( Figure 1 ) . Thus , Mcm2 RNA was reduced to ∼17% of WT . To determine if elevated GIN might be responsible for the cancer susceptibility phenotype , we measured erythrocyte MN . Whereas the percentage of micronucleated RBCs in Mcm4Chaos3/Chaos3 mice was 4 . 18±0 . 26 ( mean±SEM , N = 12 ) , Mcm4Chaos3/Chaos3 Mcm2Gt/+ mice averaged 5 . 85±0 . 47 ( N = 16 ) , indicating a synergistic increase ( P<0 . 01 ) . Overall , the data support the notion that in whole animals , reduction of MCMs to under 50% of WT causes severe developmental and physiological problems . The data reported here and elsewhere [17] , [20] support a model where phenotypic severity is proportionally related to MCM concentrations . However , our genetic experiments uncovered one notable exception: hemizygosity for Mcm3 did not cause any severe haploinsufficiency phenotypes ( increased lethality and decreased weight ) as did Mcm2/6/7 in the Mcm4Chaos3/Chaos3 background , or Mcm4Gt in trans to Mcm4Chaos3 ( Figure 3A; Figure S2 ) . Since extreme reductions of MCM3 in cultured human cells caused GIN and cell cycle arrest [16] , the absence of synthetic effects with McmChaos3 led us to hypothesize that either mice are more tolerant to lower levels of this particular MCM , or that MCM3 is present in a stoichiometric excess compared to the other MCMs , at least in a subset of cell types . To explore these issues we performed additional phenotype analyses , and also sought to uncover potential effects of MCM3 reduction by reducing other MCMs simultaneously . Strikingly , rather than exacerbating the synthetic lethality in Mcm4Chaos3/Chaos3 Mcm2Gt/+ mice , Mcm3Gt heterozygosity significantly rescued their viability to 72 . 5% from 29 . 7% ( Figure 5A and Fig S3 ) . Not only was viability rescued , but also growth ( weight ) of Mcm4Chaos3/Chaos3 Mcm2Gt/+ Mcm3Gt/+ survivors compared to Mcm4Chaos3/Chaos3 Mcm2Gt/+ animals produced from the same matings ( Figure 5B ) . Mcm3 hemizygosity also significantly rescued the near 100% lethality of Mcm4Chaos3/Gt animals ( nearly 6 fold increased viability ) , and doubled the viability of Mcm4Chaos3/Chaos3 Mcm6Gt/+ mice ( Figure 5A; Figure S3 ) . Rescue of Mcm4Chaos3/Chaos3 Mcm7Gt/+ was not observed ( not shown ) . The rescue of the reduced growth phenotype by Mcm3 hemizygosity led us to evaluate the proliferation of compound mutant cells . Whereas Mcm4Chaos3/Chaos3 and Mcm4Chaos3/Chaos3 Mcm3Gt/+ primary MEFs proliferated at identical rates , Mcm4Chaos3/Chaos3 Mcm2Gt/+ MEFs showed a severe growth defect beginning ∼5 days in culture ( Figure 5C ) . As with whole animals , MEF growth was partially but significantly rescued by Mcm3 hemizygosity . Since the Mcm4Chaos3 and Mcm2Gt alleles causes elevated GIN ( micronuclei in RBCs ) , we considered the possibility that the Mcm3 rescue effect might be related to an attentuation of GIN . Accordingly , we measured MN levels in Mcm4Chaos3/Chaos3 mice with different combinations of other Mcm mutations . As shown in Figure 5D , hemizygosity for Mcm2 and Mcm7 caused a significant elevation in MN levels , unlike Mcm3 . However , the increased MN in Mcm4Chaos3/Chaos3 Mcm2Gt/+ was not rescued by Mcm3 hemizygosity . This suggests that the synthetic lethality and mouse/cell growth defects are not related to GIN per se . However , in the course of measuring MN in enucleated peripheral blood cells , we noticed that the ratio of CD71+ cells was significantly higher in both Mcm4Chaos3/Chaos3 Mcm2Gt/+ and Mcm4Chaos3/Chaos3 Mcm7Gt/+ mice ( 3 . 3 and 6 . 2 fold , respectively; Figure 5E ) . This increase in the ratio of reticulocytes ( erythrocyte precursors; immature RBCs ) to total RBCs is characteristic of anemia . Hemizygosity for Mcm3 , which alone had no effect on CD71 ratios of Chaos3 mice , corrected completely this abnormal phenotype in Mcm4Chaos3/Chaos3 Mcm2Gt/+animals ( Figure 5E ) . Because MCM2-depleted mice were reported to have stem cell defects [20] , and Mcm4Chaos3/Chaos3 Mcm#Gt/+ mice had clear developmental abnormalities , we examined the efficiency of reprogramming mutant MEFs into induced pluripotent stem cells ( iPS ) . The efficiency was quantified using either : 1 ) iPS-like colony formation , or 2 ) cells counts of SSEA1 and LIN28 positive cells by flow cytometry . Both gave similar results . Mcm4Chaos3/Chaos3 Mcm2Gt/+ cells were severely compromised in the ability to form iPS cells compared to Mcm4Chaos3/Chaos3 ( ∼200 fold less efficient; Figure 5F ) . However , additionally reducing Mcm3 by 50% increased iPS formation from both Mcm4Chaos3/Chaos3and Mcm4Chaos3/Chaos3 Mcm2Gt/+ MEFs by ∼2 . 5 and 10 fold , respectively . Finally , we found that reduced MCM3 levels could rescue the cancer susceptibility of two different Chaos3 models . As shown earlier ( Figure 4 ) , Mcm4Chaos3/Chaos3 Mcm2Gt/+ mice were highly cancer-prone with an average latency of <4 months . When a dose of Mcm3 was removed from mice of this genotype , lifespan was extended dramatically in both sexes as a consequence of delayed cancer onset , and the cancer spectrum shifted from lymphoma/thymoma towards mammary tumors ( Figure 6A ) . Additionally , hemizygosity of Mcm3 delayed ( or eliminated ) the onset of mammary tumorigenesis in Mcm4Chaos3/Chaos3 females by ∼4 or more months ( Figure 6B ) . However , although Mcm3 hemizygosity rescued viability of Mcm4Chaos3/Gt mice ( Figure 5A ) , these animals were cancer prone with a shorter latency ( by ∼6 months ) and different spectrum ( primarily lymphomas ) than Mcm4Chaos3 homozygotes . We considered two possibilities to explain the surprising phenotypic rescues of reduced MCM genotypes ( Mcm4Chaos3/Chaos3 ; Mcm4Chaos3/Chaos3 Mcm2/6Gt/+ ; Mcm4Chaos3/Gt ) by additional MCM3 reduction ( Mcm3Gt/+ ) . One is that the phenotypes are related to altered stoichiometry of MCM monomers , and that disproportionally high amounts of MCM3 relative to MCM4 and MCM2/6/7 have a dominant negative effect . However , as demonstrated above , levels of MCM3 are proportionally reduced in Mcm4Chaos3/Chaos3 cells ( Figure 1 ) . The second possibility is that decreased levels of MCM3 leads to a favorable change in the amounts or subcellular localization of MCMs . Various experiments have indicated that MCM2-7 hexamers or subcomplexes must be assembled in the cytoplasm before nuclear import in yeast [4] , and in mice , nuclear import appears to require MCM2 and MCM3 [24] . MCMs shuttle between the nucleus and cytoplasm during the cell cycle in S . cerevisiae . Although in most other organisms MCMs are reported to be predominantly and constitutively nuclear localized throughout the cell cycle , dynamic redistribution between the nucleus and cytoplasm has been observed in hormonally-treated mouse uterine cells [25] . In budding yeast , nuclear export is dependent upon Mcm3 , which has a nuclear export signal ( NES ) that is recognized by Cdc28 to promote export of MCM2-7 [22] . Analysis of mouse and human MCM3 using NES prediction software ( www . cbs . dtu . dk/services/NetNES/ ) [26] revealed the presence of homologously-positioned , leucine-rich potential NESs ( Figure 7A ) . Therefore , we hypothesized that the rescue of phenotypes by Mcm3 hemizygosity is due to decreased MCM protein export from the nucleus , or alternatively , increased nuclear import or stabilization that allows greater access of all MCMs for licensing chromatin . To explore this hypothesis , we performed Western blot analysis of MCM levels in Mcm4Chaos3/Chaos3 MEFs with or without the Mcm3Gt and/or Mcm2Gt alleles , and examined the effects of Mcm3 dosage on the levels of nuclear and chromatin-bound MCM2 and MCM4 . The results are presented in Figure 7B . In all cases , the genetic reductions of Mcm2 and Mcm3 led to corresponding decreases in the cognate mRNA levels ( Figure 7C ) , with only minor additional decreases of other MCM mRNAs ( beyond that already caused by homozygosity for Mcm4Chaos3 ) occuring in the context of Mcm2 hemizygosity ( similar to Mcm2Gt/+ MEFs in Figure 2B ) . The overall levels of total , nuclear , and chromatin-bound MCM2 and MCM4 were unaffected by hemizygosity of Mcm3 in Mcm4Chaos3/Chaos3 cells ( Figure 7B ) . When Mcm2 levels were genetically reduced by half , a condition causing the severe phenotypic effects described earlier , this caused a marked decrease in the level of chromatin-bound MCM3 and MCM4 ( in addition to MCM2 itself ) , although total and nuclear MCM3/4 levels were affected to a lower degree or not at all . Strikingly , the decreased levels of chromatin-bound MCM2/3/4 in Mcm4Chaos3/Chaos3 Mcm2Gt/+ MEFs were reversed by Mcm3 heterozygosity , but levels of total MCM2 and MCM4 were not restored . The increase of chromatin-bound MCMs occured despite the presence of less MCM3 , suggesting that MCM3 is present at levels in excess of that needed to bind chromatin , presumably for pre-RC formation in the context of the MCM2-7 hexamer . In conclusion , a 50% reduction in total MCM3 increases MCM2/4 loading onto chromatin when MCM2 is otherwise limiting , and this rescue is associated with amelioration of several phenotypes . We found that elevation of nuclear MCMs in the Mcm3Gt/+ MEFs was often ( as shown in Figure 7B ) , but not consistently elevated across samples by Western analysis ( not shown ) . Therefore , we quantified MCM2 during the cell cycle by flow cytometric analysis of nuclei from 7 replicate MEF cultures . Similar to WT MEFs ( examples in Figure 1B ) , NIH3T3 cells showed a decrease of nuclear MCM2 during S phase progression ( Figure 7D , left panel ) . However , all genotypes with in the Mcm4Chaos3/Chaos3 background had a reduced decline . Thus , for comparative quantitation across genotypes , we compared the levels of MCM2 levels at the beginning of G1 vs . that in S phase ( regions used for these calculations are indicated in the left panel ) , using the calculation described in the Figure 7 legend . The data are graphed in the right panel . The data revealed that regardless of genotype , the difference in average amounts of nuclear MCM2 at the beginning and end of G1 ( ΔG1 ) did not vary . Compared to Mcm4Chaos3/Chaos3 , cells lacking 1 dose of Mcm2 had relatively lower levels of S phase MCM2 ( ΔS ) compared to early G1 . Additional removal of an Mcm3 dose partially rescued the ΔS value , indicating that these cells had ∼16% more nuclear MCM2 in S phase compared to Mcm4Chaos3/Chaos3 cells hemizygous for Mcm2 alone , despite overall reduced MCM2 levels in the cell ( Figure 7B , left panel ) .
MCM2-7 proteins exist abundantly in proliferating cells and are bound to chromatin in amounts exceeding that required to license all replication origins that initiate DNA synthesis [9]–[12] , [14] . The role of excess chromatin-bound MCM2-7 has been a mystery referred to as the “MCM paradox” [27] , perpetuated by observations that drastic MCM reductions in certain systems can be compatible with normal DNA replication or cell proliferation [13] , [28]–[30] . However , these circumstances are not universal , and reductions are not entirely without consequences . Early studies showed that a reduction in MCMs resulted in decreased usage of certain ARSs [12] and conferred genome instability [31] in yeasts . In cell culture systems , depletion of certain MCMs have been found to cause cell cycle defects , checkpoint abberations and GIN [13] , [16]–[17] , [29] , [32] . Recent work has shed light on aspects of the MCM paradox . Using Xenopus egg extracts attenuated for licensing by addition of geminin ( an inhibitor of CDT1 , which is required for MCM loading onto origins ) , one study proposed that excess chromatin-bound MCM2-7 complexes license “dormant” origins that can be activated to rescue stalled or damaged replication forks , a situation that can become important under conditions of replication stress [11] . Similar results were subsequently reported for human cells depleted of MCMs by siRNA [15]–[16] , and for replication stressed MCM2-deficient MEFs [21] . Our finding that nuclear MCM2 levels decrease as S-phase progresses , and moreso in WT than in Mcm4Chaos3/Chaos3 MEFs , is consistent with the dormant origin hypothesis . The decrease may reflect displacement of dormant hexamers by active replisomes , followed by subsequent degradation or nuclear export . If WT nuclei have more dormant licensed origins than Chaos3 mutants , then WT cells would be expected show a greater loss of MCMs . The isolation of Mcm4Chaos3 provided the first demonstration that mutant alleles of essential replication licensing proteins can cause GIN and cancer [17] . Diploid budding yeast containing the same amino acid change in scMcm4 as the mouse Mcm4Chaos3 exhibited Rad9-dependent G2/M delay ( Rad9 is a DNA damage checkpoint protein ) , elevated mitotic recombination , chromosome rearrangements , and intralocus mutations [19] ( Li , X . and Tye , B . , personal communication ) . One explanation for these outcomes is that the Chaos3 mutation impairs MCM4 biochemically in a manner leading to elevated replication fork defects , and that these defects lead to the GIN and cancer phenotypes . Alternatively , and/or in addition , the observed associated pan-reductions of MCMs in mouse cells [17] raised the possibility that decreased replication licensing might be the primary or ancillary cause for the mouse phenotypes . The subsequent finding that mice ( Mcm2IRES-CreERT ) containing ∼1/3 the normal level of MCM2 had GIN and and cancer lent support for the idea that reductions in MCMs contribute to the Chaos3 phenotypes [20] . Although amounts of all MCMs were not investigated in Mcm2IRES-CreERT/IRES-CreERT mice , 65% reduction of MCM2 caused a reduction of dormant replication origins in MEFs that were replication stressed by hydroxyurea [21] . In Mcm4Chaos3/Chaos3 mice , we hypothesize that in the context of Mcm2 , 6 or 7 heterozygosity , which further reduces overall and chromatin-bound MCM levels below that already caused by Mcm4Chaos3 ( measured to be <20% of WT mRNA levels for Mcm2 ) , MCMs are reduced to a degree that compromises cell proliferation . This then translates into the various developmental defects and increased cancer susceptibility we observed . Whatever the exact mechanistic cause of these phenotypes , it is clear that the phenotypes are related to reduction of one or more MCMs below a threshold level that is <50% . The severe developmental consequences of MCM depletion in mice suggests that certain cell types in the developing embryo are highly sensitive to the effects of replicative stress , and/or that relatively minor cell growth perturbations of such cells are not well-tolerated in the context of complex , rapidly-occuring developmental events . The molecular basis for these phenotypes does not appear to be directly related to GIN , because whereas Mcm3 hemizygosity rescued several phenotypes , and delayed cancer latency in Mcm4Chaos3/Chaos3 mice , it did not concommitantly decrease MN . This suggests that phenotypes such as decreased proliferation and embryonic death are caused by genetically-induced replication stress , moreso ( or in addition to ) than GIN alone . Our genetic studies indicate that there is a quantitative MCM threshold required for embryonic viability , as demonstrated by the synthetic lethalities we observed when combining homozygosity of Mcm4Chaos3 with Mcm2Gt , Mcm6Gt or Mcm7Gt heterozygosity , but not in the heterozygous single mutants . Additionally , the Mcm4Chaos3/Gt genotype , which reduced MCM levels below 50% , caused embryonic and neonatal lethality [17] . Underscoring the exquisite sensitivity of whole animals to subtle perturbations in the DNA replication machinery were the remarkable phenotypic rescues ( viability , growth , iPS efficiency , etc . ) by Mcm3 hemizygosity . The decreased MCM dosage led to increases in S phase nuclear MCMs and chromatin-bound MCMs , presumably reflecting increased replication origin formation . The various single and compound mutants described here and elsewhere [20] , which show that 50% reductions of any one MCM is well-tolerated but decreases of ∼2/3 are not , supports the idea of a threshold effect , and suggests that the threshold lies somewhere between 1/3 and 1/2 of normal MCM levels ( at least in the cases of MCM2 , MCM6 and MCM7 ) . These results also emphasize the importance of relevant physiological models , both in general and with respect to the MCMs . RNAi knockdown of MCM3 in human cells to ∼3% normal levels was still compatible with normal short-term proliferation , although the cells had GIN and high sensitivity to replication stress [16] . It is doubtful such a drastic situation would be recapitulated in vivo ( it would likely result in embryonic lethality as in Mcm3Gt/Gt mice ) . Nevertheless , it is noteworthy in that study that MCM3 depletion was better tolerated than knockdowns of any other member of the replicative helicase . The finding that reductions in MCM3 rescued MCM2/4/6 depletion phenotypes lends insight into dynamics and regulation of mammalian DNA replication . In budding yeast , MCMs shuttle between the nucleus and cytoplasm during the cell cycle . MCM2-7 multimers must be assembled in the cytoplasm before being imported into the nucleus during G1 phase [4] . The MCM2-7 importation is dependent upon synergistic nuclear localization signals ( NLS ) on Mcm2 and Mcm3 [22] . In order to prevent over-replication of the genome , MCMs are exported from the nucleus during S , G2 and M [4] . This export is dependent upon Mcm3 , which has a nuclear export signal ( NES ) that is recognized by Cdc28 to promote MCM2-7 export in a Crm1-dependent manner [22] . In contrast to budding yeast , MCMs that have been studied ( MCM2/3/7 ) are primarily nuclear-localized throughout the cell cycle in metazoans and in fission yeast [4] . Upon dissociation from chromatin during S phase , MCM2-7 complexes are reported to remain in the nucleus but are sequestered via attachement to the nuclear envelope or other nuclear structures [24] , [33]–[35] . Interestingly , mcm mutations in fission yeast that disrupt intact MCM2-7 heterohexamers triggers active redistribution of MCMs to the cytoplasm [36] . Additionally , re-distribution of MCMs between the cytoplasmic and nuclear compartments has been observed in hormonally-treated mouse uterine cells [25] . Our observations support the idea that intracellular re-distribution of MCMs also occurs in mammals , and that it is an important regulatory process . Staining of MCM2 in intact nuclei of normal NIH 3T3 fibroblasts and MEFs show a steady decline ( but not elimination ) as S phase progresses . Furthermore , it appears that the process of nuclear MCM2 elimination during S phase is regulated , since in situations of decreased MCMs ( as in the Mcm4Chaos3/Chaos3 mutant ) , there is decreased loss of nuclear MCM2 during S phase . Three lines of experimentation implicate MCM3 as playing a key role in regulating intracellular MCM localization: 1 ) Rescue of reduced-MCM phenotypes by genetic reduction of MCM3; 2 ) Increased S-phase nuclear MCM2 by Mcm3 hemizygosity in MCM-depleted cells ( Figure 7D ) ; and increased chromatin-bound MCM2/4 by Mcm3 hemizygosity in MCM-depleted cells . Our data suggests that MCM3 acts as a negative regulator that prevents re-assembly or reloading of MCM complexes as they dissociate from DNA during replication . As described earlier , mouse and human MCM3 have predicted NESs in similar positions of their primary amino acid sequences as do the yeast genes . Thus , one explanation for these phenomena is that decreased MCM3 suppresses MCM2-7 nuclear export , which occurs normally and which may be accentuated by the Chaos3 mutation in a fashion analogous to mcm mutant fission yeast discussed above [36] . This would effectively increase the amounts of MCMs available for replication licensing . More work is required to determine if the rescue mechanism is indeed related to a decrease in MCMs export , as opposed to direct or indirect involvement in other events such as increased nuclear import or enhanced chromatin loading . With respect to the early lymphoma susceptibility phenotype in Mcm4Chaos3/Chaos3 Mcm2Gt/+ mice , it is unclear whether the type of tumor is dictated primarily by the particular Mcm depletion ( in this case MCM2 , thus resembling Mcm2IRES-CreERT2/IRES-CreERT2 animals ) , the genetic background , or the age of particular cancer onset ( if animals die of thymic lymphoma at an early age , they will be unable to manifest later-arising mammary tumors ) . The compound mutant mice used for the aging aspects of this study were bred to at least the N3 generation in strain C3H . Mcm4Chaos3/Chaos3 mice congenic in this background are predisposed exclusively to mammary tumors , whereas lymphomas were observed in mutants of mixed background [17] . Presently , we favor the idea that genetic background and age of tumor type onset are primary determinants of the cancers that arise in the mice we have studied thus far . Genetic background has also been reported to influence tumor latency in MCM2-deficient mice [21] . The MCM2-7 pan-reduction in Chaos3 cells is consistent with other studies involving mutation or knockdown of a single MCM in mammalian cells [16] , [20] , [29] , [37] . In these examples of parallel MCM decreases , the general assumption is that there is hexamer destabilization or impaired MCM chromatin loading followed by degradation of monomers . However , we found that the protein decreases are related to decreased mRNA levels . These large ( ∼40% ) decreases do not appear to be attributable to transcriptional alterations from cell cycle disruptions ( these cells have a small elevation in the G2/M population ) , but rather occur at the post-transcriptional level ( unpublished observations ) . Since we also found that MEFs carrying only 1 functional Mcm2 allele caused ∼20% decreases of Mcm3-7 mRNAs , it is possible that mRNA downregulation drove MCM reductions in these other model systems . However , the mechanism for coordinated mRNA regulation , and what triggers it , is a mystery that we are currently investigating . Our data contribute to a growing body of data that replication stress , which can occur via perturbations of the DNA replication machinery , plays a significant role in driving cancer [38]–[41] . While the Mcm4Chaos3 mutation is an unique case , the deleterious consequences of MCM reductions suggest that genetically-based variability in DNA replication factors can have physiological consequences . Such variability in functions or levels may be caused by Mendelian mutations or multigenic allele interactions . Mutations affecting transcriptional activity of one or more Mcms , which might occur in non-coding cis-linked sequences or unlinked transcription factors , could have such effects . This has implications for cancer genome resequencing projects , whereby such mutations would not be obviously associated with MCM expression . The allelic collection we generated , when used alone or in combination with each other or Mcm4Chaos3/Chaos3 mice , allow the generation of mouse models with a graded range of MCM levels . These should be valuable for investigations into the impact of replication stress on animal development , cancer formation , and cellular homeostasis .
MEFs from 12 . 5- to 14 . 5-dpc embryos were cultured in DMEM+10% FBS , 2 mM GlutaMAX , and penicillin-streptomycin ( 100 units/ml ) . Assays were conducted on cells at early passages ( up to P3 ) . For cell proliferation assays , 5×104 cells were seeded per well of a 6 well plate . They were then cultured and harvested at the indicated time points to perform cell counts . Doxycycline inducible lentiviral vectors [42] were prepared by co-transecting viral packaging plasmids psPAX2 and and pMD2 . G along with vectors encoding rtTA , Oct4 , Sox2 , Klf4 , or c-Myc ( plasmids were obtained from Addgene . org , serial numbers 12259 , 12260 , 20323 , 20322 , 20324 , and 20326 ) into 293T cells using TransIT-Lt1 transfection reagent ( Mirus ) . Viral supernatants were collected at 48 and 72 hours , and concentrated using a 30kd NMWL centrifugal concentrator . MEFs from 13 . 5d embryos , up to P3 , were seeded to gelatin coated tissue plates at a density of 6 . 75×103 cells/cm2 and allowed to attach in standard MEF media for 24 hours before infection with lentiviral vectors . After 24 hours incubation the culture media was changed to KO-DMEM supplemented with 15% KO serum replacement ( Gibco ) , recombinant LIF , 2 µg/mL doxycycline ( Sigma ) , 100 µm MEM non-essential amino acids solution , 2mM GlutaMax , 100 units/mL penicillin and 100 µg/mL streptomycin ( Gibco ) . The induction media was refreshed daily for 13 days until the cells were passaged to 100 mm plates prepared with irradiated feeders . Cells were cultured for an additional 10 days in the induction media in the absence of doxycyline before iPS colony counting , cell counts , and flow cytometry . For flow cytometric quantification of iPS cells derived from reprogramming of MEFs , ∼1×106 cells were trypsinized for 10 minutes , then washed twice with cold PBS . They were gently but completely resuspended in 1ml of 4% paraformaldehyde in PBS at room temperature for 30 minutes . The fixed cells were pelleted by centrifugation at 500×G for 2 minutes and washed twice with 10 ml TBS-TX ( 0 . 1% Triton X-100 ) buffer . For antibody staining , the cells were blocked with 1ml TBS-TX buffer with 1% BSA for 15 min at room temperature , then stained with primary “stemness” antibodies ( monoclonal anti-SSEA1 , Millipore; rabbit polyclonal anti-LIN28 , Abcam ) for 60 min , washed twice , then secondary antibody was applied for 60 minutes . Immunolabeled cells were analyzed by flow cytometry using a 488nm laser . Secondary antibodies were goat anti-mouse IgG-FITC ( South Biotech ) and goat anti-rabbit IgG-594 ( Molecular Probes ) . Cells were considered to be iPS cells if they were LIN28/SSEA1 positive . Calibration of the flow cytometer and gates were set using untransfected MEFs as negative controls , and v6 . 4 ES cells as positive controls . For quantification by colony formation , plates containing the passaged reprogrammed cells were examined microscopically at 20× , and 4 fields were scored and averaged . Colonies were considered as iPS clones based on morphological criteria: well defined border , three-dimensionality , and tight packing of cells . Micronucleus assays , which include CD71 staining , were performed essentially as described [43] . MEFs were plated at 4×106 cells/150 mm culture dish for 60 hr , trypsinized , then resuspended in 1ml PBS . To the suspension was added TX-NE ( 320 mM sucrose , 7 . 5 mM MgCl2 , 10 mM HEPES , 1% Triton X-100 , and a protease inhibitor cocktail ) . The cells were gently vortexed for 10 seconds and incubated on ice for 30 min . Dounce homogenization was unnecessary . Nuclei were then pelleted by centrifugation at 500×G for 2 min and washed twice with 10 ml TX-NE , then resuspended in 1ml TX-NE . Nuclei yield and integrity was monitored microscopically with trypan blue staining . The nuclei were fixed by adding 15ml cold methanol for 60 min on ice . The fixed nuclei were pelleted by centrifugation at 500×G for 2 min , then washed twice with 10 ml TBS-TX ( 0 . 1% TX-100 ) . 1×106 nuclei were placed into 1 . 5ml tubes in 1ml TBS-TX buffer+1% BSA for 15 min at room temperature . The primary antibody ( Rabbit anti-mouse MCM2 ) was added for 60 min , then secondary antibody ( FITC goat anti-rabbit ) was added for 60 min . Finally , the nuclei were stained with propidium iodide ( PI ) , and RNAse treated ( batches optimized empirically ) for 30 mins . Immunolabeled nuclei were analyzed by flow cytometry ( using a BD FACSCalibur cytometer with CellQuest software ) , exciting the PI and FITC with a 488nm laser . ES cell lines containing gene trap insertions in Mcm genes were obtained from Bay Genomics [Mcm3 ( RRR002 ) , Mcm6 ( YHD248 ) , Mcm7 ( YTA285 ) ] or the Sanger Institute [Mcm2 ( ABO178 ) ] . The Mcm4 line was previously reported [17] . Allele names are abbreviated as , for example , Mcm3Gt instead of the full name Mcm3Gt ( RRR002 ) Byg . All of the original ES cells were of strain 129 origin , and the alleles were backcrossed into C3HeB/FeJ for ≥4 generations . To identify the exact insertion sites of the gene trap vectors , a “primer walking” procedure was used . This involved priming PCR reactions with :1 ) a fixed vector primer , and 2 ) one of a series of primers series corresponding to the intron in which the vector presumably integrated . PCR products were then sequenced . Genotyping of gene-trap-bearing mice was performed either by PCR amplification of the neomycin resistance gene within the vector , or by using insertion-specific assays ( Table S2 ) . Cytosolic and chromatin-bound protein was extracted as described [44] . Antibody binding was detected with a Pierce ECL kit . Band were quantified using NIH Image J software . Antibodies- aMCM2: ab31159 ( Abcam ) ; aMCM3: 4012 ( Cell Signaling ) ; aMCM4: ab4459 ( Abcam ) ; aMCM5: NB100-78261 ( Novus ) ; aMCM6: NB100-78262 ( Novus ) ; aMCM7: ab2360 ( Abcam ) ; aBeta-actin: A1978 ( Sigma ) ; aTBP: NB500-700 ( Novus ) . Total RNA from P1 MEFs was DNAse I treated , then cDNA was synthesized from 1 µg of total RNA using the Invitrogen SuperScript III ReverseTranscriptase kit with the supplied Olige-dT or random-hexamer primers . qPCR reactions were performed in triplicate on 1 ng or 10 ng of cDNA by using the SYBR power green RT-PCR Master kit ( Applied Biosystems; 40 cycles at 95°C for 10 s and at 60°C for 1 min ) , and real-time detection was performed on an ABI PRISM 7300 and analyzed with Geneamp 5700 software . The specificity of the PCR amplification procedures was checked with a heat-dissociation step ( from 60°C to 95°C ) at the end of the run and by gel electrophoresis . Results were standardized to β-actin . The PCR primers are listed in Table S1 .
|
Proper replication of the genome is essential for maintenance of the genetic material and normal cell proliferation . DNA replication can be compromised by exogenous factors and genetic disruptions . Such compromise can lead to disease such as cancer , which is characterized by genomic instability ( an elevated mutation rate ) . Because the DNA replication apparatus is essential , relatively little is known about how genetic variants impact the health of whole animals . In this report , we studied mice bearing combinatorial mutations in a component of the replication apparatus , the MCM2-7 helicase . MCM2-7 is a complex of 6 proteins that are essential for initiating DNA replication along chromosomes , and to unwind the DNA during DNA replication . We find that although cells have excess amounts of MCM2-7 to support proliferation under normal circumstances , that incremental MCM depletions have multiple drastic effects upon the whole animal , including embryonic lethality , stem cells defects , and severe cancer susceptibility . Additionally , we report that mouse cells regulate and coordinate the levels of usable MCM proteins , both at the level of synthesis and also by regulating access to chromatin . The implication is that genetic variants that impact MCM levels , even to a minor degree , can translate into disease .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biochemistry/replication",
"and",
"repair",
"genetics",
"and",
"genomics/cancer",
"genetics",
"genetics",
"and",
"genomics/animal",
"genetics",
"genetics",
"and",
"genomics/chromosome",
"biology"
] |
2010
|
Incremental Genetic Perturbations to MCM2-7 Expression and Subcellular Distribution Reveal Exquisite Sensitivity of Mice to DNA Replication Stress
|
Network motifs have been identified as building blocks of regulatory networks , including gene regulatory networks ( GRNs ) . The most basic motif , autoregulation , has been associated with bistability ( when positive ) and with homeostasis and robustness to noise ( when negative ) , but its general importance in network behavior is poorly understood . Moreover , how specific autoregulatory motifs are selected during evolution and how this relates to robustness is largely unknown . Here , we used a class of GRN models , Boolean networks , to investigate the relationship between autoregulation and network stability and robustness under various conditions . We ran evolutionary simulation experiments for different models of selection , including mutation and recombination . Each generation simulated the development of a population of organisms modeled by GRNs . We found that stability and robustness positively correlate with autoregulation; in all investigated scenarios , stable networks had mostly positive autoregulation . Assuming biological networks correspond to stable networks , these results suggest that biological networks should often be dominated by positive autoregulatory loops . This seems to be the case for most studied eukaryotic transcription factor networks , including those in yeast , flies and mammals .
Gene regulatory networks ( GRNs ) are believed to play a central role in organismal development and evolution [1]–[3] . Recent theoretical and experimental studies have revealed that GRNs have many interesting quantitative and qualitative features , including scale-free structure [4] , recurring motifs [5] , robustness [6] , and evolvability [7] . Here we focus on a very specific and common network motif , autoregulation [8] , and its contribution to stability and mutational robustness [9] . A direct autoregulation motif in transcriptional GRNs consists of a regulator that binds to the promoter region of its own gene , thus regulating its own transcription . It constitutes the simplest case of a feedback mechanism . Two thirds of E . coli's transcriptional factors ( TFs ) are believed to be autoregulated [10] . The fraction of autoregulated TFs is lower for yeast ( 10% [11] ) , but extensive autoregulation at the post-transcriptional level has been suggested [12] . Two rules relating the presence of feedback loops in GRNs to their dynamical properties have been proposed [13]: ( i ) a necessary condition for multistability ( i . e . , the existence of several stable fixed points in the dynamics ) is the existence of a positive circuit in the regulatory network ( the sign of a circuit being defined as the product of the signs of its edges ) ; and ( ii ) a necessary condition for the existence of an attractive cycle in the dynamics is the existence of a negative circuit . These two types of dynamical properties have been associated with important biological phenomena: cell differentiation and stochastic switching in the first case [14] , homeostasis [9] and periodic behaviors ( e . g . , cell cycle [15] and circadian rhythms [16] ) in the second . Although these conditions are necessary , they are often not sufficient to define network dynamics , which can depend on other details of the GRN model [13] . For example , negative autoregulation ( NAR ) , the shortest negative circuit possible , has been traditionally associated with robustness of gene expression to noise [9] . However , if the NAR feedback contains a long delay , noise may be amplified [17] . Moreover , both positive and negative feedback circuits are usually embedded in larger networks , and the relative contributions of multiple positive and negative feedback loops to the dynamics of a whole network are largely unknown [13] , [14] , [18]–[20] . Here , we investigate the relationship between the sign of autoregulation and the stability and mutational robustness of genetic networks . We study this in the context of a widely used gene network model [21]–[23] , related to the modeling framework of Boolean networks [24] . We find that stability and robustness are highly correlated with the sign of autoregulation , and that selection for stability leads to positive autoregulation . Despite these positive associations , we show that selection does not maximize robustness and that it is possible to engineer networks with higher robustness by manipulating their diagonal and off-diagonal elements . We also show that autoregulation is conserved over time and that evolved networks are a special subset of stable networks ( networks that show fixed point dynamics ) with high robustness . Finally , we discuss some implications for biological systems and compare our results with biological networks of different organisms .
To study how stability , robustness and autoregulation change during evolution , we use a standard model for GRN . In one generation , we assume that that the phenotype of an organism S ( t ) develops over time t , starting from an initial phenotype S ( t = 0 ) , under the influence of a gene-interaction network W . In general , phenotypes are thought of as expression levels of the genes of the organism at time t . Thus , they are vectors of dimension N , , with binary entry values , where N is the number of genes of the organism . Phenotypes S ( t ) change by the action of a gene-interaction network that drives their development , and is represented by an matrix , W , whose elements , wij , denote the effect on gene i of the product of gene j . These interaction weights wij are nonzero and binary , . Thus , all genes either repress or activate each other's expression . In this study , we assume that size of the gene interaction network is N = 10 genes . The matrix W is not necessarily symmetric . Diagonal elements , wii , represent autoregulation , i . e . , the action of the ith gene on itself . Each network W determines the dynamics of the phenotype S ( t ) in a series of development steps . The repeated application of such development steps on a phenotype results in deterministic , discrete-time dynamics of S ( t ) , modeled by the set of nonlinear coupled difference equations: ( 1 ) where sgn ( 0 ) = 1 . This spin glass or neural network-type model [22] represents a subclass of Random Boolean Networks [24] known as Random Threshold Networks [25] . When simulating development , the network is updated synchronously , that is , only values of si from time step t are used for the calculation of si ( t+1 ) ( see [26]–[28] for asynchronous updates . ) We refer to Equation ( 1 ) as the development process ( see [23] , [29] for model illustration , biological motivation and assumptions ) . The development process can be extended to include sparse networks G . Sparse networks are used to model gene interactions in which only a fraction of the genes repress or activate a fraction of all the other genes , in contrast to fully connected networks W , where all genes have some effect on all other genes . Let G denote an interaction network represented by a , N = 10 square matrix whose entries , gij , take the values of {−1 , 0 , 1} . The parameter c , the density of the network , determines the proportion of non-zero matrix elements . When simulating sparse networks ( see Results ) , we chose c = 0 . 2 ( due to similarity to the biological networks in Table 1 ) and a regular , directed graph topology , where all genes have degree 2 . This means that all genes in a network are regulated by two genes and also regulate two other genes . To generate a random network Wr the matrix-elements wij are sampled from {−1 , 1} with equal probability ( 0 . 5 per element and entry ) . Additionally , we can generate stable networks Ws with a pre-selection procedure . In this procedure , a random network Wr and random initial state pair are first generated . This pair undergoes the development process . If no fixed point is attained , a new pair is sampled and developed . This step is repeated until some ( W , S ( 0 ) ) - pair generates a fixed point . The final network Ws , is a stable network . This notion of stability refers to an individual level stability , which differs from a notion of population level stability that will be introduced below . In this study , we refer to stability as the property of a network , while strictly speaking , it is a property of a W , S ( 0 ) pair . However , we have previously shown [30] that the network is by far the most important determinant of stability . If a network is stable/unstable with a random initial state , it most likely remains stable/unstable with any other initial state . For this reason , we classify networks as stable or unstable , even if we just solve Equation ( 1 ) for one possible initial state . Two types of simulation experiments are our primary focus in this study . First , experiments in which a population of organisms undergoes the development process only , which we refer to as non-evolved . Secondly , experiments with multiple generations ( evolved ) , where after each development process ( one generation ) the composition of the population of organisms is additionally altered by evolutionary mechanisms . The development process is completed after all organisms have reached some stage of development: either a fixed point or a cycle . We implemented standard evolutionary mechanisms , such as selection , mutation and recombination . After these evolutionary forces have acted on the population , a new development process starts in the next generation with identical initial phenotypes for each organism . In this study , we set the population size to n = 500 across all experiments ( unless otherwise noted ) . This population size remains constant during evolutionary simulations ( Wright-Fisher model with sampling with replacement ) . To study how selection affects evolving populations , we implemented different types of selection or selection models . The mutation and recombination mechanisms applied were the same for all evolved populations . Selection mechanisms modify the number of copies of one specific network within the population depending on the fitness of the phenotype that specific network has generated through development . In a selection mechanism , one phenotypic state can be marked as the optimal state , with the highest possible fitness . If such an optimal state , Sopt ( ∞ ) is specified , the fitness of a network with attractor S ( ∞ ) is given by: ( 2 ) where d is the normalized Hamming distance and σ>0 determines selection strength [23] . Small values of σ imply strong selection against deviations from the optimal state . Large values minimize the fitness difference between phenotypes . The Hamming distance d corresponds to the number of differing expression states of individual genes between two phenotypic states [31] , subsequently normalized to the interval [0 , 1] in this study . Equation ( 2 ) is valid under the assumption that Sopt ( ∞ ) and S ( ∞ ) are attractors with identical , optimal cycle lengths lopt . The cycle length of the attractor with highest fitness is denoted with lopt . The fitness of attractors of length l≠lopt depends on the selection model . We use attractor length and cycle size or period as synonyms . We implemented selection models similar to those used by other authors [23] , [29] and also introduced new ones . In these models , the fitness of a developed organism depends on two parameters: selection strength , σ , and optimal period , lopt . Selection model 1 ( selecting for stability ) : lopt = 1 , fitness ( l≠lopt ) = 0 σ = 0 . 1 ‘target’ model Selects for fixed points and an optimal gene expression state . Fitness is given by Equation ( 2 ) for fixed points and is 0 for cycles . σ = ∞ ‘no target’ model Fitness is 1 for all fixed points and 0 for cycles . Selection model 2 ( selecting against stability ) : lopt>1 , fitness ( l≠lopt ) = 0 σ = ∞ , lopt = 2 , 3 , … , 7 ( cycles ) We generalize the ‘no target’ model to select for cycles . Fitness is 1 for cycles of length l = lopt and 0 otherwise , including fixed points . We try different lopt>1 . Selection model 3 ( neutral for stability ) : σ = 0 . 1 , fitness = max ( fitness ( S ) ) for all S in S ( ∞ ) , any l ( S represents any state in the attractor S ( ∞ ) ) We generalize the ‘target’ model to not require stability . When l = 1 , we have the ‘target’ model as a special case and fitness is given by Equation ( 2 ) . When l>1 , fitness is the maximum fitness given by Equation ( 2 ) for all states in the cycle . The attractor S ( ∞ ) can be a fixed point or a cycle . Selection model 4 ( random sampling ) : σ = ∞ , fitness = 1 , any l No selection . We take this as the null model . For each selection model , we generated between z = 100 and z = 300 independent populations ( depending on the model ) . Specifically: z = 200 for the ‘target’ model; z = 300 for the ‘no target’ model; z = 200 for Selection model 2; z = 100 for Selection models 3 and 4 . We denote such an aggregation of populations as a set of populations and z as its set size . Each evolved population has a different initial state , but all individuals within the same population have the same initial state . Mutations randomly change the sign of wij at a rate μ = 0 . 1 per network per generation . All matrix entries , wij , including diagonal elements , wii , have equal probability of changing sign , namely μ/N2 = 0 . 001 per generation . For sparse networks we use a probability for changing sign of μ/ ( c N2 ) . To model recombination we follow the methods in [23] , where full chromosome segregation ( no crossover ) is implemented . The two offspring of a randomly chosen pair of recombinant parents are generated by randomly taking half the rows from each parent matrix . This procedure is performed on the entire population . We define a population-level stability ( henceforth referred to as stability if not otherwise stated ) as the fraction of networks that are stable ( individual-level ) in a given population [30]: ( 3 ) where nf≤n is the number of times the attractor is a fixed point , and n is population size , that is , the number of network matrices . Stability takes values between 0 and 1 . Similarly , we define the robustness of a population as the fraction of all possible mutated networks in a population that reach the same fixed point attractor as their un-mutated originals [23] , conditional on the fact that the attractor did not become a limit-cycle . Specifically , we estimate robustness by looping through the population of networks and mutating every element of each network matrix W ( changing the sign of wij for binary matrices ) , thus generating N2 single-mutants per network . Then , the networks undergo the development process starting from the same initial phenotypes as their originals , and are further analyzed . For a single network , we define individual-level viability as the fraction of single-mutants that attain a fixed point: ( 4 ) where nfixed<N2 is the number of times the N2 single-mutants have still generated a fixed point . With this metric , we can now define individual-level robustness: ( 5 ) and n = ≤nfixed is the number of times the same attractor state as the one attained by the un-mutated original is reached starting with identical initial conditions ( i . e . , the mutant has the same phenotype as its wildtype ) . The population-level viability and robustness measures are computed from the averages of all networks in the entire set of populations . Both robustness and viability take values between 0 and 1 . In normalizing robustness by nfixed instead of N2 , we attempt to decouple the effects of stability and robustness . In the vast majority of cases , mutations that change the stability of a network do not affect its robustness score . Exceptions to this are the rare occasions when nfixed = 0 ( robustness is not defined ) , or when nfixed is low ( robustness can only take a few specific values ) . In an extension of the fraction of activating connections-statistic [32] , we found it useful to measure properties of diagonal and off-diagonal elements of a matrix W separately , thereby decoupling the effects of direct autoregulation and off-diagonal regulation . For a single network matrix W , we define: ( 6 ) where N+p and N+q are the number of positive diagonal and off-diagonal elements of W , respectively . Both p and q are always positive and take values between 0 and 1 . We call p the sign of autoregulation , because autoregulation is predominantly positive when p>0 . 5 ( we call this positive autoregulation ) , and mostly negative when p<0 . 5 ( we call this negative autoregulation ) . The metrics p and q measure direct regulatory influence . However , network dynamics can also be affected by long-range interactions . To assess the role of such long-range regulation , we introduce a metric of indirect positive autoregulation r , which measures the fraction of autoregulatory paths over two genes that are positive ( i . e . , gene A activates gene B , which activates gene A ) . For a single network W , we define: ( 7 ) where N+r is the number of positive off-diagonal elements of WWT ( WT is the transpose of W ) . Because WWT is symmetrical , it suffices to count the fraction of positive entries in either of the triangles of the matrix . We also define metrics to assess the population-average of gene interaction strengths: ( 8 ) where n+ij is the number of positive elements in position i , j across all n networks in a population , abs is the absolute value function , and t1 and t2 are two different evolutionary time points . Here , oij , is referred to as average positive ij- interaction strength , and measures how much gene i activates gene j on average , whereas the conservation statistics between population-averaged gene interaction strengths measures how much these interaction strengths are maintained over time . In evolutionary experiments , conservation , p , q , and r are averaged over all individuals and all populations . The code utilized in this paper can be downloaded from https://github . com/rpinho/phd .
To study the relationship between the sign of autoregulation ( p ) and stability during development , we devised two experiments . First , for each p = 0 , 0 . 1 , … , 1 , a pair consisting of one random network and one random initial condition was sampled ( RNRC setup; see Methods ) . Equation ( 1 ) was then evaluated for each pair: if the attractor was a fixed point , the network was considered stable . Instead , if the solution to Equation ( 1 ) was a limit cycle , the network was considered unstable . This process was repeated n = 105 times for each p . Figure 1A shows that individual-level stability is strongly associated with p . Stable networks have significantly higher values of p than unstable networks ( median of 0 . 9 compared to 0 . 4 for unstable networks; Mann-Whitney U p-value ∼0 , Figure 1A ) . This positive association was also observed for population-level stability in a second experiment . We subdivided the networks generated in the first experiment into populations of identical p , and measured the average population-level stability for each p . We observed that the fraction of stable networks increases rapidly with higher values of p ( Figure 1B ) . These results also indicate that p and stability are strongly associated . We next studied how p changes when explicitly selecting for and against individual-level stability in evolutionary simulations . To this end , we founded sets of populations with random networks and the same initial state for each population ( RNIC setup; see Methods ) [23] , [29] . The average p was set to p = 0 . 5 at generation 0 . We then evolved all populations under the six different selection models ( including mutation and recombination ) described in the Methods . For all of these selection models , we followed the evolution of the sign of autoregulation p over 106–107 generations ( until equilibrium was attained ) . Consistent with the observations for non-evolving networks , positive autoregulation is strongly favored during evolution , both under the ‘target’ and ‘no target’ models ( Figure 2A ) . However , the evolution of p follows a complex , non-linear pattern . After a sharp initial increase over the first ∼50 generations , p reaches its maximum when population-level stability is above 95% ( stability-metric not shown in the Figure ) , and starts to decay slowly to a stable evolutionary equilibrium of p∼0 . 8 from t∼103 generations for both models ( Figure 2A ) . At the peak , a fraction of up to p∼0 . 95 , or 19/20 surviving matrices show positive autoregulation for all genes under the ‘no target’ selection model . Interestingly , selecting for cycles of length l = 2 ( i . e . against individual-level stability ) , has the opposite effect on evolving networks: p decreases sharply down to ∼0 . 3 ( Figure 2B ) , leading to negative autoregulation . A similar , but less pronounced pattern is observed when selecting for longer cycles with lengths l>2 ( Figures 2B and S2 ) . As expected , a neutral model with no selection for individual-level stability or a specific target produces random networks , with values of p centered on p∼0 . 5 ( Figure 2C ) . However , mean values of p<0 . 5 also evolve when not selecting for any particular attractor length , but still selecting for a specific target ( Figure 2C ) . Thus , the selection for stability leads to positive autoregulation . These results suggest that selection may act over direct autoregulatory motifs ( i . e . the diagonal elements of the GRN matrix ) to promote individual-level stability . If this is the case , positive diagonal elements should be overrepresented across evolved populations relative to off-diagonal elements , since selection for individual-level stability could be achieved by maximizing autoregulation p . To test this hypothesis , we calculated the average interaction weights in populations evolved from a RNIC setup ( the number of populations , z , was 100 , and the population size n = 500 ) . The metric is the average value of the wij entry of individual networks W , taken across a set of populations ( including individuals within the populations ) as well as evolutionary time . Extreme values of indicate that the matrix element i , j is identical across individuals , different populations and different generations , whereas a neutral value of means the matrix element i , j fluctuates randomly in individuals and populations and is not conserved over time . We found that , in our evolution experiments , the averaged diagonal elements attained higher values than the off-diagonal elements , consistent with stronger selection for positive autoregulation acting on the diagonal elements ( Figure S3 ) . This is further supported by the observation that selecting for individual-level stability leads to positive values of , whereas selecting against individual-level stability leads to negative values . Selecting neither for nor against stability , but still selecting for a specific target , also yields negative . To increase our confidence that the value of p = 0 . 8 emerging under the no-target model is maintained by selection for stability , we compared the effects of diagonal and off-diagonal mutations on individual-level stability in the evolved networks . We hypothesized that since single mutations in diagonal elements could result in unstable networks more often than in off-diagonal elements , networks carrying diagonal entry mutations would therefore be weeded out at higher rates . Because p and stability are correlated , p would then be maintained at a high value . In a first approach , we sampled n = 105 stable networks at t = 106 generations ( equilibrium reached ) evolved under the ‘no-target’ model . The overall fraction of networks that survive after acquiring single mutations ( viability in Equation ( 4 ) ) is high ( 95% ) . Consistent with our hypothesized maintenance mechanism , the population-level viability of networks is significantly higher for off-diagonal elements ( median of 0 . 97 compared to 0 . 93 for diagonal; Mann-Whitney U p-value∼0 , Figure 3A ) . Intriguingly , the difference in viability is even higher for t = 48 generations , when p is close to maximum ( Figure 3B ) , and the evolutionary dynamic has not yet reached equilibrium . These results suggest that a mutation of a diagonal element is more likely to lead to a cycle than mutation of off-diagonal elements . In a second approach , we studied the conservation of values in diagonal versus non-diagonal elements between two given time points . To this end , we sampled n = 105 evolved stable networks at t1 = 105 and t2 = 106 generations from the ‘no-target’ and the random models . Subsequently , we counted the fraction of positive elements oij , ( Equation ( 8 ) ) across all networks and for each position in the gene regulatory matrix ( wij ) for both t1 and t2 , and estimated conservation values as in Equation ( 8 ) . For the no-target model , we found that diagonal entries wii with positive sign are significantly more conserved over time than off-diagonal matrix elements ( medians of 0 . 86 and 0 . 70 , respectively; Mann Whitney U p-value∼0 , Figure 4A ) . These diagonal elements under the no-target model are also more conserved when compared to diagonal elements evolved under the random model ( medians of 0 . 86 and 0 . 68 respectively; Mann-Whitney U p-value∼0 , Figure 4A ) . We found no significant differences in conservation of the off-diagonal entries between the ‘no-target’ and the random models ( one-sided Mann-Whitney U p-value∼0 . 1 , Figure 4A ) . This finding provides further evidence that positive autoregulation is maintained by selection for stability . The time course of p displays an intriguing complexity . After the stability metric reaches its maximum and ceases to change , p keeps evolving and decreases to a lower equilibrium value ( Figure 2A ) . To investigate this behavior , we asked which other network parameters may also affect the evolution of p . Since Wagner [23] has previously shown that during network evolution robustness is also ( indirectly ) selected for when selecting for stability , we studied how this robustness compares with p over the course of evolution . Intriguingly , during the simulation experiments under the no-target model , we observed that robustness ( Equation ( 5 ) ) increases with time and appears to coevolve with p , reaching its maximum at the same time point at which p reaches equilibrium ( Figure 5 ) . This association , however , is also not linear: shortly after stability has reached equilibrium , robustness still increases despite the fact that p has started decreasing . Because robustness and p reach equilibrium at about the same time , we hypothesized that p could be adapting under indirect selection for robustness . In that case , the equilibrium value of p∼0 . 8 would favor higher robustness ( or perhaps: “maximize robustness” ) . To test this , we generated groups of 105 random stable networks for p = {0 . 1 , 0 . 2 , 0 . 3 , … , 1} , and calculated the robustness for each group . Robustness was assessed after running one single development process for each network in the RNRC setup . Only stable networks were considered for this analysis . Surprisingly , we found that , similarly to stability , robustness is also positively associated with p and does not have a maximum at an intermediate p∼0 . 8 ( Figure 6A ) . Contrary to our hypothesis , robustness of stable , non-evolved networks is maximized by p = 1 ( see Figure S4 for random networks not pre-selected for stability ) . This positive association was also observed when representing the data differently: stable networks binned by increasing average values of robustness also show increasing p ( Figure 6B ) . This general positive association is inconsistent with the hypothesized relationship between p and robustness after stability reaches its maximum . However , a more in-depth analysis of robustness of evolving networks at different time points reveals a completely different picture to the situation in non-evolved matrices ( Figures 6C and D ) . At both early ( t = 48 , pmax ) and late ( t = 106 , pequilibrium ) evolutionary stages , robustness is maximal in evolving matrices with values of p<1 . At early time points , when p has reached its maximum , the relationship between p and robustness is fully inverted compared to non-evolved networks , with lower p having significantly higher robustness , thus suggesting strong selection for lower p to increase robustness ( Figure 6C ) . Strikingly , at equilibrium values , robustness is non-monotonic in p and is maximal for p∼0 . 7–0 . 8 , coinciding with the equilibrium value of p ( Figure 6D ) . Therefore , these results strongly suggest that , as we hypothesized , it is the maximization of robustness during evolution that determines the equilibrium value of p . The above results may seem contradictory: whereas p and robustness show a positive association in stable networks generated completely at random ( i . e . for all non-evolved stable networks ) , this association is non-monotonic for stable networks selected by evolution ( i . e . evolved networks ) , where robustness is maximized for more intermediate values of p ( 0 . 7–0 . 8 ) . The solution to this apparent paradox might lie in the fact that evolved networks constitute only a subset of all stable networks . Remarkably , the average robustness for the subset of evolved networks is twice as high as that of non-evolved networks with similar p ( compare Figures 6A and 6D ) , suggesting that the relationship between p and robustness is modulated by other matrix characteristics on which selection can act . In order to investigate this possibility , we studied the evolution of the sign of off-diagonal elements ( q ) . There are more off-diagonal than diagonal elements; thus the former offer many more targets for mutation . However , a mutation in the off-diagonal has a smaller effect on q than a mutation on the diagonal has on p . For this reason , q may seem to evolve at lower rates than p . More importantly , off-diagonal elements represent regulation of other genes and can form larger and more complex motifs than autoregulatory loops of size one . For this reason , they are harder to study and to interpret . However , under a random model , it is clear that the expected average value of q equals 0 . 5 . The evolution of q occurs over a much smaller range than that of p , with values spanning from 0 . 5 to ∼0 . 63 . However , it also shows a non-linear pattern of co-evolution with the other variables ( Figure 5 ) . q increases up to its maximum approximately until p stabilizes , and then it starts to slowly decrease to its equilibrium value at t∼105 . It reaches equilibrium around the same time as robustness . These co-evolutionary dynamics suggests that stability and robustness may not only depend on p , but also on q . Therefore , although robustness is maximized by high p for stable networks with q = 0 . 5 , the same is not necessarily true when q>0 . 5 . This is the case at q>0 . 65 , for which robustness is higher for p = 0 . 7–0 . 8 than for p = 1 ( Figure S5 ) . Interestingly , these values correspond closely to ( p = 0 . 8 , q = 0 . 63 ) of evolved networks at generation ∼1000 , when both stability and p reach their equilibrium values . A prediction from our results is that certain combinations of p and q are more likely to provide stable networks . To test this , we combined the off-diagonal elements ( determining q ) of stable networks with low p and a q similar to that of equilibrium ( q = 0 . 53 or 0 . 54 ) , with diagonal elements of high p ( p = 0 . 9 or 1 . 0 ) . We call these networks “engineered” . As before , we generated groups of 105 random stable networks for different values of ( p , q ) and calculated the robustness for each group . Robustness was assessed after running one single development process for each network in the RNRC setup . Only stable networks were considered for this analysis . The engineered networks resulted in extremely stable and robust networks ( Figure S6A ) ; importantly , randomly sampled stable networks with the same average p and q are not nearly as stable and robust to mutations ( Figure S6B ) . These observations support the idea that features in the topology of off-diagonal elements of these matrices ( i . e . , how genes regulate one another ) may buffer the destabilizing effects of mutations . These findings also show that it is possible to engineer networks more robust than those evolved under selection for stability . Thus , robustness is not fully maximized during evolution . In fact , when the founding populations ( t = 0 ) have a mean of q = 0 . 9 ( rather than being normally distributed around 0 . 5 , see Methods ) , robustness decreases throughout evolution ( Figure S7 ) . Direct regulatory influence of genes on one another can explain qualitatively why p and q are being maximized at the beginning of evolution experiments . Their subsequent decline below the maximum values is related to constraints imposed by the indirect selection on robustness . To further elucidate how these constraints operate , we investigated how long-range interactions embedded in the matrix of direct influences ( direct interaction , i . e . , W ) of an organism could contribute to the settling of p and q below their maximum values . To this end , we repeated the evolutionary experiments tracking the measure for length-2 autoregulatory interaction r , which measures the frequency of networks that contain self-reinforcing interaction loops ( gene A activates gene B , which reactivates A , see Methods ) . We found that r is maximized early and attains values above 0 . 5 throughout evolution , indicating positive long-range autoregulation ( Figure S8 ) . Additionally , r lags behind the evolution of p and q , adapting to selective pressures at lower rates . To understand why engineered networks are more robust to mutations than random stable networks with the same values of p and q , we measured r for networks similar to the ones shown in Figure S6 . We find that r is larger for the engineered networks ( Figure S9 ) , which explains why engineered networks are more robust for the same values of p and q .
In this study , we have shown that stability and robustness positively correlate with autoregulation in a Boolean network model of gene regulation , where stable networks have mostly positive autoregulation ( p>0 . 5 ) . During evolution in the no-target model , selecting for stability leads to indirect selection for robustness . Strong selection for stability is expressed in the adaptation of direct autoregulatory network properties summarized by p , which is maximized early in evolution . The subsequent decline of p is explained by additional autoregulatory effects stemming from long-range gene interactions that allow maintenance of high stability values , while simultaneously increasing robustness . We have limited this study to small networks of 10 genes , comparable to some sub-circuits in genomes found in organisms , summarized in Table 1 . We hypothesize that larger gene numbers would lead to similar results . In previous work we have shown that stability decreases with network size , which we simulated for up to N = 104 for sparse networks ( c = 0 . 2 ) with scale-free topology [30] . We expect such a decrease in stability with N to increase the direct selective pressure on stability , as well as the indirect selective pressure on robustness in an evolutionary experiment . This is supported by the finding that large networks show a increase in robustness after selection for a target compared to small networks [23] . We have also neglected the role of bistability in the evolution of the networks . In other models of gene-regulatory networks it has been shown that mutational robustness correlates with the robustness of phenotypes to changes in initial conditions of the networks Ri [33] . If a similar correlation exists for the model presented in this study , we would expect indirect selection for less multi-stable networks due to the indirect selection for higher Ri . Networks with high Ri are expected to have large basins of attraction , decreasing the number of possible fixed points and thus multi-stability . The model presented here deviates in some important aspects from Wagner's model [23] and Siegal & Bergman's model [29] by which it is inspired . In particular , our model only includes binary matrix elements , whereas [23] , [29] allow for real valued entries . Also , in contrast to [29] , the normalization function used is not a sigmoidal , but a sign function . That results in the state vectors having real values in [29] , whereas in our model we only allow for binary states . The motivation to deviate from [23] lies in the focus on the sign of autoregulation . Since previous work [30] has shown that the behaviors of real-valued and binary-valued networks show little or no qualitative difference in the context of the questions asked here , it is technically more feasible to implement the easier , binary form of networks . Furthermore , most of the knowledge about gene regulatory networks exists in binary form , given as qualitative information about activation or repression interactions between genes . Thus , to make comparisons with the available data calculated on the basis of binary data , it was justified to limit the study to binary networks . The assumption that a population of organisms consists of random networks or has random initial conditions is unrealistic . We use random samples of networks or initial phenotypic states because we are interested in the general , overall behavior of populations with respect to some metrics . Random sampling allows us to obtain an unbiased sample of all possible networks , and to capture a part of the heterogeneity in their behavior . To satisfy more realistic assumptions , a sub-space of phenotypes that corresponds to more realistic biological phenotypes needs to be specified . How to achieve this is currently unknown , and such a restriction would have amounted to studying random initial conditions . The Boolean network model of gene regulation has recently been shown to predict specific patterns of protein and gene activity observed in a wide diversity of biological systems , including yeast [34] , [35] and mammalian [36] cell cycles , embryonic segmentation in D . melanogaster [36] , [37] , and flower development in A . thaliana [38]–[40] . Assuming biological networks correspond to stable networks [23] , [29] , [34] , our results suggest that biological networks should often be dominated by positive autoregulatory loops ( i . e . have high p ) . This seems to be the case for most eukaryotic transcription factor networks ( including yeast , flies and mammals ) , with various studies showing values of p ranging from 0 . 76 to 1 ( Table 1; with the exception of early sea urchin developmental gene regulatory networks ) , and with autoregulatory loops being highly conserved across vertebrates [41] . Moreover , in some cases , the presence of strong positive autoregulatory loops seems to be crucial to achieving a stable biological state . For example , in mammalian embryonic stem cells , the core pluripotency network of Oct4 , Sox2 and Nanog ( plus Klf4 and Esrrb [42] ) forms a tight autoregulated circuit , in which each gene activates its own expression as well as the expression of the others , and these interactions are crucial to maintaining a stable pluripotent state [43] . Furthermore , this autoregulatory circuit is likely behind the capacity of somatic cells to be reprogrammed into induced pluripotent stem ( iPS ) cells when reprogramming factors are expressed exogenously [44] . On the other hand , negative autoregulation seems to dominate in the bacterium E . coli ( p = 0 . 26 ) [45] . Stewart and coworkers [46] have recently suggested that this difference may be due to the presence/absence of sexual reproduction . To test this hypothesis , we reproduced our simulations for evolution without recombination ( see Methods ) under the no-target model , as a proxy for a model with asexual reproduction , but obtained essentially the same equilibrium values of p , despite divergent intermediate evolutionary dynamics and robustness at equilibrium ( Figure S10 ) . Another caveat may lie in the density of the networks employed in our simulations . Biological networks are often sparse [47] , and may vary between species as well as for different gene regulatory subcircuits within species; however , we have used fully connected networks in our analyses . Thus , we tested Boolean networks with the same average connectivity as some biological networks ( average degree of 2 , [47] ) ( see Methods , Sparse Networks ) . The evolutionary simulations were conducted under the RNIC setup ( no pre-selected networks , see Methods ) , in a similar fashion to the previous simulations . We obtained similar results for the long-term evolution of q , and for p in sparse networks without recombination , while we obtained even larger values of p at equilibrium for sparse networks with recombination ( Figure S11 ) , suggesting that connectivity density has a minor impact on the evolution of these parameters . These results are aligned with our previous study showing that network density and topology have only a small effect on the stability of networks of 10 genes [30] . Finally , the differences between eukaryotic and bacterial autoregulation values may also relate to the distinct regulatory processes of bacteria ( e . g . common presence of operons ) and eukaryotes ( e . g . more widespread post-transcriptional regulation ) . As new circuits of transcription factor networks are elucidated in detail , the roles of negative and positive autoregulation in organismal development and evolution should be more clearly understood .
|
Multicellular organisms show an incredible diversity of cell types in their different tissues . Functional classes of cells can be attributed to the activation and repression of genes , which enable each cell type to support different functions within the organism . These patterns of activity have been studied by means of gene regulatory networks ( GRNs ) . How these gene networks generate stable phenotypic states is thought to underlie the development and evolution of organisms . The pathways to these states are influenced by the autoregulatory properties of these networks . The stability and robustness of gene networks are used to investigate how such states are maintained . This study sheds light on how these properties relate to one another . By simulating the evolution of these networks , we show that genes depend on positive self-regulation to remain stable and robust when faced with random mutations or environmental perturbations . Assuming biological networks correspond to stable networks , our results suggest that biological networks should often be dominated by positive autoregulatory loops . This seems to be the case for most studied eukaryotic transcription factor networks , including those in yeast , flies and mammals .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"developmental",
"biology",
"theoretical",
"biology",
"gene",
"regulatory",
"networks",
"natural",
"selection",
"neutral",
"theory",
"biology",
"and",
"life",
"sciences",
"population",
"genetics",
"computational",
"biology",
"evolutionary",
"biology",
"evolutionary",
"processes",
"evolutionary",
"developmental",
"biology"
] |
2014
|
Stability Depends on Positive Autoregulation in Boolean Gene Regulatory Networks
|
Seemingly minor details of mathematical and computational models of evolution are known to change the effect of population structure on the outcome of evolutionary processes . For example , birth-death dynamics often result in amplification of selection , while death-birth processes have been associated with suppression . In many biological populations the interaction structure is not static . Instead , members of the population are in motion and can interact with different individuals at different times . In this work we study populations embedded in a flowing medium; the interaction network is then time dependent . We use computer simulations to investigate how this dynamic structure affects the success of invading mutants , and compare these effects for different coupled birth and death processes . Specifically , we show how the speed of the motion impacts the fixation probability of an invading mutant . Flows of different speeds interpolate between evolutionary dynamics on fixed heterogeneous graphs and well-stirred populations; this allows us to systematically compare against known results for static structured populations . We find that motion has an active role in amplifying or suppressing selection by fragmenting and reconnecting the interaction graph . While increasing flow speeds suppress selection for most evolutionary models , we identify characteristic responses to flow for the different update rules we test . In particular we find that selection can be maximally enhanced or suppressed at intermediate flow speeds .
The study of the success of mutants in evolving populations is a well-established focus area of research in computational biology . Approaches to this problem range from mostly theoretical work to direct application to specific biological systems [1–26] . The simplest way of modelling evolution is to dispense entirely with the notion of space and population structure , and to assume that the only relevant factors are the frequencies of the different types of individuals in the population [27–29] . Each individual in such an unstructured population can interact with all other individuals at all times . If individuals are distributed in space , and have a limited range of interaction , the population becomes structured . Not every individual can interact with every other individual at all times . It is then helpful to consider the interaction graph of the population [25 , 30–34] . Nodes of these networks represent individuals , and links between nodes stand for potential interactions . Evolutionary events take place between pairs of individuals connected by a link . The case of an unstructured population is recovered if links exist between any two individuals at all times; the interaction graph is then said to be complete . Evolution on simple graphs has been characterised mathematically ( see for instance Refs . [5–8 , 30–32] and references therein ) , and theoretical modelling has been linked to biological systems , see e . g . [24–26] . One main conclusion of this line of work is that population structure has the potential to change the dynamics of evolutionary processes [25–27 , 32–45] . For example , species that would be selected against in an unstructured population are found to organise in clusters on networks , and in this way they can coexist with fitter types , or even eradicate a resident species . In many stylised models of evolution birth and death events are coupled: when one individual dies , another one is born . This maintains a constant overall population size and facilitates the mathematical analysis [27–29] . It is recognised that the sizes of real populations fluctuate . Nevertheless , models with constant population size carry biological relevance; for example , they have been used to describe cancer cell populations [9–15] , competition in microbial systems [16–23] and the evolution of cooperation [46–48] . Analytical results are available for models with constant population size , and serve as important benchmark; the effects of gradually introducing additional features can be tested against this baseline . In this context it has been shown that certain interaction graphs can promote selection , while others suppress it [24–26 , 36–43] . One key quantity used to characterise selection is the probability with which an invading mutant reaches fixation in an otherwise wildtype population . Specifically , selection is said to be ‘amplified’ or ‘suppressed’ when the fixation probability of a fitter mutant is higher or lower than in unstructured populations , respectively . The availability of analytical results for stylised models is particularly convenient for making such comparisons . The effects of a given population structure on the success of an invading mutant can depend on the microscopic process chosen to model evolution . For example , the order of birth and death events can reverse the effect of the population structure [41 , 43 , 49] . One can further distinguish between models with global and with local selection [1 , 2 , 27] , and again the outcome of evolution can be different depending on what type of model is used . Choosing the right evolutionary model for a given biological system is therefore an important and intricate task . Further complications arise if the members of the populations are in motion . The interaction graph then becomes dynamic , making mathematical approaches more difficult . At the same time motion is a ubiquitous feature of biological systems , present for example due to self-propulsion of microswimmers by means of flagella [50] , or advection of bacteria in a fluid environment [51] . The movement of the population has been found to modify the performance of a mutant . For example , differences in fixation probabilities have been found in static versus stirred populations of Escherichia coli [52–54] . Recognising that motion is an important aspect of evolutionary systems , the purpose of our study is to investigate how the success of an invading mutant is affected by dynamic population structure . In order to be able to compare our results with those in unstructured populations and in populations with static interaction networks , we restrict the analysis to the most common models used in this context . Specifically , we focus on birth-death and death-birth processes in populations of constant size . A common way of implementing motion in models of evolution is migration; in these models individuals move to neighbouring sites on the interaction graph [24 , 55–60] . Alternatively , adaptive networks have been considered; in these networks individuals can re-wire their links to other members of the population , usually with preference for links between individuals of similar types [34 , 61–65] . A separate approach assumes that organisms move randomly [66 , 67] . Work based on passive motion in flows includes [68]; the evolution of movement in algae in water columns has been studied in [69] . Much of this existing work on evolution in systems with mobility is based on deterministic models , continuous both in space and time; see however [70] for a stochastic approach . Work using continuous deterministic or stochastic reaction-diffusion-advection equations to describe populations of mobile individuals can be found in [71–73] . In this paper , we study populations that are not self-propelled and use the type of motion one could expect in dynamic gaseous or aqueous environments . Specifically , the motion is due to a flow of the medium in which the population resides . In other words , it is the type of motion one would expect when populations are stirred or mixed by external forces . The movement is not constrained by the current interaction network , and the interaction graph itself is dynamic . Similar implementation of motion can be found in refs . [60 , 71–79] . Using this type of motion , and systematically studying its effects at different flow speeds , allows us to connect known results for static heterogeneous graphs and for well-stirred populations . We focus on the rate of successful fixation of a single invading mutant in populations of discrete individuals . We find that the way in which the flow affects its success depends on the choice of the evolutionary update rules . Specifically we find differences between birth-death dynamics and death-birth processes , and our study shows that it is important to consider whether selection is global or local in the evolutionary model . We note that the distinction between death-birth versus birth-death dynamics is not usually possible in models based on continuous population densities . We identify three main factors contributing to the effects flow has on the evolution of mutants in discrete populations: how well connected the initial mutant is with the rest of the population , the opportunities mutants have to organise in clustered groups , and how long individuals remain connected for as the flow moves them in space . These factors are influenced by the speed of the flow and , depending on the evolutionary update rule , they can amplify or suppress selection relative to unstructured populations . We speculate that this may be used to discriminate between different stylised models . In some experimental settings flow can be controlled externally , or situations without flow can be compared to those with fast flows . If such data is available , systematically studying the behaviour of different computational models of evolution in flowing populations can help to select the update mechanism which best captures the features of the biological system at hand .
We use the same setup as ref . [79] , and consider a population of fixed size N composed of individuals of two species ( wildtype and mutant ) . Unless specified otherwise , we use N = 100 . Individuals take positions in space within the two-dimensional domain 0 ≤ x , y < 1 with periodic boundary conditions . Particles are subject to a continuous-time flow , moving them around in space , and to evolutionary dynamics , which change the frequencies of the two species in the population . The motion of the particles is simulated through the so-called parallel shear flow [80 , 81]; we discuss the validity of our results for different flow fields in the Discussion section . The velocity field of this flow is periodic in time , except for a random phase described below . During the first half of each period particles are moved vertically; the speed of each individual depends on the horizontal component of their position . During the second half of the period individuals move horizontally , with speeds dependent on their vertical positions . We write vx ( x , y , t ) and vy ( x , y , t ) for the velocity components of a particle at position ( x , y ) at time t . Specifically , we use v y ( x , y , t ) = 0 , v x ( x , y , t ) = V maxsin [ 2 π y + ψ ] , for t ∈ [ n T , n T + T / 2 ) , v x ( x , y , t ) = 0 , v y ( x , y , t ) = V maxsin [ 2 π x + ψ ] , for t ∈ [ n T + T / 2 , ( n + 1 ) T ) , with n = 0 , 1 , 2 , … . The constant Vmax sets the amplitude of the flow , and T the period . The phase ψ is drawn randomly from the interval [0 , 2π ) at the beginning of each half-period . Due to this random phase , the flow mimics chaotic motion; the trajectories of individuals who are initially close to each other diverge over time . At long times , the distribution of individuals moved by this flow is uniform in space [80 , 81] . The evolutionary process is implemented through coupled birth and death events . The order in which reproduction and removal take place is important , and so we will distinguish between birth-death and death-birth processes . The evolutionary dynamics occur on an undirected interaction graph , dynamically generated by the flow . Specifically , we will say that one individual is a neighbour of another if they are within a distance R of each other . In the main text we focus on a scenario in which mutation is separate from reproduction , and therefore we initialize our simulations with a single initial mutant chosen uniformly at random . The case of mutation during a reproduction event is more adequately simulated by temperature initialization , which we discuss in part D of S1 Text in the Supporting Information . Individuals are in continuous motion , but evolutionary events occur at discrete times , t = Δt , 2Δt , … in our model . Simulations are then implemented as follows: Due to the coupled birth and death events the size of the population in the model is constant over time . This allows for comparison against analytical results for populations on complete graphs as discussed below . It also facilitates the memory allocation required for the simulations; the individuals’ locations , species type and the adjacency matrix of the population are arrays of fixed size . In each evolutionary step two individuals are chosen . For simplicity , we will say that an individual is ‘picked’ when it is chosen at random , disregarding fitness differences , or that it is ‘selected’ when the choice of the individuals is made through competition , i . e . , proportional to fitness . For the latter case we focus on frequency-independent selection; we set the wildtype fitness to one , and write r for the fitness of the mutant species . Consider for example a group of nw wildtype individuals and nm mutants . A mutant would be selected to reproduce from this group with probability rnm/ ( rnm + nw ) , or a wildtype with probability nw/ ( rnm + nw ) . If selection is for death we proceed similarly , but with r replaced by 1/r . In this way , mutants are less likely to die than wildypes if r > 1 . For r < 1 the mutant species is selected against . The simulation results shown in this paper focus on advantageous mutants; we set r = 1 . 05 throughout . Selection proportional to fitness can take place either in step 3 of the above algorithm ( when an individual is chosen from the entire population ) or in step 4 ( when it is chosen from the neighbours of an individual ) . We refer to these cases as global and local selection , respectively . Since we distinguish between birth-death and death-birth processes , four combinations are possible: global birth-death ( Bd ) , global death-birth ( Db ) , local birth-death ( bD ) and local death-birth ( dB ) . The capital letter in these acronyms indicates that selection dependent on fitness occurs in the respective step . In principle , one could also consider processes in which individuals are chosen proportional to fitness in both steps of the algorithm ( BD , DB ) [1 , 2] . In order to be able to disentangle the effects that the flow has on fixation probabilities due to local or global selection , we limit the discussion in the main text to scenarios in which selection acts either globally or locally , but not both . The BD and DB processes are discussed in part C of S1 Text in the Supporting Information . We illustrate the different evolutionary processes in Fig 1 . The upper two rows correspond to processes in which competition takes place among the entire population ( global selection ) . In the lower two rows the first node is picked irrespective of fitness , and competition takes place only among the neighbours of this node ( local selection ) . A step-by-step description of each of the processes can be found in the figure caption . One important characteristic of the flow is the typical timescale over which the set of neighbours of a given individual is renewed . More precisely , the set of neighbours of a given individual at time t , and at a later time t + τ , will be uncorrelated provided τ is sufficiently large ( see ref . [79] , and part A of S1 Text in the Supporting Information ) . This renewal time is in turn determined by the parameters Vmax , T and R; following refs . [82 , 83] , we use Vmax = 1 . 4 and T = 1 throughout , and choose an interaction radius of R = 0 . 11 . This choice of parameters leads to an estimate for the network renewal time of τ ≈ 6 . 4 ( see part A of S1 Text in the Supporting Information for details ) . That is , the set of neighbours of one individual at one time is uncorrelated from its set of neighbours approximately six and a half flow periods earlier . It remains to specify how frequent evolutionary events are , i . e . to define the time step Δt in the simulation described above . We treat this as a model parameter , and use S = NΔt/T to quantify the number of generations elapsed in one flow period . Thus , S indicates the speed of the flow relative to that of evolution . For small S , individuals move relatively little between evolutionary events ( ‘slow flow’ ) . Large values of S describe fast flows . From here on , we will refer to S as the speed of the flow , and investigate the outcome of evolution for different choices of this parameter . The flow speed S is understood throughout as relative to the rate of evolutionary events . We note that the inverse of S is related to the Damköhler number in fluid dynamics [84–86] .
We first address the case in which the initial coordinates of each individual are drawn from a uniform distribution on the domain 0 ≤ x , y < 1 . The initial interaction graph is then a random geometric graph ( RGG ) [87] . For any non-zero flow rate ( S > 0 ) any member of the population can eventually interact with any other individual , even if they were not connected on the initial interaction graph . This is due to the mixing properties of the flow , and means that no individual can indefinitely remain isolated from the rest of the population . As a consequence , the final outcome of the evolutionary process is either fixation or extinction of the mutant . The fixation probability , ϕ , for a beneficial mutation is depicted in Fig 2 as a function of the flow speed , S . We show simulation results for the four different evolutionary processes bD , dB , Bd , and Db . Each data point is obtained from an ensemble of realisations . For comparison , we also show the fixation probability on a complete graph , ϕCG . By definition , ϕCG is independent of the flow speed , as all individuals interact with all others at all times . On complete graphs the fixation probability for global and local selection processes differ by a small amount [1] . Several interesting features can be observed in Fig 2: For slow flows , the order of reproduction and removal is found to have a strong effect on the fixation probability , and it is less relevant whether selection takes place in the first or the second step of each evolutionary event . For the local and global death-birth processes ( dB , Db ) the fixation probability is lower than on a complete graph , as shown by the green and blue lines in Fig 2 . Conversely , both Bd and bD show a higher fixation probability than on complete graphs ( red and purple lines ) . In the limit of fast flows , however , the outcome of evolution is mostly determined by whether selection is global or local , and not by the order of the reproduction and removal events ( birth-death vs . death-birth ) . Specifically , when selection acts locally the fixation probability of the mutant is lower than on a complete graph ( purple and blue lines ) . In contrast , when selection is global the fixation probability is the same as on a complete graph ( red and green lines ) . These observations indicate unique responses of the fixation probability to the flow speed for the different processes . For the Db process ( continuous green line in Fig 2 ) the mutant’s probability of success increases with the speed of the flow . For the Bd process ( continuous red line ) , the fixation probability decreases with increased flow speed , but is always greater than or equal to the one on a complete graph , ϕ ≥ ϕCG . In contrast , the fixation probability for a dB process ( dashed blue line ) is always smaller than ϕCG . Finally , for the bD process ( dashed purple line ) the fixation probability is higher than on a complete graph when the flow is very slow , but decreases at higher flow speeds and eventually becomes lower than on the complete graph . The bD process is the only case in which we observe a transition from amplification to suppression of selection ( relative to the complete graph ) as the flow speed is increased . In order to gain some insight into these observations , we first describe the dynamics in the limit of fast flows , summarising the results of ref . [79] . Then we discuss the no-flow limit , and subsequently the transition between the two extremes , at intermediate flow speeds . Our model describes a population in constant motion . It is then natural to assume that the positions of the individuals at the time the initial mutation occurs is drawn from the stationary distribution of the flow . For the periodic parallel shear flow this is the uniform distribution , used as an initial condition in the previous section . However , exploring different starting positions allows us to gain further insight into the effect of the flow on fixation probabilities . The data shown as lines in Fig 2 was obtained from simulations with random initial positions ( RGGs ) and non-vanishing flows . For this setup the interaction graph may not be connected , but fixation or extinction will still occur , provided there is non-zero flow . In order to explore the no-flow limit , in Figs 3a and 4 we focused on static heterogeneous graphs instead; studying fixation in the strict absence of flow only makes sense when the interaction graph consists of one single connected component , and so we restricted the discussion to connected random geometric graphs ( CRGGs ) . As a result , comparison with the data in Fig 2 is difficult; in particular , we note the quantitative differences between the square markers , obtained from static connected graphs , and the limiting values of the data shown as lines in Fig 2 , obtained from slowly moving populations started from RGGs . For comparison , we show data obtained from mobile populations , but started on CRGGs , in Fig 5 . The limiting values of the fixation probabilities for very slow flows ( end of the tick lines on the left-hand side of the figure ) now agree quantitatively with those obtained from static CRGGs ( square markers ) . The simulation data from Fig 2 , from simulations with unrestricted random initial positions , is also shown in Fig 5 ( thin lines ) . If the flow is sufficiently fast , initial conditions are immaterial . On the contrary , for slow flows the fixation probability , ϕ , for simulations started from unrestricted random graphs is different from that for connected initial conditions . For birth-death processes , ϕ is greater for the unrestricted case than for the connected one . The opposite is observed for death-birth processes . This indicates that the initial condition can have a significant effect on the outcome when then flow is slow . As briefly mentioned before , the fragmented nature of the unrestricted setup can isolate groups of nodes from the rest of the population . As the evolutionary dynamics proceed , this promotes the formation of clusters , i . e . parts of the graph in which all individuals are of the same species . Clusters arise from the spread of the mutation from individual mutants seeded in different parts of the system to their local neighbours . We note that this is not restricted to discrete individual-based models , but that similar domain formation can , in principle , be expected in models with continuous population densities and in continuous time . The degree of clustering can be quantified through the fraction of active links in the network , that is , the proportion of links between mutants and wildtypes among all links in the graph , Lact/Ltot . A small fraction of active links is an indicator of clustering . We show measurements of the fraction of active links in Fig 6 for both unrestricted and restricted random initial conditions ( thin dotted lines and thick continuous lines , respectively ) . The data indicates that the fraction of active links is significantly larger when simulations are initialised on CRGGs than when started on RGGs . The amplification or suppression of selection ( for birth-death and death-birth processes , respectively ) can then be supported by a similar argument to the one presented for the degree of the initial mutant . A smaller number of active links has the same effect as poor connectivity of the initial mutant; it does not affect the probability that the individual chosen in the first step of the evolutionary process is a mutant or a wildtype , but it reduces the probability that the individual chosen in the second step is of the opposite species ( see also part B of of S1 Text in the Supporting Information ) . In the early stages of the evolutionary process mutants are a minority , and are therefore less likely to be chosen in the initial step . A large number of active links then increases the chances that the neighbour of the initial individual is a mutant . Under birth-death processes this means that mutants are more likely to die; for death-birth processes they have more opportunities to reproduce . Therefore , a connected initial configuration ( CRGGs ) , leading to a larger fraction of active links than arbitrary RGGs , reduces the fixation probability of a mutant under birth-death processes , and increases it for death-birth processes . This is in line with the results on the left-hand side of Fig 5; the fixation probabilities for RGGs ( dotted lines ) are higher than their counterparts on CRGGs for birth-death processes ( red and purple lines ) , but lower for death-birth processes ( green and blue lines ) . This argument is only valid when mutants are less abundant than wildtypes . The effect is reversed at later stages of the evolutionary process ( if mutants become a majority ) . However , the results presented in Fig 5 suggest that there is a net advantage for the mutant in having fewer active links , for birth-death processes , or in having increased inter-species connectivity , for death-birth mechanics . The inset in Fig 6 helps to understand this further . It shows the conditional fixation probability of the mutant species , given that a state with i mutants has been reached . The shape of the curves indicates that increasing the number of mutants in the population has stronger repercussions on the fixation probability when mutants are a minority ( i/N ≤ 0 . 5 ) than when they are the majority ( i/N ≥ 0 . 5 ) . For death-birth processes , the selective effect due to increased active links drives the population composition to states with approximately equal frequencies of the two species . However , the mutants have more to gain ( in terms of fixation probability ) when their numbers are small than what they may lose when they are abundant . For birth-death processes , on the other hand , a large number active links acts in the opposite way; it hinders the spread of the mutant species when they are a minority and encourages it once they are abundant . Since more is lost in the early invasion than what can be gained at later stages , the overall fixation probability is lower than when there are fewer active links . The net effect of fragmentation ( i . e . , a reduced number of active links ) is therefore amplification of selection for birth-death processes , and suppression for death-birth update rules . The amplification/suppression effect caused by the fragmented nature of the network can also be noticed at intermediate flow speeds . In this regime , the flow is sufficiently fast to disrupt the initial network structure before the evolutionary process reaches its conclusion ( fixation or extinction of the mutant ) ; disconnected components then develop . At the same time the flow is also slow enough to allow the formation of organised clusters of mutants and wildtypes through the evolutionary dynamics . Indeed , for simulations started on connected graphs a minimum in the fixation probability as a function of the flow speed is discernible for the Db process ( thick green line in Fig 5 ) , and we also notice a shallow maximum for the Bd and bD processes ( thick red and purple lines , respectively ) . The fragmentation from an initially connected network increases the fixation probability for birth-death processes and decreases it for death-birth processes . Movement of the population , and the resulting mixing between evolutionary events counteracts this amplification or suppression , driving fixation probabilities to their fast-flow limits . The balance of these two effects leads to the extrema in Fig 5 . Regular lattices are particularly convenient for the study of fixation probabilities . The nodes are distributed equidistantly in space , and they all have the same number of neighbours . This means that analytical results can be obtained in the absence of flows . For example , the isothermal theorem [30] applies; the fixation probabilities of the global birth-death and death-processes are the same as those for complete graphs; only small deviations from ϕCG are expected for local-selection processes [1] . In order to relate the success of mutants in populations advected by flows to these benchmark results , we show the outcome of simulations in which individuals are initially placed on the nodes of a regular lattice in Fig 7 . Broadly , three different regimes can be distinguished:
We studied evolutionary dynamics in populations immersed in flows . In computer simulations , we measured the effect that the speed of the motion has on the success of an invading mutant , and found that the outcome of evolution can be affected by seemingly minor details of the model used to describe evolution . Our results highlight the importance of including motion in the modelling of evolutionary dynamics . Just as static population structure can generate amplification or suppression of selection , we find that flow can act against or in favour of mutant invasion . While the models we study are stylised , we can identify general emerging principles . For instance , for the majority of evolutionary processes we observe a decrease in fixation probability when populations are in motion . This observation could be useful , for example , in industries where mutations are detrimental for the desired product but beneficial to the mutant , such as in microalgae , bacteria , fungi and yeast , relevant for the production of biodiesel [94–97] . Another example are the features we found to dominate fixation probability in the limits of very slow or very fast flows . If populations are mostly static in an experiment , our results indicate that whether selection acts locally or globally is a more important factor than the order of birth and death events . This is an important consideration for the choice of model to describe a particular system . If an experiment involves populations in motion , on the other hand , it is more important to decide whether to use a birth-death or a death-birth process as a model; in what step of evolutionary events competition takes place are less relevant in such situations . We note that the distinction between birth-death and death-birth processes requires an individual-based modelling approach; for example , no particular order of death and birth events is usually specified in models using continuous reaction-diffusion-advection equations . It is appropriate to briefly comment on the limitations of our study . We focused on the periodic parallel shear flow in our simulations . However , we note that most features of the amplification or suppression of selection arise from the mixing of the population ( i . e . , the renewal of the set of neighbours of any one individual ) , and the heterogeneity of the interaction network . Both of these features can be expected in most real flows . While there may be quantitative differerences , we believe that the essence of our findings—modified selection strength with changing flow speed—is relevant beyond the exemplar of the shear flow . This is supported by observations in our earlier work [79] , in which we obtained analytical results for the limit of fast flows , and demonstrated that these predictions are independent of many details of the flow field . At the same time , our work also opens up another view on the enhancement and suppression of selection due to mixing . We controlled the degree of mixing by changing the flow speed . That is to say , the neighbourhood of any one individual becomes increasingly less correlated in time as the flow speed increases . There might be other ways to generate such uncorrelated neighbourhoods . For example , it is known that certain flows enhance diffusive mixing [98] . One may therefore speculate that it might also be possible to obtain minimal or maximal fixation probability as the shape of the flow , and with it the degree of mixing , is varied at fixed flow speed . Our study is also limited to frequency-independent selection; natural extensions would include more complex fitness functions to better model the experimental situation in ref . [53] , where frequency-dependent fitness was identified for static conditions . While dilution techniques or resource-limited environments can be used to keep the population approximately constant in experiments without significantly modifying the mutants’ success [99] , we note that future modelling work might relax the assumption of a fixed population size . This may be useful to explore the effects of demographic stochasticity . It is then also possible to introduce a carrying capacity . Such models can then account for localised regions of high densities , generated by the flow , and leading to areas in which the carrying capacity is exceeded , and birth is suppressed . This in turn can result in the collapse of that group of individuals , a potential effect not captured by our approach . Despite the fact that direct measurements of the success of a specific mutation are not necessarily easy to perform , recent advances in technology make direct measurements of the fixation probability of a specific mutation feasible [23] . Experimental evidence of differences in fixation probabilities in static and in stirred populations can already be found in the literature [52–54] . In these studies , cultures of E . coli were grown in a continuously stirred liquid medium , on Petri dishes mixed every 24 hours , and on static Petri dishes . The structure and cluster formation of the cultures were found to have different dynamics under the different mixing conditions . The authors of ref . [54] , for example , find that the ability to adapt , as measured by reproduction rates , is greater in the continuously-stirred case than in the case of only occasional mixing . This suggests a lower fixation probability in the slowly moving medium . Experimental observations of this kind highlight the relevance of studying the effects of motion on the mechanics of fixation . Our work provides an avenue to understanding the key factors affecting fixation in models of mobile populations .
|
Whether a mutation spreads in a population or not is one of the most important questions in biology . The evolution of cancer and antibiotic resistance , for example , are mediated by invading mutants . Recent work has shown that population structure can have important consequences for the outcome of evolution . For instance , a mutant can have a higher or a lower chance of invasion than in unstructured populations . These effects can depend on seemingly minor details of the evolutionary model , such as the order of birth and death events . Many biological populations are in motion , for example due to external stirring . Experimentally this is known to be important; the performance of mutants in E . coli populations , for example , depends on the rate of mixing . Here , we focus on simulations of populations in a flowing medium , and compare the success of a mutant for different flow speeds . We contrast different evolutionary models , and identify what features of the evolutionary model affect mutant success for different speeds of the flow . We find that the chance of mutant invasion can be at its highest ( or lowest ) at intermediate flow speeds , depending on the order in which birth and death events occur in the evolutionary process .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"bacteriology",
"infographics",
"organismal",
"evolution",
"interaction",
"networks",
"population",
"dynamics",
"microbiology",
"fungal",
"evolution",
"microbial",
"evolution",
"population",
"biology",
"computer",
"and",
"information",
"sciences",
"mycology",
"molecular",
"biology",
"data",
"visualization",
"natural",
"selection",
"graphs",
"biology",
"and",
"life",
"sciences",
"evolutionary",
"biology",
"bacterial",
"evolution",
"evolutionary",
"processes"
] |
2019
|
Motion, fixation probability and the choice of an evolutionary process
|
Grade of membership models , also known as “admixture models” , “topic models” or “Latent Dirichlet Allocation” , are a generalization of cluster models that allow each sample to have membership in multiple clusters . These models are widely used in population genetics to model admixed individuals who have ancestry from multiple “populations” , and in natural language processing to model documents having words from multiple “topics” . Here we illustrate the potential for these models to cluster samples of RNA-seq gene expression data , measured on either bulk samples or single cells . We also provide methods to help interpret the clusters , by identifying genes that are distinctively expressed in each cluster . By applying these methods to several example RNA-seq applications we demonstrate their utility in identifying and summarizing structure and heterogeneity . Applied to data from the GTEx project on 53 human tissues , the approach highlights similarities among biologically-related tissues and identifies distinctively-expressed genes that recapitulate known biology . Applied to single-cell expression data from mouse preimplantation embryos , the approach highlights both discrete and continuous variation through early embryonic development stages , and highlights genes involved in a variety of relevant processes—from germ cell development , through compaction and morula formation , to the formation of inner cell mass and trophoblast at the blastocyst stage . The methods are implemented in the Bioconductor package CountClust .
We assume that the RNA-seq data on N samples has been summarized by a table of counts CN×G = ( cng ) , where cng is the number of reads from sample n mapped to gene g ( or other unit , such as transcript or exon ) [19] . The GoM model is a generalization of a cluster model , which allows that each sample has some proportion ( “grade” ) of membership , in each cluster . For RNA-seq data this corresponds to assuming that each sample n has some proportion of its reads , qnk coming from cluster k . In addition , each cluster k is characterized by a probability vector , θk⋅ , whose gth element represents the relative expression of gene g in cluster k . The GoM model is then c n 1 , c n 2 , ⋯ , c n G ∼ Multinomial c n + , p n 1 , p n 2 , ⋯ , p n G , ( 1 ) where p n g ≔ ∑ k = 1 K q n k θ k g . ( 2 ) The number of clusters K is set by the analyst , and it can be helpful to explore multiple values of K ( see Discussion ) . To fit this model to RNA-seq data , we exploit the fact that this GoM model is commonly used for document clustering [8] . This is because , just as RNA-seq samples can be summarized by counts of reads mapping to each possible gene in the genome , document data can be summarized by counts of each possible word in a dictionary . Recognizing this allows existing methods and software for document clustering to be applied directly to RNA-seq data . Here we use the R package maptpx [20] to fit the GoM model . Fitting the GoM model results in estimated membership proportions q for each sample , and estimated expression values θ for each cluster . We visualize the membership proportions for each sample using a “Structure plot” [21] , which is named for its widespread use in visualizing the results of the Structure software [7] in population genetics . The Structure plot represents the estimated membership proportions of each sample as a stacked barchart , with bars of different colors representing different clusters . Consequently , samples that have similar membership proportions have similar amounts of each color . See Fig 1 for example . To help biologically interpret the clusters inferred by the GoM model we also implemented methods to identify , for each cluster , which genes are most distinctively differentially expressed in that cluster; that is , which genes show the biggest difference in expression compared with the other most similar cluster ( see Methods ) . Functions for fitting the GoM model , plotting the structure plots , and identifying the distinctive ( “driving” ) genes in each cluster , are included in our R package CountClust [22] available through Bioconductor [23] .
We begin by illustrating the GoM model on bulk RNA expression measurements from the GTEx project ( V6 dbGaP accession phs000424 . v6 . p1 , release date: Oct 19 , 2015 , http://www . gtexportal . org/home/ ) . These data consist of per-gene read counts from RNA-seq performed on 8 , 555 samples collected from 450 human donors across 53 tissues , lymphoblastoid cell lines , and transformed fibroblast cell-lines . We analyzed 16 , 069 genes that satisfied filters ( e . g . exceeding certain minimum expression levels ) that were used during eQTL analyses by the GTEx project ( gene list available in http://stephenslab . github . io/count-clustering/project/utilities/gene_names_all_gtex . txt ) . We fit the GoM model to these data , with number of clusters K = 5 , 10 , 15 , 20 . For each K we ran the fitting algorithm three times and kept the result with the highest log-likelihood . As might be expected , increasing K highlights finer structure in the data , and for brevity we focus discussion on results for K = 20 ( Fig 1 ( a ) ) , with results for other K shown in S1 Fig . For comparison we also ran several other commonly-used methods for clustering and visualizing gene expression data: Principal Components Analysis ( PCA ) , Multidimensional Scaling ( MDS ) , t-Distributed Stochastic Neighbor Embedding ( t-SNE ) [24 , 25] , and hierarchical clustering ( Fig 2 ) . These data present a challenge to visualization and clustering tools , because of both the relatively large number of samples and the complex structure created by the inclusion of many different tissues . Indeed , neither PCA nor MDS provide satisfactory summaries of the structure in these data ( Fig 2 ( a ) and 2 ( b ) ) : samples from quite different tissues are often super-imposed on one another in plots of PC1 vs PC2 , and this issue is only partly alleviated by examining more PCs ( S2 Fig ) . The hierarchical clustering provides perhaps better separation of tissues ( Fig 2 ( d ) ) , but producing a clear ( static ) visualization of the tree is difficult with this many samples . By comparison t-SNE ( Fig 2 ( b ) ) and the GoM model ( Fig 1 ( a ) ) both show a much clearer visual separation of samples by tissue , although they achieve this in very different ways . The t-SNE representation produces a two-dimensional plot with 20–25 visually-distinct clusters . In contrast , the GoM highlights similarity among samples by assigning them similar membership proportions , resulting in groups of similarly-colored bars in the structure plot . Some tissues are represented by essentially a single cluster/color ( e . g . Pancreas , Liver ) , whereas other tissues are represented as a mixture of multiple clusters ( e . g . Thyroid , Spleen ) . Furthermore , the GoM results highlight biological similarity among some tissues by assigning similar membership proportions to samples from those tissues . For example , samples from several different parts of the brain often have similar memberships , as do the arteries ( aorta , tibial and coronary ) and skin samples ( sun-exposed and un-exposed ) . Although it is not surprising that samples cluster by tissue , other results could have occurred . For example , samples could have clustered according to technical variables , such as sequencing batch [26] or sample collection center . While our results do not exclude the possibility that technical variables could have influenced these data , the t-SNE and GoM results clearly demonstrate that tissue of origin is the primary source of heterogeneity , and provide a useful initial assurance of data quality . While in these data both the GoM model and t-SNE highlight the primary structure due to tissue of origin , the GoM results have at least two advantages over t-SNE . First , the GoM model provides an explicit , quantitative , estimate of the mean expression of each gene in each cluster , making it straightforward to assess which genes and processes drive differences among clusters; see Table 1 ( and also S1 Table ) . Reassuringly , many results align with known biology . For example , the purple cluster ( cluster 18 ) , which distinguishes Pancreas from other tissues , is enriched for genes responsible for digestion and proteolysis , ( e . g . PRSS1 , CPA1 , PNLIP ) . Similarly the yellow cluster ( cluster 12 ) , which primarily distinguishes Cell EBV Lymphocytes from other tissues , is enriched with genes responsible for immune responses ( e . g . IGHM , IGHG1 ) and the pink cluster ( cluster 19 ) which mainly appears in Whole Blood , is enriched with genes related hemoglobin complex and oxygen transport ( e . g . HBB , HBA1 , HBA2 ) . Further , Keratin-related genes characterize the skin cluster ( cluster 6 , light denim ) , Myosin-related genes characterize the muscle skeletal cluster ( cluster 7 , orange ) , etc . These biological annotations are particularly helpful for understanding instances where a cluster appears in multiple tissues . For example , the top genes in the salmon cluster ( cluster 4 ) , which is common to the Gastroesophageal Junction , Esophagus Muscularis and Colon Sigmoid , are related to smooth muscle . And the top genes in the red cluster , highlighted above as common to Breast Mammary tissue , Adipose Subcutaneous and Adipose Visceral , are all related to adipocytes and/or fatty acid synthesis . A second advantage of the GoM model is that , because it allows partial membership in each cluster , it is better able to highlight partial similarities among distinct tissues . For example , in Fig 1 ( a ) the sky blue cluster ( cluster 13 ) , appears in testis , pituitary , and thyroid , reflecting shared hormonal-related processes . At the same time , these tissues are distinguished from one another both by their degree of membership in this cluster ( testis samples have consistently stronger membership; thyroid samples consistently weaker ) , and by membership in other clusters . For example , pituitary samples , but not testis or thyroid samples , have membership in the light purple cluster ( cluster 2 ) which is driven by genes related to neurons and synapsis . In the t-SNE results these three tissues simply cluster separately into visually distinct groups , with no indication that their expression profiles have something in common ( Fig 2 ( b ) ) . Thus , although we find the t-SNE results visually attractive , this 2-dimensional projection contains less information than the Structure plot from the GoM ( Fig 1 ( a ) ) , which uses color to represent the samples in a 20-dimensional space . In addition to these qualitative comparisons with other methods , we also used the GTEx data to quantitatively compare the accuracy of the GoM model with hierarchical clustering . Specifically , for each pair of tissues in the GTEx data we assessed whether or not each method correctly partitioned samples into the two tissue groups; see Methods . ( Other methods do not provide an explicit clustering of the samples—only a visual representation—and so are not included in these comparisons . ) The GoM model was more accurate in this test , succeeding in 88% of comparisons , compared with 79% for hierarchical clustering ( S3 ( a ) vs S3 ( c ) Fig ) . Although the analysis of all tissues is useful for assessing global structure , it may miss finer-scale structure within tissues or among similar tissues . For example , here the GoM model applied to all tissues effectively allocated only three clusters to all brain tissues ( clusters 1 , 2 and 9 in Fig 1 ( a ) ) , and we suspected that additional substructure might be uncovered by analyzing the brain samples separately and using more clusters . Fig 1 ( b ) shows the Structure plot for K = 6 on only the Brain samples . The results highlight much finer-scale structure compared with the global analysis; see Table 2 . Brain Cerebellum and Cerebellar hemisphere are essentially assigned to a separate cluster ( lime green ) , which is enriched with genes related to cell periphery and communication ( e . g . PKD1 , CBLN3 ) as well as genes expressed largely in neuronal cells and playing a role in neuron differentiation ( e . g . CHGB ) . The spinal cord samples also show consistently strong membership in a single cluster ( yellow-orange ) , the top defining gene for the cluster being MBP which is involved in myelination of nerves in the nervous system [27] . Another driving gene , GFAP , participates in system development by acting as a marker to distinguish astrocytes during development [28] . The remaining samples all show membership in multiple clusters . Samples from the putamen , caudate and nucleus accumbens show similar profiles , and are distinguished by strong membership in a cluster ( cluster 4 , bright red ) whose top driving gene is PPP1R1B , a target for dopamine . And cortex samples are distinguished from others by stronger membership in a cluster ( cluster 2 , turquoise in Fig 1 ( b ) ) whose distinctive genes include ENC1 , which interacts with actin and contributes to the organisation of the cytoskeleton during the specification of neural fate [29] . In comparison , applying PCA , MDS , hierarchical clustering and t-SNE to these brain samples reveals less of this finer-scale structure ( S4 Fig ) . Both PCA and MDS effectively cluster the samples into two groups—those related to the cerebellum vs everything else . Hierarchical clustering also separates out the cerebellum-related tissues from the others , but again the format seems ill-suited to static visualization of more than one thousand samples . For reasons that we do not understand t-SNE performs poorly for these data: many samples are allocated to essentially identical locations , and so overplotting obscures them . Recently RNA-sequencing has become viable for single cells [30] , and this technology has the promise to revolutionize understanding of intra-cellular variation in expression , and regulation more generally [31] . Although it is traditional to describe and categorize cells in terms of distinct cell-types , the actual architecture of cell heterogeneity may be more complex , and in some cases perhaps better captured by the more “continuous” GoM model . In this section we illustrate the potential for the GoM model to be applied to single cell data . To be applicable to single-cell RNA-seq data , methods must be able to deal with lower sequencing depth than in bulk RNA experiments: single-cell RNA-seq data typically involve substantially lower effective sequencing depth compared with bulk experiments , due to the relatively small number of molecules available to sequence in a single cell . Therefore , as a first step towards demonstrating its potential for single cell analysis , we checked robustness of the GoM model to sequencing depth . Specifically , we repeated the analyses above after thinning the GTEx data by a factor of 100 and 10 , 000 to mimic the lower sequencing depth of a typical single cell experiment . For the thinned GTEx data the Structure plot for K = 20 preserves most of the major features of the original analysis on unthinned data ( S5 Fig ) . For the accuracy comparisons with hierarchical clustering , both methods suffer reduced accuracy in thinned data , but the GoM model remains superior ( S6 Fig ) . For example , when thinning by a factor of 10 , 000 , the success rate in separating pairs of tissues is 0 . 32 for the GoM model vs 0 . 10 for hierarchical clustering . Having established its robustness to sequencing depth , we now illustrate the GoM model on two single cell RNA-seq datasets: data on mouse spleen from Jaitin et al [32] and data on mouse preimplantation embryos from Deng et al [33] .
Our goal here is to highlight the potential for GoM models to elucidate structure in RNA-seq data from both single cell sequencing and bulk sequencing of pooled cells . We also provide tools to identify which genes are most distinctively expressed in each cluster , to aid interpretation of results . As our applications illustrate , the results can provide a richer summary of the structure in RNA-seq data than existing widely-used visualization methods such as PCA and hierarchical clustering . While it could be argued that the GoM model results sometimes raise more questions than they answer , this is exactly the point of an exploratory analysis tool: to highlight issues for investigation , identify anomalies , and generate hypotheses for future testing . Our results from different methods also highlight another important point: different methods have different strengths and weaknesses , and can compliment one another as well as competing . For example , t-SNE seems to provide a much clearer indication of the cluster structure in the full GTEx data than does PCA , but does a poorer job of capturing the ordering of the developmental samples from mouse pre-implantation embryos . While we believe the GoM model often provides a richer summary of the sample structure , we would expect to use it in addition to t-SNE and PCA when performing exploratory analyses . ( Indeed the methods can be used in combination: both PCA and t-SNE can be used to visualize the results of the GoM model , as an alternative or complement to the Structure plot . ) A key feature of the GoM model is that it allows that each sample has a proportion of membership in each cluster , rather than a discrete cluster structure . Consequently it can provide insights into how well a particular dataset really fits a “discrete cluster” model . For example , consider the results for the data from Jaitin et al [32] and Deng et al [33]: in both cases most samples are assigned to multiple clusters , although the results are closer to “discrete” for the latter than the former . The GoM model is also better able to represent the situation where there is not really a single clustering of the samples , but where samples may cluster differently at different genes . For example , in the GTEx data , the stomach samples share memberships in common with both the pancreas ( purple ) and the adrenal gland ( light green ) . This pattern can be seen in the Structure plot ( S4 Fig ) but not from other methods like PCA , t-SNE or hierarchical clustering ( Fig 2 ) . Fitting GoM models can be computationally-intensive for large data sets . For the datasets we considered here the computation time ranged from 12 minutes for the data from [33] ( n = 259;K = 6 ) , through 33 minutes for the data from [32] ( n = 1 , 041;K = 7 ) to 3 , 370 minutes for the GTEx data ( n = 8 , 555;K = 20 ) . Computation time can be reduced by fitting the model to only the most highly expressed genes , and we often use this strategy to get quick initial results for a dataset . Because these methods are widely used for clustering very large document datasets there is considerable ongoing interest in computational speed-ups for very large datasets , with “on-line” ( sequential ) approaches capable of dealing with millions of documents [49] that could be useful in the future for very large RNA-seq datasets . A thorny issue that arises when fitting clustering models is how to select the number of clusters , K . Like many software packages for fitting these models , the maptpx package implements a measure of model fit that provides one useful guide . However , it is worth remembering that in practice there is unlikely to be a “true” value of K , and results from different values of K may complement one another rather than merely competing with one another . For example , seeing how the fitted model evolves as K increases is one way to capture some notion of hierarchy in the clusters identified [21] . More generally it is often fruitful to analyse data in multiple ways using the same tool: for example our GTEx analyses illustrate how analysis of subsets of the data ( in this case the brain samples ) can complement analyses of the entire data . Finally , as a practical matter , we note that Structure plots can be difficult to read for large K ( e . g . K = 30 ) because of the difficulties of choosing a palette with K distinguishable colors . The version of the GoM model fitted here is relatively simple , and could certainly be embellished . For example , the model allows the expression of each gene in each cluster to be a free parameter , whereas we might expect expression of most genes to be “similar” across clusters . This is analogous to the idea in population genetics applications that allele frequencies in different populations may be similar to one another [50] , or in document clustering applications that most words may not differ appreciably in frequency in different topics . In population genetics applications incorporating this idea into the model , by using a correlated prior distribution on these frequencies , can help improve identification of subtle structure [50] and we would expect the same to happen here for RNA-seq data . Finally , GoM models can be viewed as one of a larger class of “matrix factorization” approaches to understanding structure in data , which also includes PCA , non-negative matrix factorization ( NMF ) , and sparse factor analysis ( SFA ) ; see [51] . This observation raises the question of whether methods like SFA might be useful for the kinds of analyses we performed here . ( NMF is so closely related to the GoM model that we do not discuss it further; indeed , the GoM model is a type of NMF , because both grades of membership and expression levels within each cluster are required to be non-negative . ) Informally , SFA can be thought of as a generalization of the GoM model that allows samples to have negative memberships in some “clusters” ( actually , “factors” ) . This additional flexibility should allow SFA to capture certain patterns more easily than the GoM model . For example , a small subset of genes that are over-expressed in some samples and under-expressed in other samples could be captured by a single sparse factor , with positive loadings in the over-expressed samples and negative loadings in the other samples . However , this additional flexibility also comes at a cost of additional complexity in visualizing the results . For example , S10 , S11 and S12 Figs show results of SFA ( the version from [51] ) for the GTEx data and the mouse preimplantation data: in our opinion , these do not have the simplicity and immediate visual appeal of the GoM model results . Also , applying SFA to RNA-seq data requires several decisions to be made that can greatly impact the results: what transformation of the data to use; what method to induce sparsity ( there are many; e . g . [51–54] ) ; whether to induce sparsity in loadings , factors , or both; etc . Nonetheless , we certainly view SFA as complementing the GoM model as a promising tool for investigating the structure of RNA-seq data , and as a promising area for further work .
We use the maptpx R package [20] to fit the GoM models ( 1 and 2 ) , which is also known as “Latent Dirichlet Allocation” ( LDA ) . The maptpx package fits this model using an EM algorithm to perform Maximum a posteriori ( MAP ) estimation of the parameters q and θ . See [20] for details . In addition to the Structure plot , we have also found it useful to visualize results using t-distributed Stochastic Neighbor Embedding ( t-SNE ) , which is a method for visualizing high dimensional datasets by placing them in a two dimensional space , attempting to preserve the relative distance between nearby samples [24 , 25] . Compared with the Structure plot our t-SNE plots contain less information , but can better emphasize clustering of samples that have similar membership proportions in many clusters . Specifically , t-SNE tends to place samples with similar membership proportions together in the two-dimensional plot , forming visual “clusters” that can be identified by eye ( e . g . http://stephenslab . github . io/count-clustering/project/src/tissues_tSNE_2 . html ) . This may be particularly helpful in settings where no external information is available to aid in making an informative Structure plot . To help biologically interpret the clusters , we developed a method to identify which genes are most distinctively differentially expressed in each cluster . ( This is analogous to identifying “ancestry informative markers” in population genetics applications [55] . ) Specifically , for each cluster k we measure the distinctiveness of gene g with respect to any other cluster l using KL g [ k , l ] : = θ k g l o g θ k g θ l g + θ l g - θ k g , ( 3 ) which is the Kullback–Leibler divergence of the Poisson distribution with parameter θkg to the Poisson distribution with parameter θlg . For each cluster k , we then define the distinctiveness of gene g as D g [ k ] = min l ≠ k KL g [ k , l ] . ( 4 ) The higher Dg[k] , the larger the role of gene g in distinguishing cluster k from all other clusters . Thus , for each cluster k we identify the genes with highest Dg[k] as the genes driving the cluster k . We annotate the biological functions of these individual genes using the mygene R Bioconductor package [56] . For each cluster k , we filter out a number of genes ( top 100 for the Deng et al data [33] and GTEx V6 data [57] ) with highest Dg[k] value and perform a gene set over-representation analysis of these genes against all the other genes in the data representing the background . To do this , we used ConsensusPathDB database ( http://cpdb . molgen . mpg . de/ ) [58] [59] . See Tables 1 , 2 and 3 for the top significant gene ontologies driving each cluster in the GTEx V6 data and the Deng et al data respectively . We compared the GoM model with a distance-based hierarchical clustering algorithm by applying both methods to samples from pairs of tissues from the GTEx project , and assessed their accuracy in separating samples according to tissue . For each pair of tissues we randomly selected 50 samples from the pool of all samples coming from these tissues . For the hierarchical clustering approach we cut the dendrogram at K = 2 , and checked whether or not this cut partitions the samples into the two tissue groups . ( We applied hierarchical clustering using Euclidean distance , with both complete and average linkage; results were similar and so we showed results only for complete linkage . ) For the GoM model we analysed the data with K = 2 , and sorted the samples by their membership in cluster 1 . We then partitioned the samples at the point of the steepest fall in this membership , and again we checked whether this cut partitions the samples into the two tissue groups . S3 Fig shows , for each pair of tissues , whether each method successfully partitioned the samples into the two tissue groups . We used “thinning” to simulate lower-coverage data from the original higher-coverage data . Specifically , if cng is the counts of number of reads mapping to gene g for sample n for the original data , we simulated thinned counts tng using t n g ∼ B i n ( c n g , p t h i n ) ( 5 ) where pthin is a specified thinning parameter . Our methods are implemented in an R package CountClust , available as part of the Bioconductor project at https://www . bioconductor . org/packages/3 . 3/bioc/html/CountClust . html . The development version of the package is also available at https://github . com/kkdey/CountClust . Code for reproducing results reported here is available at http://stephenslab . github . io/count-clustering/ .
|
Gene expression profile of a biological sample ( either from single cells or pooled cells ) results from a complex interplay of multiple related biological processes . Consequently , for example , distal tissue samples may share a similar gene expression profile through some common underlying biological processes . Our goal here is to illustrate that grade of membership ( GoM ) models—an approach widely used in population genetics to cluster admixed individuals who have ancestry from multiple populations—provide an attractive approach for clustering biological samples of RNA sequencing data . The GoM model allows each biological sample to have partial memberships in multiple biologically-distinct clusters , in contrast to traditional clustering methods that partition samples into distinct subgroups . We also provide methods for identifying genes that are distinctively expressed in each cluster to help biologically interpret the results . Applied to a dataset of 53 human tissues , the GoM approach highlights similarities among biologically-related tissues and identifies distinctively-expressed genes that recapitulate known biology . Applied to gene expression data of single cells from mouse preimplantation embryos , the approach highlights both discrete and continuous variation through early embryonic development stages , and genes involved in a variety of relevant processes . Our study highlights the potential of GoM models for elucidating biological structure in RNA-seq gene expression data .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"blastocysts",
"sequencing",
"techniques",
"population",
"genetics",
"neuronal",
"differentiation",
"neuroscience",
"cell",
"differentiation",
"developmental",
"biology",
"molecular",
"biology",
"techniques",
"population",
"biology",
"embryos",
"rna",
"sequencing",
"research",
"and",
"analysis",
"methods",
"embryology",
"rna",
"structure",
"animal",
"cells",
"gene",
"expression",
"molecular",
"biology",
"biochemistry",
"rna",
"cellular",
"neuroscience",
"cell",
"biology",
"nucleic",
"acids",
"neurons",
"genetics",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"evolutionary",
"biology",
"macromolecular",
"structure",
"analysis"
] |
2017
|
Visualizing the structure of RNA-seq expression data using grade of membership models
|
The intestinal epithelium plays a critical role in host-microbe homeostasis by sensing gut microbes and subsequently initiating proper immune responses . During the neonatal stage , the intestinal epithelium is under immune repression , allowing the transition for newborns from a relatively sterile intra-uterine environment to one that is rich in foreign antigens . The mechanism underlying such immune repression remains largely unclear , but involves downregulation of IRAK1 ( interleukin-1 receptor-associated kinase ) , an essential component of toll-like receptor-mediated NF-κB signaling . We report here that heterogeneous nuclear ribonucleoprotein I ( hnRNPI ) , an RNA binding protein , is essential for regulating neonatal immune adaptation . We generated a mouse model in which hnRNPI is ablated specifically in the intestinal epithelial cells , and characterized intestinal defects in the knockout mice . We found that loss of hnRNPI function in mouse intestinal epithelial cells results in early onset of spontaneous colitis followed by development of invasive colorectal cancer . Strikingly , the epithelium-specific hnRNPI knockout neonates contain aberrantly high IRAK1 protein levels in the colons and fail to develop immune tolerance to environmental microbes . Our results demonstrate that hnRNPI plays a critical role in establishing neonatal immune adaptation and preventing colitis and colorectal cancer .
Increasing evidence indicates that proper host-microbe interaction in the gastrointestinal tract is critical for the balance of immune tolerance and active immune responses [1 , 2] . Dysregulated host response to gut microbiota is the major cause of autoimmune diseases , inflammatory disorders and cancer [1 , 3–6] . The intestinal epithelium , which lines the gastrointestinal tract , plays a fundamental role in controlling the host-microbe interaction [7] . Structurally , the intestinal epithelium acts as a physical barrier to separate luminal contents from immune cells situated in the lamina propria . A mucus layer formed by goblet cells covers the intestinal epithelium and protects it from direct attack of foreign antigens [8] . Moreover , junction complexes located between the intestinal epithelial cells ( IECs ) control the paracellular permeability of the intestinal epithelium , which is critical to prevent the invasion of pathogens and other luminal contents across the epithelial layer [9 , 10] . In addition to acting as a physical barrier , IECs play an active role in immune defense by expressing a variety of molecules that recognize and subsequently kill pathogens , and initiating the innate and adaptive immune responses . One of the most important pathways that act in IECs is Toll-like receptor ( TLR ) -mediated NF-κB signaling . A number of TLRs are expressed in the IECs [11] . Upon binding with their ligands , which are the conserved molecular motifs on microorganisms , TLRs activate a series of downstream signaling cascades and subsequently activate NF-κB signaling [11–13] . A key event in transmitting signals from TLRs to NF-κB signaling is IRAK1-induced degradation of IκB , the cytosolic inhibitors of NF-κB signaling . This consequently releases NF-κB subunits from a cytoplasmic inhibitory complex , which allows them to translocate into the nucleus to induce transcription of pro-inflammatory genes . Prolonged or excessive TLR-mediated NF-κB signaling activation is a major cause of inflammatory disorders and inflammatory bowel disease-associated colorectal cancer [14–17] . Thus , understanding mechanisms by which TLR -mediated NF-κB signaling is precisely controlled in the IECs is critical for elucidating the etiology of gastrointestinal inflammatory disorders and its associated cancers . Recent studies indicate that TLR -mediated NF-κB signaling is suppressed in the intestinal epithelium during the neonatal stage [18 , 19] . Upon birth , newborns undergo a transition from a sterile intra-uterine environment to one that is rich in environmental microbes . To accommodate the colonization of the commensal intestinal microorganisms , the intestinal epithelium of the newborn undergoes a series of dynamic changes in gene expression to suppress the TLR -mediated NF-κB signaling activity . One of the most important events is downregulation of the IRAK1 protein level in the IECs shortly after birth [18] . This downregulation is at least partly through TLR4 signaling-mediated continuous proteolytic degradation of IRAK1 during the neonatal stage [19] . A low level of IRAK1 protein in IECs is essential for inhibiting excessive immune response to newly arrived gut microbes and facilitating microbe colonization in the neonate [18 , 19] . Recent studies show that miR-146a is essential to maintain the low level of IRAK1 protein in the neonatal IECs [19] . However , the detailed molecular mechanism by which IRAK1 is downregulated by miR-146a remains elusive . In addition , it is unclear whether other inhibitory mechanisms are involved during neonatal immune adaption . We previously reported that hnRNPI , an RNA binding protein , is an important regulator of intestinal epithelium renewal and calcium-mediated egg activation in zebrafish [20 , 21] . hnRNPI , also known as polypyrimidine tract-binding protein ( PTB ) , plays important roles in alternative splicing and other post-transcriptional regulatory events [22 , 23] . A number of hnRNPI targets are abnormally spliced in intestinal inflammatory and neoplastic diseases [24–31] , suggesting that hnRNPI-dependent post-transcriptional control may play important roles in pathogenesis of these diseases . To determine hnRNPI functions in mammalian intestinal homeostasis and more importantly , to understand how malfunction of this protein contributes to inflammation and colorectal cancer , we have generated IEC-specific hnRNPI knockout mice . We show here that ablation of hnRNPI in the IECs induces spontaneous colitis in mice followed by development of invasive colorectal cancer at a young age . We further show that inflammation occurs shortly after birth in the knockout neonate , which is accompanied by hyperactive NF-κB signaling in the colonic epithelial cells . We provide evidence that downregulation of IRAK1 protein expression is disrupted in the knockout neonatal colon , whereas expression levels of TLRs remain unaffected . Thus , our results reveal a novel role of hnRNPI in establishing neonatal immune adaptation , which is at least partly through the control of the IRAK1 protein level .
Our previous studies in the zebrafish hnRNPI mutant indicate that hnRNPI plays a key role in balancing IEC proliferation and differentiation [20] . To determine the functions of hnRNPI in mammalian intestinal homeostasis , we examined its expression in the mouse intestine . We found that hnRNPI protein is highly accumulated in the nuclei of IECs as well as cells situated in the lamina propria ( Fig 1D and 1E , S2–S5 Figs ) . To determine the role of hnRNPI in the mammalian IECs , we generated a floxed mouse allele of hnRNPI , in which two loxP sites flank the DNA region of exon 3 to exon 8 of the hnRNPI locus ( Fig 1A and 1B ) . This allows deletion of the three most abundant isoforms of hnRNPI upon the Cre recombination [32] . By breeding the hnRNPI floxed allele with a Cre line controlled by the villin promoter [33] , we generated the hnRNPIflox/flox; VillinCre/+ ( hereafter IEC-specific hnRNPI knockout ) mice . The villin promoter directs expression of the Cre recombinase in the IECs as early as at embryonic day 12 . 5 [33] , which allows epithelium-specific deletion of hnRNPI at late embryogenesis in the knockout mice . As expected , the expression of hnRNPI protein was dramatically downregulated in the IECs of the hnRNPI knockout mice , but not that of the control mice ( Fig 1C–1E ) . The IEC-specific hnRNPI knockout mice were born at the Mendelian ratio , but appeared smaller than their littermates ( Fig 2A ) . Their body weight at weaning is significantly less than that of their wild-type littermates ( Fig 2E ) . Severely affected mutants , which weighed 50% less than wild-type littermates , died within three days after weaning , likely due to their malfunction in digesting solid food ( mouse labeled as KO2 in Fig 2A is representative ) . Among the remaining knockout mice , over 60% of them developed rectal prolapse within 80 days after birth ( Fig 2B–2D , n = 41 ) . We performed histological analysis on the 3-week to 12-week old knockout mice and found that 100% of them developed moderate to severe degree of inflammation in the colonic epithelium and their colons often appeared shortened ( Fig 2C , n = 26 , 13 of them < 4 weeks ) . There is no gender-based difference in the development of colon inflammation in these mice . Histological features of the inflamed colonic epithelium in the knockout mice include crypt elongation and abscesses , loss of goblet cells , inflammatory cell infiltrate , and impaired surface integrity ( Fig 3A , 3B , 3E and 3F ) . This is accompanied by hyperproliferation of IECs ( Fig 3C and 3D ) . Large numbers of infiltrated inflammatory cells including Ly6G positive neutrophils , F4/80 positive macrophages , and CD4 positive T cells , were detected in the lamina propria ( Fig 3G–3L ) . These phenotypes highly resemble the pathological features of human ulcerative colitis . Mice with a lower body weight displayed more severe colitis . While the colitis phenotype was observed in all parts of colon in the knockout mice , epithelium in the distal colon is more severely affected . In contrast , the epithelium of the small intestine did not display histologically detectable inflammation in these mice . In line with the morphological appearance of colitis in the knockout mice , the expression of proinflammatory cytokines and chemokines , including IL6 , IL1β , Cxcl1 , and Ccl2 , are dramatically increased in the colonic epithelial cells of the knockout mice ( Fig 3M ) . Thus , IEC-specific depletion of hnRNPI results in early onset of spontaneous colitis . Intriguingly , we observed multiple adenomatous lesions in the colonic epithelium of the knockout mice . We analyzed total 30 mice aged between P22 to P230 , and found 60% of them developed colon adenomas at variable degrees of dysplasia . Among them , the youngest mice that had adenomas were at the age of P23 . These lesions display hyperproliferation of colonic epithelial cells ( Fig 4A and 4B ) . Nuclear accumulation of β-catenin and p65 , the respective hallmarks of active Wnt signaling and NF-κB signaling , is prominent in the lesions ( Fig 4C–4H ) . This indicates that colon adenomatous lesions in these mice are in the precancerous condition . All adenomatous lesions in the colonic epithelium , however , are restricted to the mucosa ( Fig 4A ) . In striking contrast , we found that lesions developed in the epithelium of the prolapsed rectum were highly invasive , and had spread through the muscularis mucosae into the submucosa ( compare Fig 4J to 4I ) . We examined total 17 knockout mice aged between P42 to P230 that developed rectal prolapse , and found 15 of them ( 88% ) developed invasive adenocarcinomas in the rectal epithelium . Among those with rectal adenocarcinomas , the youngest knockout mouse was at the age of P50 . Similar to the lesions in the colon , these rectal carcinomatous lesions contain highly active Wnt signaling and NF-κB signaling ( Fig 4K–4R ) . To determine when colitis development in the IEC-specific hnRNPI knockout mice is initiated , we performed histological analysis on the colonic epithelium from the knockout mice and their sibling littermates at P7 and P14 . Severe inflammation was observed in the colonic epithelium of the knockout mice at both P7 and P14 ( the histology of wild-type and knockout colon at P14 is shown in Fig 5A and 5B respectively ) . At P14 , the colonic epithelium of the knockout mice displayed impaired intestinal epithelium junctional complexes as shown by the zonula occludens ( ZO ) -1 staining ( compare Fig 5F to 5E ) . Consistently , we detected bacteria infiltration in the mutant intestinal epithelium ( compare Fig 5H to 5G ) , indicating destruction of the epithelial barrier . Furthermore , a large number of infiltrated innate and adaptive immune cells were seen in the lamina propria ( Fig 5I–5N ) . These defects are accompanied by hyperproliferation of IECs ( compare Fig 5D to 5C ) . We further examined the colonic epithelium during the first week of life . While the colonic epithelium from the knockout neonates appear histologically normal at P0 , P1 and P2 ( the colon histology of P2 neonates is shown in S1 Fig ) , inflamed colonic epithelium was readily observed in the knockout neonates at P3 ( compare Fig 6B to 6A ) . Significant increase in the numbers of proliferating epithelial cells and inflammatory cells was detected in the knockout neonates at P3 ( Fig 6C–6H ) . Consistent with the histological observation , the expression levels of proinflammatory cytokines and chemokines including IL6 , IL1β , TNFα and Cxcl2 are dramatically upregulated in the colon of the knockout mice at P3 , while those at P0 and P1 remain unchanged ( Fig 6I ) . At P2 , a slight increase in the expression of Cxcl2 , IL1β , and TNFα was detected in the knockout colon ( Fig 6I ) . It appears that the immune response was initiated at the molecular level at P2 in these mutants . Interestingly , we observed a tight correlation of colonic inflammation onset and decline in weight gain in the knockout neonates . As shown in Fig 6J , a sharp decline in the weight gain occurred in the knockout neonates at P3 , a time point when colonic inflammation was first observed histologically . In IECs , TLRs mediated NF-κB signaling plays an essential role in sensing luminal bacteria and initiating subsequent immune reactions [11] . During the neonatal stage , activity of NF-κB signaling is suppressed transiently , allowing microbe colonization and development of immune tolerance . This occurs at least in part through downregulating the expression of IRAK1 protein in the IECs [18 , 19] . Given the aforementioned hyperactive inflammatory responses in the knockout neonatal colon , we examined the activity of NF-κB signaling in the neonatal colonic epithelium by assessing the cellular localization of the NF-κB subunit p65 . Indeed , whereas p65 is localized to the cytoplasm of colonic epithelial cells in the wild-type control mice , nuclear translocation of p65 is prominent in the knockout colon ( compare the top and bottom panels in Fig 7A ) . This observation prompted us to determine whether hyperactive NF-κB signaling in the colonic epithelial cells of the knockout neonates is due to altered expression levels of TLRs and/or IRAK1 . We first examined mRNA expression levels of TLR2 , 4 , and 5 , three TLRs that are known to be expressed in the mouse colon [11] . No increase in mRNA expression levels of TLR2 , 4 , and 5 was detected in the knockout mice at both P0 and P4 ( Fig 7B ) . In striking contrast , increased protein expression of IRAK1 was detected in the colon of the knockout neonates at P0 ( Fig 7C ) . This increase was more dramatic by P3 ( Fig 7C ) . Notably , IRAK1 protein expression in the knockout fetuses remains unchanged ( Fig 7C ) , indicating that hnRNPI is required to downregulate IRAK1 protein level postnatally . Interestingly , we found that the mRNA expression level of IRAK1 was not upregulated in the knockout colon at both P0 and P3 ( Fig 7D ) . This suggests that hnRNPI regulates IRAK1 at the post-transcriptional level . Taken together , the above observations demonstrate that hnRNPI plays an essential role in downregulating expression of IRAK1 protein in the neonatal colon and is essential for neonatal immune adaptation .
Precisely controlled host-microbe interactions are crucial for human overall health and well-being . Neonatal immune adaptation is the first and fundamentally important step in establishing host- microbe homeostasis . During the neonatal stage , the innate immune activity in the digestive tract must be temporally suppressed to accommodate the large number of newly arrived microbes . Recent studies show that this temporal suppression is at least in part through downregulating the expression of IRAK1 protein in the IECs upon birth , a process that requires the presence of miR-146a in the neonatal IECs [19] . The mechanism by which miR-146a downregulates IRAK1 is unknown . It is also unclear if other mechanisms are involved in this process during neonatal immune adaptation . Here , we report that deletion of hnRNPI in the IECs impairs downregulation of IRAK1 in the neonatal colon . We show that the expression level of IRAK1 protein , but not its mRNA , is upregulated in the neonatal colon upon deletion of hnRNPI in the IECs , suggesting that hnRNPI-mediated IRAK1 downreulation occurs at the post-transcriptional level . Interestingly , IEC-specific deletion of hnRNPI does not affect the protein level of IRAK1 in the fetal colon . These findings are consistent with the recent observation that IRAK1 is downregulated in the neonatal intestine through post-transcriptional regulation , and this process requires microbial stimulation at birth and postnatally [18 , 19] . In line with the finding that deletion of hnRNPI increases IRAK1 expression in the neonatal colon , we found that NF-κB signaling is highly active in the neonatal colon of the mutant mice . This is accompanied by the induction of colonic inflammation in the knockout neonates , which becomes detectable histologically and molecularly within the first three days after birth . The timing of colon inflammatory response is coincident with the transition of neonates from a sterile intra-uterine environment to one that is rich in foreign antigens , suggesting that mutant neonates fail to develop immune tolerance . We also observed a significant decline in weight gain in the hnRNPI mutant neonates at P3 , which is concomitant to the induction of colon inflammation . It is highly likely that the slow weight gain in the mutant neonates is caused by malnutrition in these mice due to the impaired host-microbe interactions . Collectively , these findings uncover an important function of hnRNPI in suppressing the expression of IRAK1 protein in the neonatal colon and establishing host-microbe homeostasis upon birth in the intestine . Mechanistically , how does hnRNPI regulate the expression of IRAK1 ? Several splicing variants of IRAK1 with variable stability and activity in mediating TLR-induced NF-κB signaling have been identified in mice and humans [34–37] . It is tempting to speculate that hnRNPI may down-regulate IRAK1 through regulating alternative splicing of IRAK1 or its upstream regulators . Alternatively , hnRNPI , which has the ability to regulate translational efficiency through binding 3’ UTR of its targets [23 , 38] , may directly repress IRAK1 translation , or indirectly alter the translation of its regulators . Interestingly , the 3’ UTR of IRAK1 contains multiple sites resembling the consensus hnRNPI binding sequences . It will be of great interest to determine if hnRNPI physically interacts with IRAK1 3’UTR and regulates IRAK1 translation . Down-regulation of IRAK1 in the neonatal intestine requires continuous proteasome or lysosome-dependent proteolytic degradation [19] . Thus , it is also possible that hnRNPI may regulate the expression of proteins that alter IRAK1 protein turnover . Of note , hnRNPI is capable of modulating the activity of microRNAs in disease pathogenesis and many important biological processes [30 , 39–42] . It has been reported that miR-146a controls both translation and degradation of epithelial IRAK1 in the neonatal intestine [19] . It would be interesting to determine if hnRNPI regulates the biosynthesis or activity of miR-146a and/or other microRNAs in establishing neonatal immune tolerance . Further studies are required to distinguish these possibilities . Our results reveal that hnRNPI-deficient mice develop invasive colorectal cancer at a very young age ( as early as at P50 ) . This observation is consistent with the findings that hnRNPI is aberrantly expressed in colorectal cancer cells [24 , 31] , and a number of hnRNPI targeting genes are abnormally spliced in colorectal cancer [24–30] . While it is possible that the colorectal cancer development in the hnRNPI-deficient mice may be a consequence of impaired neonatal host-microbe homeostasis , it is more likely that hnRNPI plays additional roles in preventing colitis and colorectal cancer development in adulthood . In agreement with this view , we found that Wnt signaling , a major driver of colorectal cancer , is hyper-active in the hnRNPI-deficient colonic epithelial cells . It has been reported that Wnt ligands are expressed in the IECs and intestinal stromal cells [43 , 44] . Stromal cells-derived Wnts , but not epithelial cells-produced Wnts , are indispensable for intestinal homeostasis [45] . We thus assessed the expression of six Wnts that are expressed in the colonic stroma [43 , 44] . These include Wnt2b/Wnt4/Wnt5a , which are highly expressed in the colon mesenchyme , and Wnt5b/Wnt10b/Wnt16 that are expressed at low levels [43 , 44] . We observed a trend of increase in the expression of Wnt2b in the hnRNPI-deficient mice ( S6 Fig ) . However , this increase is not statistically significant . We did not detect any statistically significant changes in the expression of Wnt4 , Wnt5a , and Wnt5b ( S6 Fig ) . In the case of Wnt10b and Wnt16 , the expression was decreased ( S6 Fig ) . Currently we do not understand the significance of the decrease in the expression of Wnt10b and Wnt16 . Nonetheless , these results seem to suggest that hnRNPI suppresses Wnt signaling in IECs through a mechanism independent of downregulating stromal Wnt ligands . As an RNA-binding protein , hnRNPI exerts its function by controlling post-transcriptional events . Its effects on signaling pathways are highly context- and species- specific . hnRNPI inhibits Notch signaling in Drosophila wing disc [46] and during zebrafish intestinal homeostasis [20] . In mouse IECs , rather than inhibiting Notch signaling ( S7 Fig ) , hnRNPI suppresses NF-κB and Wnt signaling . In the future , it will be of great interest to identify direct targets of hnRNPI and investigate the detailed molecular mechanisms by which hnRNPI influences major signaling pathways . In summary , we report for the first time that hnRNPI-mediated post-transcriptional regulation is fundamentally important for establishing neonatal immune tolerance . The IEC-specific hnRNPI knockout mice represent a valuable animal model for studying regulatory mechanisms governing the establishment of neonatal immune tolerance at the post-transcriptional level .
Generation of the hnRNPIflox/flox; VillinCre/+ mice: hnRNPI targeted ES cells ( KOMP Repository ) were used for blastocyst injection ( performed by the Transgenic Core Facility at the Research Institute at the Nationwide Children’s Hospital ) . Male chimeras were bred with wild type C57BL6 females for germline transmission . To obtain hnRNPIflox mice , germline transmitted mice were bred with the ACT-FLPe ( the Jackson Laboratory ) mice to delete the neo-cassette . The hnRNPIflox mice were crossed with the villin-cre mice ( gift from Dr . Noah Shoyer ) to generate the hnRNPIflox/flox; VillinCre/+ mice . Primers used for genotyping the hnRNPI floxed allele are: F1: 5’–CCCATAACTGTCCATAGACC -3’ , and B1: 5’ -TGTTGGTAATGCCAGCACAG -3’ . All mice with one exception used in this report are from the cross of the hnRNPIflox/flox; VillinCre/+ mice with the hnRNPIflox/flox mice . The hnRNPIflox/flox; VillinCre/+ mice were used in the knockout group and the hnRNPIflox/flox mice were used in the control group . An exception to this is the mice used in the adult weight statistical analysis . These mice were derived from the cross of the hnRNPIflox/+; VillinCre/+ mice with the hnRNPIflox/flox mice . In this experiment , the wild-type group includes the hnRNPIflox/+ mice and the hnRNPIflox/flox mice , heterozygotes are the hnRNPIflox/+; VillinCre/+ mice , and the knockout mice are the hnRNPIflox/flox; VillinCre/+ mice . Colons were isolated , fixed , paraffin-embedded , and sectioned according to standard protocols . Intestine sections ( 5 μm ) were processed for hematoxylin and eosin staining or for immunostaining . Immunohistochemistry was performed with R . T . U . vectastain kit ( Vector Laboratories ) with DAB substrate . Sections were counterstained lightly with Hematoxylin afterwards . For immunofluorescence staining , secondary antibodies used are goat anti-rabbit AlexaFluor 488 and donkey anti-rat AlexaFluor 594 ( Invitrogen ) . Sections were counterstained with 4’ , 6-diamidino-2-phenylindole ( DAPI ) . Primary antibodies used are: mouse anti-ki67 ( BD Pharmingen , 550609 ) , rat anti-CD4 ( Ebioscience Inc , 14-9766-80 ) , rat anti-Ly6G ( BD Pharmingen , 551459 ) , rabbit anti-F4/80 ( Novus Biologicals Inc , NBP2-12506 ) , rat anti-ZO-1 ( Developmental Studies Hybridoma Bank , R26 . 4C ) , rabbit anti-hnRNPI ( gift from Dr . Douglas Black ) , rabbit anti-p65 ( Santa Cruz Biotechnology , sc-372 ) , rabbit anti-p65 ( Cell signaling , 8242 ) , rabbit anti- β-catenin ( gift from Dr . Peter Klein ) . Goblet cell secreted mucins were identified by sequentially incubating deparaffinized sections in pH 2 . 5 alcian blue ( 1 hour ) , periodic acid ( 7 minutes ) and Schiff's reagent ( 10 minutes ) . After the staining , acidic mucins are stained “blue” and neutral mucins are stained red . Fluorescence in situ Hybridization ( FISH ) was performed to detect eubacteria infiltration in the colon . Paraffin sections ( 10 μm ) were dewaxed and incubated with the commercially synthesized universal eubacterial probe EUB 338 ( 5′-GCTGCCTCCCGTAGGAGT-3′ ) conjugated with Alexa Fluor 488 as described [47] . A complimentary probe ( 5′-ACTCCTACGGGAGGCAGC-3′ ) ) conjugated with Alexa Fluor 488 was used as a negative control . Sections were counterstained with DAPI afterwards . Images were taken from a Compound microscope ( Leica ) with digital camera or a Nikon A1R confocal microscope and processed using Adobe Photoshop . Colonic epithelial cells were isolated as described [48] . RNAs were extracted from colonic epithelial cells isolated from adult mice and the postnatal day 4 ( P4 ) neonatal mice or whole colon tissues from the P0 , P1 , P2 , and P3 mice . RNA extraction was done using TRIzol reagent according to standard protocols . Real-time PCR reactions were performed blindly in triplicate or duplicate using SYBR green master mix ( Applied Biosystem ) on an Applied Biosystem's 7500 Real-time PCR system . PCR primers are: Il6: 5′- CCGGAGAGGAGACTTCACAG -3′ and 5′- CAGAATTGCCATTGCACAAC -3′; Il1β: 5′-CAACCAACAAGTGATATTCTCCATG-3′ and 5′-GATCCACACTCTCCAGCTGCA-3′; Cxcl2/MIP-2: 5′- GTGAACTGCGCTGTCAATGC -3′ and 5′- GCTTCAGGGTCAAGGCAAAC -3′; Tnfα: 5′-AGGGATGAGAAGTTCCCAAATG-3′ and 5′-TGTGAGGGTCTGGGCCATA-3′; Ccl2: 5′-AGGTCCCTGTCATGCTTCTG-3′ and 5′-TCTGGACCCATTCCTTCTTG-3’; Cxcl1: 5′-GCCAATGAGCTGCGCTGTCAATGC-3′ and 5′-CTTGGGGACACCCTTTTAGCATCTT-3’; Tlr2: 5′-GCTACCTGTGTGACTCTCCG-3′ and 5′- CGCCCACATCATTCTCAGGT-3′; Tlr4: 5′-GCTTTCACCTCTGCCTTCAC-3′ and 5′-AGGCGATACAATTCCACCTG-3′; Tlr5: 5′-CCAGCCCCGTGTTGGTAATA-3′ and 5′-TTTCTGAAAGCCCCTGGACC-3′; IRAK1: 5′-GGCTCAACTAGCTTGCTGCT-3′ and 5′-TAGTGCCTCCCTGGGTACAG-3′; and Gapdh: 5′-TTCTTGTGCAGTGCCAGCC-3′ and 5′-CACCGACCTTCACCATTTTGT-3′ . Isolated colonic epithelial cells from adult mice or whole colon tissues from the fetuses at embryonic day 19 or neonates at P0 or P3 were homogenized in lysis buffer . Protein lysates were cleared by spinning the samples twice at 4°C . Subsequently , samples were separated on SDS-PAGE and analyzed by western blotting as described [49] . Primary antibodies used are mouse anti-hnRNPI ( Life Technologies , 324800 ) , rabbit anti-IRAK1 ( Santa Cruz , sc-7883 ) , rabbit anti-Actin ( Sigma , A2066 ) . Membranes were incubated with HRP-linked secondary antibodies and developed using ECL prime ( G&E Healthcare Life Sciences ) . Differences between the knockout mice and the control groups were assessed for significance using a one-tailed unpaired Student t-test ( Fig 3M ) . For data involving two variables , data were analyzed by two-way ANOVA using GraphPad Prism ( Figs 6I , 6J , 7B and 7D ) or R ( Fig 2E ) . Log2 conversion was used in the figures where necessary ( Figs 3M and 6I ) . The use of mice in this research was approved by University of Illinois at Urbana-Champaign Animal Care and Use Committee ( protocol #14240 and 14290 ) .
|
Precisely controlled host-microbe interactions in the gastrointestinal tract are crucial for human overall health and well-being . Dysregulated host responses to gut microbiota are the major cause of autoimmune diseases , inflammatory disorders and cancers . The intestinal epithelium lines the gastrointestinal tract and plays a critical role in sensing gut microbes and subsequently developing a balance of immune tolerance and active immune responses . During the neonatal stage , the immune system in the gastrointestinal tract must be temporally suppressed to accommodate the large number of newly arrived microbes . This process is known as neonatal immune adaptation , and is critical for the establishment of proper host- microbe interactions . We studied the function of hnRNPI in the intestinal epithelium by genetically ablating it in the intestinal epithelial cells of mouse . We found that loss of hnRNPI in intestinal epithelial cells disrupts neonatal immune adaptation , resulting in spontaneous colitis and early onset of invasive colorectal cancer . We show that hnRNPI is required for the neonatal immune suppression through decreasing the protein level of IRAK1 , an essential component of toll-like receptor-mediated NF-κB signaling . Our studies demonstrate a critical role of hnRNPI in establishing neonatal immune adaptation and preventing colitis and colorectal cancer .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"immunofluorescence",
"staining",
"immunology",
"epithelial",
"cells",
"animal",
"models",
"developmental",
"biology",
"model",
"organisms",
"signs",
"and",
"symptoms",
"experimental",
"organism",
"systems",
"digestive",
"system",
"research",
"and",
"analysis",
"methods",
"specimen",
"preparation",
"and",
"treatment",
"staining",
"inflammation",
"animal",
"cells",
"biological",
"tissue",
"mouse",
"models",
"immune",
"response",
"gastrointestinal",
"tract",
"diagnostic",
"medicine",
"anatomy",
"cell",
"biology",
"epithelium",
"biology",
"and",
"life",
"sciences",
"neonates",
"cellular",
"types",
"colon"
] |
2017
|
hnRNP I regulates neonatal immune adaptation and prevents colitis and colorectal cancer
|
Many biological and clinical outcomes are based not on single proteins , but on modules of proteins embedded in protein networks . A fundamental question is how the proteins within each module contribute to the overall module activity . Here , we study the modules underlying three representative biological programs related to tissue development , breast cancer metastasis , or progression of brain cancer , respectively . For each case we apply a new method , called Network-Guided Forests , to identify predictive modules together with logic functions which tie the activity of each module to the activity of its component genes . The resulting modules implement a diverse repertoire of decision logic which cannot be captured using the simple approximations suggested in previous work such as gene summation or subtraction . We show that in cancer , certain combinations of oncogenes and tumor suppressors exert competing forces on the system , suggesting that medical genetics should move beyond cataloguing individual cancer genes to cataloguing their combinatorial logic .
Biological complexity , it is thought , is not a simple function of the number of genes in a genome . It likely stems from a variety of factors , including the number of protein states and , as importantly , the number of combinations in which proteins assemble into functional modules [1] , [2] . In development , it is largely combinatorial modules of transcription factors that give rise to the diversity of tissues [3] . Protein combinations are equally instrumental in the pathogenesis of human disease , for instance the inappropriate fusion of Bcr and Abl that leads to chronic myelogenous leukemia [4] or the abnormal interactions acquired by the huntington protein in Huntington's Disease [5] . An intriguing question is how the states of single proteins jointly determine the higher level states of protein modules . In classic biological studies , protein modules have been shown to encode basic logic functions such as AND , OR and NOT which are further combined within larger modules to code for complex programs [6] . A canonical example is the pigment cell module in sea urchin embryos [7] . There , the SuH/Groucho repressor complex forms in the absence of Nic which , in turn , is determined by the lack of Delta signaling . Once Delta signaling is received , the SuH/Groucho repressor complex is displaced by the SuH/Nic activator complex , which activates the GCM gene to induce pigment cell specification . In this case , the module activity can be summarized using basic AND and NOT functions: Another example of network-encoded logic is the BAF chromatin remodeling complex [8] . The stem-cell specific version of the complex ( esBAF ) is characterized by presence of BRG1 but not BRM , and BAF155 but not BAF170 [9] . The neuron-progenitor version ( npBAF ) contains both BAF155 and BAF170 and also incorporates BRM and BAF60C while excluding BAF60B [10] . Pathological forms of BAF have also been characterized . For example the core subunit of the complex , SNF5 , is inactive in malignant rhabdoid tumors , a highly aggressive cancer of early childhood [11] . Given the importance of protein modules and their outputs , a major activity within the field of Systems Biology has been to identify such modules systematically through analysis of global data sets [12]–[16] . Many computational methods have been developed to integrate a panel of gene expression profiles with protein-protein interaction maps or pathway databases , with the goal of associating modules with a biological or clinical outcome [17]–[30] . Among these , several approaches have investigated how protein modules can be used to classify samples . In these methods , each module defines a set of interacting proteins whose expression levels are combined to determine the module activity , which in turn is used to predict the phenotypic class of the sample . However , with one recent exception [28] these methods have assumed that the activity of every module of interest is homogenous and follows a single general function , such as the sum of gene expression levels in a module [20] , [25] or the difference in expression levels across interacting genes in a module [15] , [27] ( Figure 1A ) . While these simple functions ( as well as more advanced frameworks [22] , [24] , [29] ) can identify coherently expressed or perturbed modules , they do not provide the rich logical framework known to occur in biological systems . Here , we develop a novel method called Network-Guided Forests ( NGF ) to learn the network modules whose logic specifies key biological and clinical outcomes . NGF integrates key ideas from Random Forests ( RF ) [31] with biological constraints induced by a protein-protein interaction network— the first use of protein networks in ensemble learning [32] . Rather than relying on a general measure of module activity , NGF fits specific logic functions to each module directly from data . In contrast to Chowdhury et al . [28] who learned network state functions to select informative gene sets that were further used to train a neural network model , the functions identified here are used directly in the classification process . NGF can also readily be applied to continuous gene expression measurements and problems with more than two classes . Using NGF , we explore the functions used in diverse biological programs related to tissue differentiation , breast cancer metastasis , or mesenchymal transformation of brain tumors . For each case a set of network modules is identified which captures known causal mechanisms of development or disease and – in contrast to classical Random Forests – provides robust biomarkers across different sample cohorts . The modules implement diverse logic functions using both coherent and opposing gene activities , in which the module output depends on expression increases for some genes and concomitant decreases for others . Notably in cancer progression , the most predictive decision functions can often be linked to interactions between known oncogenes and tumor suppressors , such that the combined activity of both types of genes determines the disease outcome .
The NGF framework learns a set of decision trees ( the “forest” ) in which each tree maps to a connected component of the protein-protein interaction network ( Figure 1B ) . The decision tree specifies a function that determines the output of the network component based on the activity of its genes . In turn , the collection of all tree outputs is used to predict the cell type or disease state of the biological sample ( the “class” ) . When binary gene activities and two-class decision problems are considered , decision trees map directly to Boolean logic functions [33] ( Figures 1C , S1 ) . In general , however , decision trees can be readily applied to continuous gene activity values and multi-class scenarios [34] . To build a decision tree , NGF selects an initial gene to partition the samples by high versus low gene expression and it scores how well this partition separates the classes . Samples for which the expression of the selected gene is high are placed in the right subtree while those for which the expression is low are placed in the left subtree . NGF then conducts a network-guided search which progressively adds new genes to the tree to improve its discrimination between classes , with new genes chosen from the network neighborhood of genes already in the tree ( Figure 1B; Materials and Methods ) . Many trees are built , starting from many different initial genes , to define the forest . By construction , decision trees include genes that influence a phenotypic outcome both individually and through multi-way interactions with other genes [35] . As in the standard Random Forests algorithm , NGF uses a permutation-based procedure to assess the importance of each gene on the classification accuracy of the forest ( Materials and Methods ) . Motivated by [36] , we also assess the importance of pairs of genes in a tree — in our study these pairs are constrained by the network neighborhood . Genes and gene pairs with significantly high importance scores are placed into clusters that capture similar patterns of presence/absence across the forest of decision trees . Each cluster aggregates genes that fall into the same network region and , in combination , have predictive power over the sample class . Hence these clusters are termed “consensus decision modules” . To apply this framework to study the logic of biological decisions , we obtained mRNA expression data from three diverse studies related to ( 1 ) Development of germ layers , ( 2 ) Breast cancer metastasis , or ( 3 ) Progression of glioma , respectively ( Materials and Methods ) . While these studies collectively span a wide range of human biology , each makes use of mRNA expression profiles to discriminate between classes of development ( study 1 ) or disease ( studies 2 and 3 ) . To provide a complementary protein network , we downloaded a set of 5227 physical interactions measured among pairs of human transcription factors , many of which have been recently reported using the mammalian two hybrid system [15] . NGF was used to combine this protein network with each expression data set to derive a forest of decision trees and corresponding network decision modules for each study ( Figure 2 ) . To allow comparison to other module-finding approaches , we also obtained a network of 57 , 228 human protein-protein interactions as used previously in [20] , [24] . Further information about each expression and network data set is provided below and in Table S1 . Tissue differentiation is largely governed by combinatorial interactions among transcription factors [1] . To identify protein modules involved in tissue development , we applied NGF to qRT-PCR expression profiles collected for 34 human tissues ( Ravasi et al . dataset [15] ) classified according to their embryonic origin: endoderm , mesoderm , non-neural ectoderm , central nervous system ( CNS ) or cell lines ( Figure 2 ) . NGF integrated these data with the transcription factor protein interaction network ( Table S1 ) to reveal a set of 16 consensus decision modules , each containing genes frequently used in combination to predict tissue origin ( Figures 2 , 3A ) . Among these modules , we recognized a number of well-established regulatory complexes with known decisive roles in development ( Table 1 ) . For instance , the single most predictive interaction identified was between HOXC8 and SMAD1 , a transcriptional heterodimer that is known to induce osteoblast differentiation [37] . Also consistent with the logic identified by NGF ( Figure 2 ) , HOXC8 is highly expressed in ectoderm and mesoderm during mouse early embryogenesis [38] . A systematic functional analysis of the modules ( Materials and Methods ) indicated that they were highly enriched for genes whose perturbation is linked to prenatal lethality or improper organ development in mammals ( Figure 3B ) , as reported in the Mouse Genome Informatics ( MGI ) database [39] — an established source of functional associations for both mouse genes and their human orthologs [13] . Gene Ontology analysis [40] indicated that the network was significantly enriched for pattern-specification homeobox genes ( 19/48 genes ) and other developmentally important gene categories , for example embryonic morphogenesis and skeletal system development ( Figure S2 ) . Furthermore , we found that the genes used by NGF to identify a particular tissue origin ( endoderm , mesoderm , ectoderm ) were generally implicated in developmental processes specific for that type of tissue ( Figure 3C and Text S1 ) . To examine the robustness of these decision modules , we investigated whether they could be reproduced from random subsets of the input gene expression profiles , as well as from an independent set of profiles . We found that the protein combinations co-occurring within the same module were highly reproducible across subsets of expression profiles , much more so than the protein combinations identified by the standard Random Forest algorithm ( Figure S3 ) . Further , NGF was used to analyze a large expression profiling study by Muller et al . [13] consisting of 153 types of multipotent stem cells , where each cell type is attributed to the mesoderm , endoderm or ectoderm . We analyzed the single proteins and protein pairs identified as being significantly predictive in the previous dataset ( Ravasi et al . ; Figure 3A ) and compared them to the same number of top scoring proteins and protein pairs identified in the dataset from Muller et al . While only two of ten significant proteins ( 20% ) were identified in common based on single feature analysis , we found that 14 of 38 proteins ( 37% ) were reproduced based on importance scores for pairs of genes ( Figures 3D , S4 ) . Among non-trivial decision modules ( i . e . , those with three or more proteins ) , five out of six ( 83% ) were recovered in both studies ( Figures 3D , S4 ) . In comparison , the standard Random Forest algorithm , which did not use the network , was not able to identify any reproducible gene combinations ( Figure 3E; Text S1 ) . Moreover , randomized runs of NGF ( in which the assignment of expression profiles to network nodes was permuted ) identified only 8% of the same genes and 3% of the same gene-gene combinations ( Figure 3E ) . Taken together , these results indicate that the tissue-specific network expression pattern identified by NGF is both biologically relevant and robust across sample cohorts . While normal developmental programs are tightly regulated , pathological states including cancer can reflect regulatory programs gone awry . To investigate how well NGF can predict cancer progression and identify robust biomarkers , we selected a cohort of 295 nonfamilial breast cancer patients ( van de Vijver dataset [41] ) , for 78 of whom metastasis has been detected during a follow-up visit within five years after surgery . The accuracy of NGF and other algorithms in classifying metastatic vs . non-metastatic samples was assessed using a five-fold cross validation scheme repeated 100 times . The average area under the ROC curve ( AUC ) for Network-Guided Forests was 0 . 74 ( Figures 4A , S5A ) , which was better by 3–6% than previously reported results for a variety of standard and network/pathway-based classification methods [24] , [25] , [27] . Interestingly , the performance of NGF was on par with regular Random Forests ( non-network-based ) , as well as with NGF applied to randomized networks in which the edges were permuted while maintaining the original degree distribution ( NGF**; Figures 4A , S5A ) . Thus , it appears that the decision tree framework used by all three methods is able to find predictive feature sets regardless of the restriction imposed by the protein-protein interaction network . However , in contrast to Random Forests we found that NGF identified many more genes with known roles in breast cancer or cancer in general ( Figure 4B ) . Closer inspection showed that known cancer genes are often not among the most differentially expressed , but are predictive in combination with their network neighbors so that they appear among the most abundant genes in the forest ( Figure 4B ) . In contrast , permuted networks identified far fewer cancer genes among the most abundant features , indicating that the network neighborhood provides crucial information which guides NGF to the biology of disease . To study the robustness of markers identified by NGF , we compared the most abundant features from the van de Vijver dataset to those found in an independent study of 106 metastatic and 180 non-metastatic breast cancer samples described by Wang et al . [42] . The correlation of the resulting gene rankings based on their occurrences in the forest was 0 . 73 for NGF versus 0 . 01 for the regular Random Forest algorithm . Altogether , 31 genes were shared among the 100 most abundant genes from the two datasets , compared to 2 common genes identified by Random Forests ( Figure 4C ) . Thus , the regularization imposed by the network serves to focus the training process on true cancer susceptibility genes , which are observed reproducibly across data sets . These general findings were also observed in a different process related to cancer progression: mesenchymal transformation of brain tissue . Mesenchymal transformation has been associated with exceedingly aggressive forms of high-grade gliomas ( HGGs ) – the most common type of brain tumor in humans . To study network activity patterns leading to the mesenchymal phenotype , we trained the NGF framework on expression profiles of 76 HGG samples previously assigned to one of three groups: proneural , proliferative or mesenchymal [43] . Proneural and proliferative samples were grouped together as “non-mesenchymal” and treated as a control group for detecting the mesenchymal network signature . As with breast cancer , we found that NGF outperformed the benchmark classifier Naïve Bayes in terms of classification accuracy and performed as well as the standard Random Forest algorithm ( Figures S5B , S6A ) . Furthermore , NGF identified more cancer susceptibility genes among the top ranked features ( Figures S6B ) . We next wished to determine whether there were particular network decision functions that were common across biological data sets or , alternatively , which functions were distinct . For this purpose , protein interactions in the decision trees were functionally categorized according to the sign of their proteins in classifying a given phenotype ( Figure 5A; Text S1 ) . The three functional combinations were: “A AND B” , “NOT A AND NOT B” and “A AND NOT B” . We asked which of these functions can best separate the samples into class-homogeneous groups and which types of functions are preferred . Indeed , we found that particular functions were overrepresented among the most predictive gene combinations and that these functions differed across the different biological processes investigated ( Figure S7 ) . Interestingly , across all cancer datasets , decision functions used to predict the more aggressive phenotype were more likely to be associated with “A AND NOT B” logic than other functions ( Figures 5A , S7 ) . Such opposing gene combinations were instrumental in many decision modules identified by NGF . For instance , in breast cancer a highly predictive consensus decision module was identified among C/EBPβ , STAT5A , and HSF1 ( Figure 2 ) – three genes whose activity has been shown to directly influence cancer progression [44]–[46] . The unfavorable metastatic phenotype is associated with high levels of C/EBPβ and HSF1 and low levels of STAT5A ( Figures 2 , S1A ) . Consistent with this prediction , upregulation C/EBPβ can induce acquisition of an invasive phenotype [44] , and expression of HSF1 is required for cellular transformation and tumorigenesis in HER2-positive breast tumors [46] . STAT5 , on the other hand , has been shown to inhibit invasive characteristics of human breast cancer cells and is often lost during metastatic progression [45] . Similarly , for the brain tumor case study , NGF identified a key logic function which associates the mesenchymal phenotype with the upregulation of STAT3 and downregulation of SS18L1 ( Figures 2 , S1B ) . STAT3 is a known oncogene recently identified as a driver of mesenchymal transformation in brain tumors [14] , while SS18L1 is a protein normally required for calcium-dependent dendritic growth and branching in cortical neurons [47] . Across all functional categories , we found that the top scoring decision functions identified in cancer were enriched for interactions between known cancer-related genes ( P = 4 . 92×10−4 and P = 1 . 94×10−3 for the mesenchymal transformation of brain tumors [43] and breast cancer metastasis [42] , respectively ) . Moreover , opposing functional combinations ( “A AND NOT B” ) predictive of the mesenchymal transformation were significantly enriched for interactions between products of oncogenes and tumor suppressors ( Figure 5B ) . In turn , the coherent combinations “A AND B” or “NOT A AND NOT B” were enriched for known interactions between oncogenes or between tumor suppressor genes , respectively ( Figure 5B; Table S2 ) . These results support a model in which the aberrant cancer-related activity is caused by combinations of oncogenes and tumor suppressors co-occurring in the same pathways [48]-[50] and suggest that decision modules reported by NGF may be an excellent means to identify such combinations for further study ( Table S2 ) .
Previous efforts to mine networks for differentially-expressed modules have assumed that module activity can be represented with a single functional form . This hypothesis is expressed in the scoring function that is applied to each module to assess its differential activity . However , our analysis of a representative sample of diseases and developmental programs indicates that the most effective decision functions are in fact not homogeneous , but involve a combination of coherent and opposing gene-gene interactions . While the biological programs covered in this paper are certainly not a comprehensive survey of molecular decision-making , it is significant that both the developmental and cancer modules lead to similar conclusions . First , the network signatures identified by NGF are robust as evidenced by their support from multiple independent datasets . Of the development modules reported by NGF , 83% are reproduced across developmental datasets , in contrast to 0% reproduced by a network-free approach . In breast cancer , we observed a 73% correlation between the features selected for breast cancer , in contrast to 1% for a network-free approach . Second , while the overall classification performance of NGF does not differ from regular Random Forests , network information does achieve sharp focus on genes and gene combinations that are close to the causes of development or disease . A known difficulty with classification using molecular profiles is that it is possible to construct many alternative classifiers all of which have equivalent performance but are based on very different sets of genes [51] , [52] . This is due to the relatively small number of samples as well as the large number of genes that are correlated with outcome . Among the many alternative classifiers , some rely on genes that are close to the true disease mechanisms , while most rely on distantly associated genes . NGF constrains the selected gene features to fall into contiguous protein interaction subnetworks . These network-derived features are more reproducible and strongly enriched in the expected gene functions: Developmental modules are highly enriched for development , and cancer modules are highly enriched for known cancer susceptibility factors . Thus the prior knowledge of the protein interactions serves to filter the set of all possible classifiers [53] allowing NGF to identify those that are based on biologically relevant markers . Finally , network analysis reveals how single factors form predictive combinations . In development , NGF identifies a concise network of HOX genes interacting with developmentally important cofactors , whose tissue-specific roles are just beginning to be illuminated [37] , [54] . In cancer , combinations of interacting oncogenes and tumor suppressors are found such that their combined activity determines disease outcome . Beyond development and cancer , it is likely that for many biological programs , molecular interaction networks will provide a useful framework to guide computational approaches towards biologically-relevant and reproducible genetic logic .
Detailed information on gene expression and protein-protein interaction datasets is provided in Table S1 . Phenotypes associated with genetic perturbations in mouse ( Figure 3B ) were downloaded from the MGI database [55] . Cancer-associated genes ( including breast and brain cancer genes ) were from the Genetic Association Database [56] and were downloaded from DAVID [57] . Lists of tumor suppressors and oncogenes were downloaded from the Cancer Genes database [58] . NGF is a network-based supervised learning algorithm that constructs an ensemble of decision trees which vote to determine the class of a sample . As in the standard Random Forests algorithm , each tree is constructed based on a bootstrap subset of samples drawn with replacement from the original training set . The individual trees are built using the recursive partitioning algorithm CART ( Classification And Regression Trees ) [59] . CART uses a measure of impurity called the Gini index to determine how well a gene and a corresponding expression threshold can differentiate samples with respect to their phenotypic class . The best such gene establishes the first split in the tree . Samples for which the expression value for the selected gene is lower than the threshold are assigned to the left child node in the tree and those with values higher than or equal to the threshold are put in the right child node . This process is iterated for each child node until the improvement in class separation ( as measured by the Gini index ) is lower than ε ( here we use ε = 0 . 02 or ε = 0 . 01 for the global and transcription factor-specific network , respectively ) . In NGF , as in Random Forests , the search process applied by CART is randomized to allow for multiple concurrent trees to be built . First , each tree root is selected as the best gene among a random subset of size √N , where N is the number of all considered genes . Then , at each subsequent node in the tree , the best splitting gene is selected among a random candidate set . NGF selects the candidate set among network neighbors of genes already present in the tree . To promote the identification of dense subnetworks , the roots are required to have at least k network neighbors ( here k = 5 ) and the candidate set of subsequent nodes is expanded iteratively , where each time the probability of selecting a given gene for the candidate set is proportional to the number of interactions it shares with genes already in the tree . NGF also requires that each gene appears at most once on each path from the root to the leaf of the tree . After the trees are constructed , the entire forest is used to determine the class of a new sample . For each tree , the sample is propagated down from the root of the tree and assigned to one of the leaves according to the series of splitting conditions along the path leading from root to leaf . The probability of a given class is determined based on the proportion of training samples that were initially assigned to this leaf . The average probability across all trees is computed and the value of this score is used to determine sample class . Different score thresholds can be used to trade-off specificity and sensitivity . Following [31] , we use samples that were not selected to construct a given tree ( so called “out-of-bag” samples ) to estimate the misclassification error of the tree and determine feature importance . Specifically , we use each tree to classify the corresponding out-of-bag samples and report the percentage of samples misclassified . Next , for each gene in the tree , we measure the increase in the misclassification error resulting from permuting the expression measurements for this gene in the out-of-bag samples . The mean increase of this error over all trees determines the importance score of each gene ( trees in which a gene was not used are counted and contribute 0 to the mean ) . An analogous approach is used to determine the importance scores for pairs of genes . For this we calculate the mean increase in tree misclassification error caused by permuting expression values of any two genes which are used by a particular tree ( see [35] , [36] , [60] for related techniques applied for standard decision tree ensembles ) . To construct network decision modules , NGF outputs the top scoring genes and gene pairs which have a False Discovery Rate ( FDR ) < 0 . 05 , where the null distribution is estimated by executing NGF 100 times on data with permuted class labels . The stability of this procedure increases with the number of trees in the forest . For datasets used here , we found that the method produces robust results provided that the forest contains > 20 , 000 trees . For gene pairs , we additionally check that the mean increase in the misclassification error for the pair is significantly greater than for any single gene in that pair in trees where both genes are present ( FDR<0 . 05 ) . Genes with significant importance scores either independently or in combination with other genes are clustered based on how often they co-appear in the same decision trees . To this end we apply the affinity propagation algorithm [61] which is implemented as a plugin for Cytoscape [62] , [63] . Gene Ontology enrichment analysis was performed using DAVID [57] . MGI phenotype enrichment and enrichment for cancer genes was calculated using Fisher's Exact Test implemented in R ( http://www . R-project . org ) . All enrichments were calculated with respect to the background of all genes present in the input protein-protein interaction network used in each study .
|
Biological outcomes are often determined by modules of proteins working in combination . In classic biological studies , these modules have been shown to encode a diverse repertoire of logic functions which provide the means to express complex regulatory programs using a limited number of proteins . Here , we integrate gene expression profiles and physical protein interaction maps to provide a systematic and global view of combinatorial network modules underlying representative developmental and cancer programs . We develop a new method that associates decision trees with concise network regions to identify network decision modules predictive of biological or clinical outcome . The resulting network signatures prove robust across different sample cohorts and capture causal mechanisms of development or disease . Furthermore , we find that the most predictive network decision functions rely on both coherent and opposing gene activities . Notably , in cancer progression the predictive gene associations often map to physical interactions between known oncogenes and tumor suppressors , where the combined activity of these genes determines disease outcome .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"cancer",
"detection",
"and",
"diagnosis",
"medicine",
"breast",
"tumors",
"cancers",
"and",
"neoplasms",
"basic",
"cancer",
"research",
"algorithms",
"oncology",
"neurological",
"tumors",
"developmental",
"biology",
"organism",
"development",
"regulatory",
"networks",
"biology",
"microarrays",
"systems",
"biology",
"computer",
"science",
"glioblastoma",
"multiforme",
"computational",
"biology",
"glioma",
"genetics",
"and",
"genomics"
] |
2011
|
Protein Networks as Logic Functions in Development and Cancer
|
Praziquantel ( PZQ ) is the drug compound of choice in the control and treatment of schistosomiasis . PZQ is administered as a racemate , i . e . 1∶1 mixture of enantiomers . The schistosomicidal activity arises from one PZQ-enantiomer , whereas the other enantiomer does not contribute to the activity . The WHO's Special Programme for Research and Training in Tropical Diseases ( TDR ) has assigned the low-cost preparation of pure schistosomicidal ( − ) -PZQ a key priority for future R&D on PZQ , but so far this transition has not happened . PZQ has two major administration drawbacks , the first being the high dose needed , and its well documented bitter and disgusting taste . Attempts of taste-masking by low-cost means have not been successful . We hypothesized that the non-schistosomicidal component in PZQ would be the main contributor to the unpleasant taste of the drug . If the hypothesis was confirmed , the two major administration drawbacks of PZQ , the high dose needed and its bitter taste , could be addressed in one go by removing the component contributing to the bitter taste . PZQ was separated into its schistosomicidal and the non-schistosomicidal component , the absolute stereochemical configuration of ( − ) -PZQ was determined to be ( R ) -PZQ by X-ray crystallography , and the extent of bitterness was determined for regular racemic PZQ and the schistosomicidal component in a taste study in humans . Finding: The schistosomicidal component alone is significantly less bitter than regular , racemic PZQ . Our hypothesis is confirmed . We propose to use only the pure schistosomicidal component of PZQ , offering the advantage of halving the dose and expectedly improving the compliance due to the removal of the bitter taste . Therefore , ( R ) -PZQ should be specifically suitable for the treatment of school-age children against schistosomiasis . With this finding , we would like to offer an additional incentive to the TDR's recommendation to switch to the pure schistosomicidal ( R ) -PZQ .
Praziquantel [1] ( PZQ ) is the drug compound of choice in the control and treatment of schistosomiasis [2] , in fact , it is the only commercially readily available drug . So far , no backup compound for PZQ of comparable efficacy and breadth of application is available . Clinically relevant resistance has not been observed , however differences in responses of PZQ-resistant and -susceptible Schistosoma mansoni to PZQ in vitro have been described [3] . PZQ is included in the WHO Model List of Essential Drugs [4] and is at the core of numerous schistosomiasis control programmes . The WHO's strategy for schistosomiasis control [5] aims at reducing morbidity through treatment with PZQ , with a focus on periodic treatment of school-age children and adults considered to be at risk . School-age children are seen as a high-risk group for schistosome infections because they are more susceptible to infection in cases where their increased nutritional needs are not adequately met , might be compromised by helminth infections in their cognitive development , and are continuously exposed to contaminated soil and water but probably less aware of the need for good personal hygiene [6] . While the safety and efficacy against all schistosoma species are outstanding , PZQ has two major administration drawbacks , the first being the high dose needed , 40 mg PZQ/kg bodyweight: Dosages in children are determined by measurement of children's heights using tablet poles , and range from one to five 600 mg-tablets for one treatment . Especially young children have been reported not to be able to swallow these 600 mg tablets [7] . The second drawback is PZQ's well documented bitter and disgusting taste , which can lead to gagging or vomiting if tablets are chewed contrary to recommendation [8] . In veterinary medicine , the oral delivery of PZQ to taste-sensitive companion animals like cats is known to be a challenge . Traditional methods of taste-masking , like the addition of aromas or sugar , are ineffective for PZQ . The bitterness of PZQ even led to PZQ's use as a bitter model drug compound in the effectiveness testing of sophisticated and expensive taste-masking techniques like micro-encapsulation [9] or drug active coating [10] . Apart from anecdotal evidence [2] , we are not aware of reports of low compliance among children treated within schistosoma programmes due to the bitter taste . However , we have to assume that the unpleasant taste of PZQ does not lead to a treatment situation which school-age children would enjoy . PZQ is administered as a racemate , i . e . 1∶1 , mixture of two compounds of identical constitution but non-superimposable mirror-image configuration , so called enantiomers . The straightforward and low-cost chemical synthesis has to be assumed as the reason for the use of the racemate , although it has been known for years that the schistosomicidal activity mainly relies in one PZQ-enantiomer , designated ( − ) -PZQ ( alternatively termed levo-PZQ , l-PZQ , sometimes L-PZQ ) , whereas the other enantiomer , designated ( + ) -PZQ ( alternatively termed dextro-PZQ , d-PZQ ) , does not contribute to the activity [11]–[13] ( Figure 1 ) . From this perspective , only half of the drug compound administered is in fact the drug active , whereas the other half must be considered molecular ballast , which has to be metabolized and excreted while not contributing to the schistosomicidal activity . To the best of our knowledge , no clinical studies in humans exist if and how non-schistosomicidal ( + ) -PZQ alone contributes to the side effects known of racemic PZQ , but this may be assumed: Upon incubation of PZQ and both enantiomers with isolated rat hepatocytes , additional metabolites were detected resulting from the non-contributing ( + ) -PZQ [14] . Various methods of producing the pure schistosomicidal component ( − ) -PZQ exist , which are considerably more expensive than the production of racemic PZQ itself . So far , the potential alone to administer half the current dose by replacing racemic PZQ by ( − ) -PZQ did not lead to a production process for ( − ) -PZQ comparable in costs for racemic PZQ . In the context of the WHO's Global Plan to combat NTDs [15] , the Special Programme for Research and Training in Tropical Diseases ( TDR ) set up an incentive for further R&D work by emphasizing the low-cost preparation of pure schistosomicidal ( − ) -PZQ ( see also the schistosomiasis research collaborative community within The Synaptic Leap [16] ) as a key priority for future R&D on PZQ [17] . Three pharmacological goals for the development were stated: ( 1 ) same dose of ( − ) -PZQ as currently in regular , racemic PZQ , with smaller tablet size and less frequent/severe adverse events , ( 2 ) higher dose of ( − ) -PZQ with similar tablet size and possibly similar adverse event profile as current treatment which could reduce the probability of or delay the development of resistance , or ( 3 ) a combination of these two objectives . As we already mentioned , a smaller tablet size would be more suitable for the treatment of children . Taking into account that the WHO's strategy specifically aims at school-age children , we were intrigued by the question whether the taste disadvantage of PZQ could be turned into an additional incentive to introduce ( − ) -PZQ against schistosomiasis as the drug active of choice . Background to our consideration was the well-documented fact that in most cases taste experiences depend on the stereochemical configuration of the agent [18] , i . e . the taste buds react enantioselectively–like all natural receptors which are composed of chiral constituents like L-amino acids . We hypothesized that ( − ) -PZQ and ( + ) -PZQ would contribute to the bitter taste to a different extent , and that the non-schistosomicidal ( + ) -PZQ would be the main or sole contributor to the disgusting taste . Surprisingly , no public knowledge exists on the tastes of the two enantiomers . We prepared schistosomicidal ( − ) -PZQ , assigned the stereochemical configuration by X-ray crystallography , and determined the extent of bitterness for regular racemic PZQ versus the schistosomicidal component ( − ) -PZQ in a taste study in humans . We chose this comparison over the comparison of ( − ) -PZQ to non-schistosomicidal ( + ) -PZQ because the latter alone does not have any role in a treatment situation . Also the pharmacological studies by others had compared racemic PZQ to ( − ) -PZQ , and not ( + ) -PZQ to ( − ) -PZQ [19] .
Although effective synthetic methods for the enantioselective preparation of PZQ have been reported [20] , we opted for the direct enantioseparation of the racemate yielding gram quantities of both optical forms . The preparative scale chromatography was performed on microcrystalline cellulose triacetate using methanol as the mobile phase , conditions under which the enantiomer having the negative optical rotation emerged first from the column [21] . After crystallisation from methanol/water , ( − ) -PZQ was obtained in enantiomeric excess >99% , as determined by HPLC ( column used Chiralcel OD-H ) . No residual other enantiomer ( + ) -PZQ was detected in this sample . X-ray structural analysis , using Cu-Kα radiation , of a monoclinic crystal in hemi-hydrate form obtained from said fraction by crystallization from methanol/water unequivocally proved the R-configuration of the molecule by measuring Friedel pairs and the Flack parameter ( x = −0 . 1 ( 3 ) ) ( Figure 2 ) . Further details of the crystal structure analysis are available on request from the CCDC ( www . ccdc . cam . ac . uk ) quoting the names of the authors and journal citation . The bitterness values of racemic PZQ and its schistosomicidal component ( R ) -PZQ were determined according to the European Pharmacopoeia [22] by comparison with quinine hydrochloride , the bitterness value of which is set at 2×105 . The bitterness value is defined by the European Pharmacopoeia as the reciprocal of the concentration of a solution in a dilution series of a compound , a liquid or an extract that still has a bitter taste . Concentrations of solutions used in the tests ranged from 1 . 69×10−8 to 1 . 0×10−4 g/mL . A test panel consisting of sixteen members was assembled . Although children comprise the treatment target group no children were included in the test panel . All panel members were adults completely untrained in performing sensory tests . To correct for individual differences in tasting bitterness amongst the panel members a correction factor was determined for each panel member by preparing dilutions of quinine hydrochloride . The mouth was rinsed with water before tasting . The dilution with the lowest concentration having a bitter taste was determined by taking 10 mL of the weakest solution into the mouth and passing it from side to side over the back of the tongue for 30 seconds . If the solution was not found to be bitter , the panellist had to spit out and wait for one minute before the mouth was rinsed again with water . After 10 minutes , the next dilution in order of increasing concentration was tasted . The correction factor k for each panel member was calculated according to the European Pharmacopoeia by k = n/5 , where n is the number of millilitres of the stock solution in the dilution of the lowest concentration that is judged to be bitter . One panel member detected bitterness already in pure water , and was therefore excluded from the test panel . Dilutions of the test compounds racemic PZQ and ( R ) -PZQ were prepared and tasted by the remaining fifteen members of the test panel in the same manner as described for quinine hydrochloride . The bitterness value as experienced by each member was calculated according to the European Pharmacopoeia taking the individual-related correction factor into account by Y×k/X×0 . 1 , where Y is the dilution factor of the dilution , and X is the number of millilitres of the respective dilution which , when diluted to 10 mL with water , still has a bitter taste . The bitterness value of the test compounds resulted from calculating the average of the individual values . Requested statement: Informed written consent was obtained from all panelists to participate in this taste study . As a taste study , and not a medical study in the sense of the WMA Declaration of Helsinki , it did not require approval of an independent review board ( highly diluted preparations were tasted and spat out–they were not ingested ) . Nevertheless , it was conducted according to the principles of the WMA Declaration of Helsinki where applicable .
The results of the determination of bitterness values are shown in Table 1 . Remarkable is the variation of the individuals' results as indicated by the relative standard deviation and the dispersion of the results in the box-and-whisker diagram ( Figure 3 ) . In contrast to the average , the medians of the results , as depicted in the box-and-whisker diagram , are different from each other . The observed variation was probably provoked by the test panel consisting of untrained members only [23] . Thirteen out of fifteen panel members found ( R ) -PZQ to taste less bitter than racemic PZQ . Although no statistical test is required or proposed by the European Pharmacopoeia , a statistical test ( using SAS software , release 9 . 1 . 3 , SAS Institute Inc . , Cary , NC , USA ) was conducted to investigate the observed difference between the compounds . Considering the small sample size and the nature of the data which does not justify the assumption of a normal distribution , a nonparametric , distribution-free method was chosen . On the 5% level of significance , Wilcoxon's Signed Rank Test ( two-sided ) resulted in a significant difference between the taste of racemic PZQ and ( R ) -PZQ ( p = 0 . 0107 ) . This result was confirmed by the Sign Test ( two-sided , p = 0 . 0018 ) . In addition to the quantitative determination of the bitterness values , qualitative taste sensations were noted by the members of the test panel for each compound . For racemic PZQ , all panel members commonly observed the sensation of an unpleasant chemical or metallic taste or a taste circumscribed best by old rubber . On the other hand , for ( R ) -PZQ the panellists commonly described the sensation of a moderate chemical taste , comparable to that of a polyethylene or a rubber pipe . Although the tastes were not recognized alike across the test panel , for the majority of the test panel we can state that ( R ) -PZQ had a less unpleasant taste compared to racemic PZQ .
The schistosomicidal component of regular PZQ , ( R ) -PZQ has a less unpleasant taste compared to racemic PZQ , which was found to be comparably bitter or unpleasant . It can be assumed that the disgusting taste of racemic PZQ stems from the non-schistosomicidal component , ( S ) -PZQ . Removing the latter from currently used racemic PZQ therefore not only offers the chance to halve the dose , with the potential to decrease the number or size of the tablets , but also addresses the second disadvantage of regular , racemic PZQ-its unpleasant taste . With this finding and its publication we would like to offer an additional incentive to focus work of the PZQ R&D community on further decreasing the cost of production of ( R ) -PZQ with the goal to switch to pure ( R ) -PZQ as a replacement for racemic PZQ for the treatment of school-age children against schistosomiasis .
|
Schistosomiasis , or Bilharzia , is a parasitic disease caused by flatworms , which affects about 200 million people worldwide . Praziquantel ( PZQ ) is the drug compound of choice in the control and treatment of this disease . Only half of the drug dose currently administered actually has activity against schistosomiasis , whereas the other half has no activity . Therefore , the WHO has assigned the low-cost preparation of the pure active component a key priority for future PZQ research and development . PZQ has two major administration drawbacks , the first being the high dose needed , the second its well documented bitter taste . Attempts of masking the unpleasant taste have not been successful . We hypothesized that the non-active component in PZQ would be the main contributor to the unpleasant taste of the drug . We determined the extent of bitterness for regular PZQ compared to the pure active component in a taste study in humans . We found that the pure active component alone is significantly less bitter than regular PZQ . This new finding should serve as an additional incentive for the PZQ research and development community to provide a low-cost , large-scale preparation route to the pure active component of PZQ .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"infectious",
"diseases/helminth",
"infections",
"pediatrics",
"and",
"child",
"health",
"chemistry/organic",
"chemistry",
"infectious",
"diseases/neglected",
"tropical",
"diseases"
] |
2009
|
Taste, A New Incentive to Switch to (R)-Praziquantel in Schistosomiasis Treatment
|
Recent advances in experimental techniques have allowed the simultaneous recordings of populations of hundreds of neurons , fostering a debate about the nature of the collective structure of population neural activity . Much of this debate has focused on the empirical findings of a phase transition in the parameter space of maximum entropy models describing the measured neural probability distributions , interpreting this phase transition to indicate a critical tuning of the neural code . Here , we instead focus on the possibility that this is a first-order phase transition which provides evidence that the real neural population is in a ‘structured’ , collective state . We show that this collective state is robust to changes in stimulus ensemble and adaptive state . We find that the pattern of pairwise correlations between neurons has a strength that is well within the strongly correlated regime and does not require fine tuning , suggesting that this state is generic for populations of 100+ neurons . We find a clear correspondence between the emergence of a phase transition , and the emergence of attractor-like structure in the inferred energy landscape . A collective state in the neural population , in which neural activity patterns naturally form clusters , provides a consistent interpretation for our results .
The past decade has witnessed a rapid development of new techniques for recording simultaneously from large populations of neurons [1–4] . As the experimentally accessible populations increase in size , a natural question arises: how can we model and understand the activity of large populations of neurons ? In statistical physics , the interactions of large numbers of particles create new , statistical , laws that are qualitatively different from the original mechanical laws describing individual particle interactions . Studies of these statistical laws have allowed the prediction of macroscopic properties of a physical system from knowledge of the microscopic properties of its individual particles [5] . By exploiting analogies to statistical physics , one might hope to arrive at new insights about the collective properties of neural populations that are also qualitatively different from our understanding of single neurons . The correlated nature of retinal ganglion cell spike trains can profoundly influence the structure of the neural code . Information can be either reduced or enhanced by correlations depending on the nature of the distribution of firing rates [6] , the tuning properties of individual neurons [7] , stimulus correlations and neuronal reliability [8 , 9] , the patterns of correlations [10] , and interaction among all these factors [11] . In addition , the structure of the decoding rule needed to read out the information represented by a neural population can be strongly influenced by the pattern of correlation regardless of whether it reduces or enhances the total information [12 , 13] . One approach to understanding the properties of measured neural activity is to study the nature of minimally structured ( ‘maximum entropy’ ) models of the probability distribution that reproduce the measured correlational structure [14–16] . These models have been shown to be highly accurate in reproducing the full statistics of the activity patterns of small numbers of neurons [15 , 17 , 18] . The hope is that even if these models underestimate the real structure of larger neural populations , the properties of the distribution which arise in these simplified models are general and of consequence to the true distribution . Maximum entropy models that constrain only the pairwise correlations between neurons are generalized versions of the Ising model , one of the simplest models in statistical physics where collective effects can become significant . The macroscopic behavior of these models varies substantially depending on the parameter regime . By fitting these models to measured neural acitivity , we can begin to explore ( by analogy ) the ‘macroscopic’ properties of the retinal population code . In particular , we can gain insight into these macroscopic properties by introducing a fictitious temperature variable into the maximum entropy model . By changing this temperature , we can continuously tune the population from a ‘high temperature’ regime where correlation has a minimal effect on the probability distribution to a ‘low temperature’ regime where correlation dominates . Furthermore , we expect to see signatures of a phase transition at the boundary between these regimes . Thus , by observing a phase transition at a particular value of the temperature variable , we can determine if the real state of the neural population more closely resembles the high or low temperature state . One macroscopic property of interest is the specific heat . Discontinuity or divergence of this property is an indicator of a second-order phase transition , which implies a qualitative change in the properties of the system . Previous studies have shown that the specific heat has a peak that grows and sharpens as more neurons are simultaneously analyzed [19–22] . Most of the literature on this topic can be divided into two camps: the ‘proponents’ who argue that this is a signature of criticality , i . e . the system is poised in between high and low temperature phases , in a manner that might optimize the capacity of the neural code [19 , 21 , 23] , and the ‘sceptics’ who argue that this is merely a consequence of ignored latent variables [24 , 25] , ignored higher order correlation structure in the data [26] , or even the presence of any correlation at all [27] . An alternative interpretation is that system is in a ‘low temperature’ state , and that the peak in the heat capacity is a signature of a first-order phase transition . The difference between the two types of transitions rests on which macroscopic properties of the system are discontinuous at the transition: first-order phase transitions have discontinuities in the entropy ( hence having an infinite heat capacity ) , while second-order phase transitions will have discontinuities , or integrable divergences , in the specific heat . The observed sharpening of the specific heat is influenced by finite-size effects which could be consistent with either a delta function ( first-order ) or a divergence ( second-order ) in the specific heat . In principle , one can use finite-size scaling arguments to argue that the sharpening is more or less consistent with one of the two possibilities . In practice however we do not think that we can convincingly distinguish between these two possibilities with our analysis of the specific heat , and so we cite other forms of evidence in favor of our interpretation . Empirically , the peak of the specific heat is found at a higher temperature than the operating point of the real system ( T = 1 ) , suggesting that the system is on the low temperature side of the phase transition . Low temperature phases in statistical physics are usually associated with structure in the distribution of states , in which the system can ‘freeze’ into ordered states . High temperature phases , in contrast , are associated with weakly correlated , nearly independent structure in the population of neurons . From this perspective , the phase transition that many studies have observed as a function of an imposed temperature variable , T , serves as an indicator of structure in the probability landscape of the neural population at the real operating point ( T = 1 ) . Maximum entropy models fit particular statistics of the distributions of experimentally measured neural neural activity [16] . Because retinal responses are specific to both the adaptational state of the retina and the ensemble of stimuli chosen to probe them , the measured pattern of neural correlation—and hence the detailed properties of the maximum entropy model—will also vary . Therefore , it is yet unclear how robust is the presence of a low temperature state to different experimental conditions . The phase transition itself arises as a consequence of the correlations between neurons . This pattern of correlation in turn has contributions from correlations in the stimulus and from retinal processing . The distribution of correlations also has a particular shape , with many weak but statistically non-zero terms . It is unclear how these different properties contribute to the nature of the structured collective state of the neural population . Here , we show that while the detailed statistics of the retinal population code differ across experimental conditions , the observed phase transition persists . We find that retinal processing provides substantial contributions to the pattern of correlations among ganglion cells and thus to the specific heat , as do the many weak but statistically non-zero correlations in the neural population . We also find that the spatio-temporal processing of the classical receptive field is not sufficient to understand the collective properties of ganglion cell populations . To address the nature of the collective state of the retinal population code , we explored how a particle representing the state of neural activity moves over the system’s energy landscape under the influence of finite temperature . We find that the energy landscape has regions that “trap” particle motion , in analogy to basins of attraction in a dynamical system . By varying the overall correlation strength , we show that the emergence of this structure is closely connected to the emergence of the measured phase transition . This emergence occurs at surprisingly low overall correlation strength , indicating that the real population is robustly within the structured regime .
A first step towards the study of collective phenomena in neural populations is to understand what is the qualitative nature or ‘phase’ of the neural population . Phases of matter occur everywhere in nature where there is some collective structure in the population . In the theory describing phase transitions in statistical physics , first-order phase transitions can occur when a particular system can decrease its free energy by transitioning to a new phase . While such a transition in our work here will occur in a region of parameter space that is not real—i . e . , is not visited by the retina experimentally—its occurence provides evidence for structure in the real experimentally measured distribution . Thus , an explanation of previously observed phase transitions is that the pattern of correlation among ganglion cells induces a highly structured phase which is qualitatively different from the phase found in the high temperature limit . From this perspective then , we ask whether this phase is robust to different adaptational states , and what are the properties of the retinal population code that give rise to that phase ? To study the emergence of a phase transition with increasing system size , we subsampled groups of N neurons and inferred models for these subsets of the full neural data . For all of these networks , we then introduced a fictitious temperature parameter , T , into the distribution , P ( R ) = ( 1/Z ( T ) ) exp ( −E ( R ) /T ) . This parameter allows us to visit parameter regimes of our model where the qualitative nature of the system changes . If the shape of the specific heat as a function of temperature exhibits a sharp peak , this indicates a phase transition—a macroscopic restructuring of the properties of the system across parameter regimes . Thus , an analysis where we vary the effective temperature allows us to gain insight into the state of the real neural population at T = 1 . As previously described [19] , we found a peak in the specific heat that sharpened and moved closer to T = 1 with increasing system size , N ( Fig 3A–3C ) . The systematic changes that we observe as a function of the system size N ( Fig 3A–3C ) indicate that correlation plays a more dominant role as the population size increases . To further understand the role of correlations , we performed a shuffle test , where we broke correlations of all orders in the data by shifting each cell’s spike train by a random time shift ( including periodic wrap-around ) that was different for each cell . Following this grand shuffle , we repeated the full analysis procedure described above ( fitting a maximum entropy model to the shuffled data and estimating the specific heat ) . We found that the heat capacity had a much lower and broader peak that did not change as a function of N . In addition , this heat capacity curve agreed closely with the analytical form of the specific heat for an independent neural population ( S6 Fig ) . This analysis demonstrates that the sharpening of the specific heat that we observed is a direct consequence of the measured pattern of correlation among neurons [Fig 3E] . The shuffled curves were noticeably different across light adapted conditions ( Fig 3D ) . This is not surprising as the analytical form for the specific heat of a network of independent neurons depends only on the average firing rate of each neuron , and these are substantially different between the two luminance conditions ( Fig 2A ) . However , the heat capacity peaks for both the dark and the light conditions became more similar with increasing N . Clearly some macroscopic properties of the network were conserved across luminance conditions for the real , correlated , data ( Fig 3 ) . The correlation structure of natural movies can in principle trigger a broad set of observed retinal adaptation mechanisms , such as adaptation to spatial contrast [49 , 50] , temporal contrast [51] , and relative motion of objects [52] . To generalize our results to these higher-order adaptive mechanisms , we ran another experiment comparing the distributions of responses of the same retina to two different natural stimuli ensembles , without a neutral density filter ( Experiment #2 , see Methods ) . These two natural movies were of grass stalks swaying in the wind ( M1 , the same movie as in the previous experiment ) , and ripples on the surface of water near a dam ( M2 ) . The first movie ( M1 ) had faster movements , larger contrasts , and fewer periods of inactivity . Likely as a consequence , we found higher firing rates in ganglion cells during M1 ( Fig 4A ) . We found statistically significant differences in the correlation coeffecients , Cij , and P ( K ) across the two stimulus conditions ( Fig 4B and 4C ) . However , the specific heats of the full networks in the two movies sharpened similarly across conditions ( Fig 4D ) , indicating that this macroscopic property of the retinal population code was also robust to different choices of naturalistic stimuli . So far , our results have demonstrated that the peak in the specific heat is due to the pattern of correlation among neurons . However , these correlations have contributions both from retinal processing , such as the high spatial overlap between ganglion cells of different functional type [53 , 54] , and from the correlation structure in the stimulus itself . In order to compare the relative importance of these two different sources of correlation among ganglion cells , we measured neural activity during stimulation with a randomly flickering checkerboard . By construction , our checkerboard stimulus had minimal spatial and temporal correlation: outside of 66 μm squares and 33 ms frames , all light intensities were randomly chosen . Returning to Experiment #1 in the light-adapted condition , we compared the response of the retina to the natural movie and the checkerboard stimuli ( Note that here we are working with the N = 111 ganglion cells that were identifiable across both conditions , a subset of the N = 128 ganglion cells we worked with in Fig 3 ) . The distribution of pairwise correlation coefficients was tighter around zero when the population of ganglion cells was responding to the checkerboard stimulus ( Fig 5A ) . The specific heat in the checkerboard was smeared out relative to the natural movie , but was still very distinct from the independent population ( Fig 5B ) . This suggested to us that most , but not all , of the contributions to the shape of the specific heat were shared across the two stimulation conditions , and therefore arose from retinal processing . A simple and popular view of retinal processing is that each ganglion cell spike train is described by the spatio-temporal processing of the cell’s classical receptive field . In this picture , correlation between ganglion cells arises largely from common input to a given pair of ganglion cells which can be described by the overlap of their receptive fields . To explore the properties of this simple model , we estimated linear-nonlinear ( LN ) models for each of the N = 111 ganglion cells in the checkerboard recording ( Methods ) . We then generated spike trains from these model neurons responding to a new pseudorandom checkerboard sequence , and binarized them into 20ms bins in the same manner as for the measured neural data . As expected , the receptive fields had a large degree of spatial overlap [53 , 55] , which gives rise to significant stimulus-dependent correlations . We found that these networks did not reproduce the distributions of correlations found in the data , instead having lower values of correlation and fewer outliers ( Fig 6A ) . The specific heat of the network of LN neurons was reduced relative to the neural data that the LN models were based upon ( Fig 6B ) . Thus , the peak in the specific heat is enhanced by the nonlinear spatial and temporal computations in the retina that are not captured by models of the classical receptive field . Because the detailed properties of the maximum entropy model depend strictly on correlations measured within the neural population , we wanted to develop a more general understanding of what aspects of the pattern of correlation were essential . To do this , we altered particular properties of the measured matrix of correlations , keeping the firing rates constant . We then inferred the maximum entropy model parameters for these new sets of constraints , and estimated the specific heat . For these manipulations , we worked with the simpler pairwise maximum entropy model . We made this choice for several reasons . First , manipulating only the pairwise correlation matrix made our analysis simpler and more elegant than also having to perturb the distribution of spike counts , P ( K ) . There is a large literature reporting values of pairwise correlation coefficients , helping us to make intuitive choices of how to manipulate the correlation matrix , while very little such literature exists for P ( K ) . Additionally , any perturbation of the correlation matrix consequently changes P ( K ) , so that attempting to change the correlation matrix while keeping P ( K ) fixed is a nontrivial manipulation . Second , in the pairwise model all effects of correlational structure are confined to the interaction matrix . This interaction matrix has been studied extensively in physics [56] , and hence there is some intuition as to how to interpret systematic changes in the parameters . Conversely , we have little intuition currently for the nature of the k-potential . Our final and most important reason was that the qualitative behavior in the heat capacity ( sharpening with system size , convergence across light and dark datasets ) is the same for both pairwise and k-pairwise models across all conditions tested ( S7 Fig ) . The correlation matrix in the retinal population responding to a natural stimulus has many weak but statistically non-zero correlations [54 , 55] , a result also found elsewhere in the brain [57 , 58] . To test their contribution to the specific heat , we kept only the largest L correlations per cell , replacing the other terms in the correlation matrix with estimates from the shuffled ( independent ) covariance matrix . If our results are based on a “small world network” of a few , strong connections [59] , then the specific heat for small values of L should begin to approximate our results for the real data . Clearly ( Fig 7A ) , even keeping the top L = 10N ( out of a total of L = 63 . 5N values ) strongest correlations did not reproduce the observed behavior . Therefore , the full “web” of weak correlations contributed substantially to the shape of the specific heat of the retinal population code . We were next interested in understanding the qualitative nature of how networks transition between the independent and fully correlated regimes . Our approach was to scale all the pairwise correlations down by a constant ( α ) , to subselect groups of neurons as previously in Fig 3 , and to follow the inference procedure described above . Specifically , we formed a new correlation matrix C i j mixed ( α ) = α C i j true + ( 1 - α ) C i j shuff . We found that the specific heat of the neural population exhibited a transition between independent and fully correlated behavior , as the correlation strength , α , ranged from 0 to 1 ( Fig 7B ) . Peaks in the heat capacity similar to the one observed in the full model emerge when α is greater than a critical value α* , which we estimated to be between 0 . 225 to 0 . 25 ( Fig 7B and 7C ) . Substantially similar behavior was observed in the dark condition as well ( Fig 7D ) . These data suggest that the low temperature phase emerges near α* . In fact this behavior constitutes another phase transition which can be observed without the introduction of a fictitious temperature parameter , in the curve of the specific heat of the real system ( T = 1 ) as a function of correlation strength α ( Fig 7C ) . There is a clear emergence of a contribution to the specific heat that depends on the system size , N . This additional contribution gives rise to a discontinuity in either the specific heat or its derivative . Similar behavior is observed in a classic model of spin glasses , the Sherrington Kirkpatrick ( SK ) model [60] , as we describe in the Discussion . Importantly , the critical value of alpha at which we see a transition to structure , α* , was substantially smaller than the measured correlation strength , α = 1 . This indicates that the population of retinal ganglion cells had a overall strength that was “safely” within the strongly correlated regime . Thus , the low temperature state is robust to changes in adaptational state or stimulus statistics that might shift the overall strength of correlations among neurons . Our hypothesis was that the emergence of a phase transition was correlated with the emergence of structure in the energy landscape . Previously , the structure of the energy landscape has been studied with zero temperature Monte Carlo ( MC ) mapping of local minima [16] , where one changes the activity state of single neurons such that the energy of the population activity state always decreases . States from the data were thereby assigned to local minima in the energy landscape , which can be thought of as a method of clustering a set of neural activity patterns into population “codewords” [16] . If each cluster encodes a specific visual stimulus or class of stimuli , then this clustering operation provides a method of correcting for errors introduced by noise in the neural response . There are two reasons why we chose to study the structure of the energy landscape at the operating point of the system ( T = 1 ) . First , when we performed zero temperature descent with our models , our primary finding was that the overwhelming majority of states descended into the silent state ( only 503 out of 1 . 75 ⋅ 105 did not descend into silence on a sample run ) . This indicated that the energy landscape had very few local minima . Thus we needed a different approach to explore the structure of the energy landscape . Second , we were interested in properties of the system ( such as the specific heat ) that were themselves temperature dependent , so it made sense to stick with the real operating point of the neural population ( T = 1 ) . When analyzing sufficiently large neural populations ( typically , “large” means N > 20 cells ) , there are too many states to simply ennumerate them all . As a consequence , the energy landscape was accessed indirectly , through a Markov Chain Monte Carlo ( MC ) sampler [61] , which simulates an exploration of phase space by defining the state-dependent transition probabilities between successive states . Provided that these transition probabilities are properly defined , the distribution of samples drawn should approach the desired ( true ) distribution with sufficient sampling . The set of these transition probabilities across all the neurons defines a ‘direction of motion’ in neural response space . We will study these directions of motion as a way to gain more insight into the properties of the energy landscape of our measured neural populations . Note , however , that MC sampling dynamics is used here as a tool to explore the geometric properties of the probability landscape of the retinal population; these are not claims of how the real dynamics of activity states change in the retina under the influence of the visual stimulus . To study the relationship between the directions of motion given by the MC sampling process and the observed phase transition , we returned to the manipulation with scaled covariances . For a given state R , the MC sampler in each model ( inferred for a particular value of the correlation strength α ) will return a vector of conditional probabilities X ( R , α ) = {xi} , where the conditional probability for each cell i’s activity is given by x i ( R , α ) = exp ( h i ( eff , α ) ( R ) ) / ( exp ( h i ( eff , α ) ( R ) ) + 1 ) ( 3 ) with an effective field , given by h i ( eff , α ) ( R ) = h i ( α ) + 2 ∑ j ≠ i J i j ( α ) r j ( 4 ) We can now ask how the shape of the energy landscape evolved with respect to α . Specifically , we first compared the similarity in direction of these ‘Monte Carlo flow’ vectors with the vectors defined for the fully correlated model ( at α = 1 ) , by calculating the average overlap between states , p , at a correlation strength α with the same states at α = 1 , X ^ ( R p , α ) · X ^ ( R p , α = 1 ) . We found that in the independent limit , α = 0 , the flow vectors pointed in substantially different directions than in the fully correlated population ( Fig 8A ) . This indicates that there is very different structure in the energy landscape in these two limits . Furthermore , as alpha increased from zero , we found a steep increase in the similarity of flow vector directions until a little past our estimate of the critical value , α ≈ 0 . 3 , at which point the slope became more and more shallow . The probability landscape roughly settled into its fully correlated ‘shape’ by α ≈ 0 . 5 , which was comparable with the point by which the specific heat had stabilized near its fully correlated value as well ( Fig 7C ) . This is consistent with a tight connection between the emergence of a phase transition and the development of structure in the probability landscape . We carried out a similar analysis to compare the amplitude of Monte Carlo flow vectors as a function of the correlation strength , α ( Fig 8B ) . At α = 0 , we found that flow vector magnitudes were very different from the fully correlated population for states with high spike count , K . The similarity in amplitude increased gradually up to α* , increased sharply from α* up to α ≈ 0 . 5 , and then changed slowly at higher values of α . Again , these results are consistent with the interpretation that the shape of the energy landscape emerged at a correlation strength near α* and that further increases in α served to ‘deepen’ the existing contours in the energy landscape . While our energy landscape does not have many true local minima , we can gain insight into the nature of the energy landscape induced by correlations by considering how long the system remains in the vicinity of a given state under T = 1 MC dynamics . Since the experimentally measured neural activity is sparse , the directions of motion are heavily biased towards silence . Regardless of initial state , the sampler will eventually revisit the silent state . To demonstrate this , we returned to the data from Experiment #1 ( M1 , light ) , and the corresponding k-pairwise model fits . Due to the addition of the k-potential in this analysis , the effective field was now h i ( eff ) ( R ) = h i + 2 ∑ j ≠ i J i j r j + λ K + 1 - λ K ( 5 ) with K = ∑j≠i rj . The effective field is derived as the difference in energy of the system when cell changes from silent to spiking ( Eq 3 ) , and so it has three contributions: one from the local field , hi , another from the sum of pairwise interactions , and a final term from the change in the k-potential , λK+1 − λK . We worked with all the states observed in the light condition that had K = 12 spiking cells ( total number of states = 3187 ) . We selected a set of initial activity states all having the same value of K , as these analyses depended strongly on K . We wanted a value of K large enough that effective fields were large , and hence collective effects of the population code were significant . At the same time , we needed K to be small enough that we could observe many such states in our sampled experimental data . Balancing these two concerns , we chose K = 12 . In order to characterize the ‘dwell time’ of a particular state , we initiated many MC sampling runs from that given state . On each of these runs , we defined the dwell time as the number of MC samples ( where all N cells were updated ) required to change 9 of the 12 originally spiking cells to silent . For each given initial state , our analysis produced a distribution of dwell times , due to the randomized order of cell choice during sampling , as well as the stochasticity inherent in sampling at T = 1 . We found significant differences in the distributions of dwell times across different initial states ( Fig 9A ) . This demonstrated that Monte Carlo flow was trapped for a longer amount of time in the vicinity of particular states , consistent with subtle attractor-like properties in the geometric structure of the probability landscape . For the same initial states , average dwell times measured on the energy landscape for independent models were almost an order of magnitude shorter , indicating that these effects were due to the measured correlations ( Fig 9B ) . We searched for a measure that could capture the variability in dwell times across initial states in the full model , based on our intuition that a state that started near a local minimum would have a long persistence time before finite temperature MC sampling would move it far away . This led us to define a persistence index ( PI ) that captured the tendency of a state to remain near its starting point under MC sampling dynamics . Specifically , for a given state R , we define P I = X ^ ( R ) · R ^ , namely the cosine of the angle between the initial state and the average next state . If the PI is close to 1 , then the direction that the state R evolves towards under MC sampling is the state itself , and hence the state will remain the same . Over all the initial states studied , we found a large positive correlation between the average dwell times and the persistence indices for the fully correlated population ( correlation coefficient of 0 . 75 , Fig 9C ) . This significant correlation justifies the use of the PI as a simpler proxy for the dwell time . In contrast , the correlation was only 0 . 38 when measured on the energy landscape of the independent model and dwell times were systematically smaller by orders of magnitude ( Fig 9C blue ) . The persistence index also allowed us to characterize the similarity of the neural code for the same natural movie between light and dark adapted conditions . If the structure in the energy landscape changed between the dark and light conditions , then one would expect the dwell times , and hence the persistence indices , to change as well . Instead , we found a strong correspondence across the light and dark experimental conditions , that was absent in the independent model ( Fig 9D , S8 Fig correlation coefficient of 0 . 90 vs 0 . 37 ) . To estimate the variability in this measure , we compared the PI across models inferred for two separate random halves of the light condition ( S8 Fig , correlation coefficient of 0 . 97 ) . Thus , the pattern of correlation measured in the two luminance conditions created similar structure in the system’s energy landscape , even though the detailed statistics of neural activity were quite different . This structure endows the population code with a form of invariance to light level that is not present at the level of individual ganglion cells .
Representing visual information with specific multineuronal activity states is an error-prone process , in the sense that responses of the retina to repeated presentations of identical stimuli evoke a set of activity states with different probabilities . A many-onto-one mapping of individual activity states to cluster identities naturally reduces this variability , thus endowing the population code with a form of error correction . In fact , this appealing and intuitive idea has been recently demonstrated for the retinal population code [62 , 63] . Clustering of activity states manifested itself in the geometric structure of the probability distribution , which we characterized by several analyses , including Monte Carlo sampling dynamics , Monte Carlo flow vectors , and persistence indices . This structure was found to be preserved across variations in ambient luminance , and was robust to minor perturbations of the correlation matrix . This robustness was not obviously evident in the lower order statistics of the distribution ( firing rates , correlations , spike count distribution ) , which were measured directly . For downstream readout mechanisms which access only the incoming retinal population code , such a robustness in the clustered organization constitutes a form of invariance in the retinal representation . The variability in the interaction matrix gives rise to the variability that we observed in measures of persistence across states with the same number of spiking cells K ( see Fig 9 ) . In our picture in which the probability distribution over all neural activity states is organized into a set of clusters , some states are ‘attractor-like’ . These states have a higher density of nearby states , which corresponds to lower energies , and thus traps states in their vicinity under MC sampling dynamics . Other states do not have this property at all , and hence the dwell time around these states under MC sampling dynamics approaches that of a network of independent neurons . This property depends crucially on the detailed structure of the pairwise interactions Jij . If for instance all the interaction matrix terms were set equal to a positive constant as in a ferromagnet , i . e . Jij = J0 , then the effective fields for all cells would have almost the same contribution from interactions , namely a quantity proportional to KJ0 . The only variability in the conditional probabilities , X ( R ) , that cells would experience would be due to the local fields and whether or not the cells were active in the state ( which reduces the effective K by one ) ; this variability would be overwhelmed with increasing K . As a result , the effective field would eventually tend to a constant for all cells at large enough K . However , we observe in our data that this is not true: the persistence index of states with the same K varied substantially ( Fig 9 ) . The effect of variability in the distribution of interaction matrix terms has been studied extensively in spin glass models [56] . In order to convey some of our intuition about the low temperature regime , we will discuss a particular example of a glassy model and its relationship to our work . But keep in mind that while we believe there is a useful analogy between some of the properties of the glassy limit of the Sherrington-Kirkpatrick ( SK ) model [60] and our measured neural distribution , we are not claiming that our distributions are fit by the SK model . The SK model is a model of all-to-all connectivity in a population , similar to the high degree of connectivity we observe in our inferred models of a correlated patch of ganglion cells . For example , cells have an average of 46 non-zero interactions per cell , out of N = 128 possible in the k-pairwise model in the light condition . The SK model itself has two regimes , characterized by the relationship between the variance ( σ J 2 ) and mean ( μJ ) of the interaction parameters , Jij . When the variability is large relative to the mean , the low temperature phase of the SK model is a spin glass phase , where the probability distribution over all activity states is characterized by an abundance of local maxima . The glassy SK model also undergoes a phase transition from a weakly-correlated paramagnetic phase to a structured spin glass phase as temperature is lowered . This transition is an ergodicity-breaking transition that is third-order ( see below ) [56 , 64] . In other words , this transition is characterized by a significant reduction in the number of states available to the system , as the glassy state confines the system to particular valleys in the energy landscape . A note is in order: ergodicity-breaking phase transitions can only be formally defined when there is no possibility that the system can escape a particular valley . This is only true mathematically in the thermodynamic limit ( N → ∞ ) , as a finite-sized system will always have a small but nonzero probability with which it can escape the phase space it is confined to . So we should always keep in mind that as we compare our data to the SK model , we need to consider the finite size limit of the SK model . By comparison , when we scaled up the correlation strength α from values below the critical correlation strength , α* , to values above , we observed a phase transition where the smooth independent distribution wrinkled to form a set of attractor-like states in the energy landscape ( Figs 7 , 8 and 9 ) , perhaps with the geometry of ‘ridges’ [65] . While this wrinkling transition does not strictly break ergodicity , like in the thermodynamic limit of the SK model , it does confine the system near attractor-like states . This confinement may in fact be quite similar to the ergodicity-breaking transition in the SK model . Indeed , our analysis of this transition suggests that it is consistent with a third-order transition , where there is a cusp in the specific heat at the critical temperature ( i . e . a discontinuity in the derivative of the specific heat , not in its value; see S8 Fig ) . This behavior is reminiscent of the fact that the ergodicity-breaking transition for a spin glass in the SK model is also a third-order phase transition ( see Figure 3 in ref . [60] ) . These similarities suggest that the geometric properties of the probability landscape of the neural population are somewhat akin to the properties of the SK model in the spin glass phase , which is appealing for the connection between phase transitions and clustering of neural activity . But again , we are not arguing that our probability distributions over neural activity are exactly reproduced by the SK model . For instance , there are no local fields in the SK model , and these play a significant role in the properties of our maximum entropy models of neural data . Fundamentally , error-correcting structure is not present in populations of independent neurons: if one neuron in an independent population misfires , that neuron’s information is lost [66] . The qualitative separability between the regimes of error-correction and independence is our proposed origin of the observed phase transition with respect to our temperature variable [19 , 21] . To summarize these ideas , we present a suggested picture of the phase transitions studied in our work , by comparison with the Ising ferromagnet and the SK spin glass models ( Fig 10 ) . In the Ising ferromagnet at low temperature ( Fig 10A ) , the constant and equal interactions cause all the cells to tend to be active or quiet simultaneously . At high temperature , fluctuations wash out the interactions and the system is in a weakly correlated ( paramagnetic ) phase . As temperature is decreased at zero applied field , the system reaches a critical point where it chooses the all-active or all-quiet half of phase space ( grey dot in Fig 10A ) . Below the critical temperature the two halves of phase space are separated by a line of first-order phase transitions that separates a ferromagnetic phase with all the spin aligned in one direction from a similar ferromagnetic phase with all the spins aligned in the opposite direction ( black line in Fig 10A ) . As described above , the SK model has both a spin glass and a ferromagnetic limit ( Fig 10B , sketched following [67] ) . In the SK model in the spin glass limit ( μJ ≪ σJ ) , decreasing the temperature at zero applied field causes the model to freeze into a spin glass phase . This ergodicity-breaking phase transition is third-order ( Fig 10B , magenta arrows ) . In our picture of the retinal population code , a third-order phase transition occurs when the correlation strength α is increased from 0 at T = 1 ( gray dot and magenta arrow in Fig 10C ) . In this phase transition , the distribution of neural activity wrinkles to form a set of ridges in the probability landscape . Because firing rates are constrained during this manipulation , this manipulation is analogous to varying temperature at zero applied field in the nearest neighbor Ising ferromagnet , and to varying temperature in the spin glass limit of the SK model . Increasing the temperature in our models of the retinal population when the correlation α is greater than α* causes the distribution to melt to a weakly correlated state in a transition that is first-order ( see Fig 7B ) . Because of the local fields in our models , variations in temperature are accompanied by changes in firing rate , and this is most similar to varying the applied field , h , in the Ising ferromagnet . The analogy between the axes in the Ising ferromagnet and our model only extends to the horizontal ( temperature ) axis of the SK model: here the phase space far from the dotted line is shown for clarity . So in summary , the main phase transition that we see when we change temperature is first-order , as is the transition between ferromagnet states in the Ising model when the applied field is changed . In both cases , there is a substantial change in the state of individual elements—the firing rate of neurons in the retinal population and the magnetization of spins in the Ising model . Furthermore , the 3rd-order phase transition that we observe when α increases above α* is reminiscent of the 3rd-order phase transition in the SK model as a function of temperature in the limit of highly variable interactions . In our case , increasing α increases the strength of correlations , while in the SK model , lowering temperature increases the impact of interactions on the network state . This analogy provides support for the conclusion that the low temperature phase of the retinal population resembles a spin glass . In statistical physics , the peak in the specific heat that we observed in our models ( Fig 3 ) could be consistent with two types of phase transition , which are classified by the order of the derivative of the free energy which exhibits a discontinuity . In statistical mechanics , every physical system can be described on a macroscopic level by a free energy function . When the system transitions between phases , some order of the derivative of this free energy function will have a discontinuity . The first-order derivative of free energy versus temperature is the entropy , so when there is a discontinuity in the entropy , the system is said to exhibit a ‘first-order’ phase transition . The second-order derivative of free energy versus temperature is proportional to the specific heat , so a system with a discontinuity in the heat capacity exhibits a ‘second-order’ phase transition . The ambiguity we suffer in interpreting our data arises due to the fact that phases and transitions are rigorously defined in the thermodynamic limit which is when the system size N is taken to infinity . At finite sizes , it may be difficult to tell these two possibilities apart . The first possibility is that the observed phase transitions exhibit a divergence in the specific heat , making them second-order . This would indicate criticality in the retinal population code [19 , 21 , 27] . The second possibility is that the entropy is discontinous and that the specific heat exhibits an infinite value only at the critical temperature ( i . e . the specific heat has a delta-function form ) . This would be a first-order transition , and it does not indicate criticality . Both of these hypotheses would be consistent with a sharpening of the specific heat as the system size increased ( Fig 3 ) . How could we distinguish definitively between these two types of transition and what are the consequences of the difference ? This issue would be resolved if we could convincingly relate the properties of our model to some well known example in physics that does have a critical point . There are two examples that we have in mind here . The first is the symmetry-breaking mechanism for criticality ( see for example chapter 142 onwards in [5] ) . For concreteness we’ll take the example of the structure in the nearest neighbor Ising ferromagnet , where all the non zero interactions Jij are a positive constant and there are no other parameters . In the high temperature state , the system has the ability to visit any one of the 2N possible states , and the average firing probability for each neuron is 0 . 5 in a single time bin . In the low temperature state , the interactions cause all the cells to behave similarly , either mostly silent or mostly active , with some fluctuations allowed by the temperature . Importantly , however , the two phase spaces centered on all-silent and all-active are separated by an energy barrier that increases with system size . Because of this property , at large enough system sizes the available phase space in the low temperature state is reduced by a factor of two . The low and high temperature phases are known as low and high symmetry states , respectively , and it is the change in symmetry at the phase transition which leads to discontinuities in the specific heat . The critical point here is thus characterized by a reduction in symmetry where the symmetry described in this example is the deviation of the average firing rate from one half . In our models , the presence of local fields means that all cells have some bias towards silence or activity . As a consequence , the symmetry in firing rate is absent regardless of the temperature , and no symmetry breaking can occur with respect to changes in firing rates . There might be some other symmetry that is broken at the transition between phases , but no one has identified it yet [19 , 21] . Failing to identify a symmetry breaking mechanism in our analysis does not prove that the peak in the heat capacity is a first-order phase transition , but it is consistent with this interpretation . The second type of critical point that we’ve considered is the type that occurs in a spin glass , as in the SK model . These critical points are also characterized by a reduction in the size of the available state space , as in the symmetry breaking example described above . Because these transitions occur between paramagnetic and glassy regimes , they are also candidates for describing the transition that we observe in Fig 3 . However , these critical points in glassy systems are not , to our knowledge , characterized by a divergence in the specific heat . The second-order transition in the SK model is discontinuous in the specific heat , but not divergent ( see Figure 3 in ref . [60] ) . If such a critical transition did occur , it would not lead to a sharpening of the specific heat as we observe in Fig 3 . Instead , it would resemble the transition we observe as a function of increasing correlation strength , α , where the specific heat is either discontinuous or has a cusp at α* ( Fig 7 and see our discussion in the supplement , S8 Fig ) . Additionally , such critical points are typically marked by a divergence in some higher order statistic , such as the nonlinear susceptibility . Our measurements of the nonlinear susceptibility have not shown such a result ( S11 Fig ) . To summarize , we have considered two natural mechanisms that could connect our observed peak in the heat capacity versus temperature to a well known critical point in statistical physics , and we find that these mechanisms are simply not consistent with the observed properties of the data . It is possible that some other analogy provides a connection between criticality and our observations , but we are not aware of it . So , taken together , we believe that these observations argue against the hypothesis that the retinal ganglion cell population is poised at a critical point . Criticality implies that there is something special in the distribution of neural responses: for example , that the specific heat is maximized with respect to some properties of the retinal circuit , and hence that “the distribution of equivalent fluctuating fields must be tuned , rather than merely having sufficiently large fluctuations” [19] . Our analysis in which we scaled the strength of all pairwise correlations by a constant factor is not consistent with the notion of fine tuning . Specifically , we could decrease pairwise correlations by a factor of more than 2 without significant changes in the specific heat ( Fig 7C ) or the geometric structure of the probability distribution ( Fig 8 ) . Additionally , we actually observed higher peaks in the specific heat at T ≠ 1 in the partially correlated networks than in the fully correlated networks ( Fig 7 ) . This fact also argues against the idea that the heat capacity of the system is strictly maximized by some principle requiring fine tuning . The interpretation of the peak in the heat capacity as a first-order phase transition also provides an explanation for the proximity of the system to the transition . Our argument is that correlations in the distribution create a phase which is qualitatively different than the high temperature phase . This qualitative difference then requires that there be a a transition between these two regimes . Furthermore , the sparseness of neural activity implies that the zero-temperature limit of the model is the all-silent state . Consequently , the average firing rates must follow a sigmoidal function of temperature , starting at zero for T = 0 , rising for T > 0 , and then saturating at 0 . 5 firing probability for T → ∞ . This sigmoid sharpens into a step with increasing system size , with the width of the step corresponding to the area over which finite-size effects smear out the phase transition ( see Fig 11 ) . Such a step-like change in the firing rate is most consistent with a first-order phase transition ( where first-order derivatives of the free energy , such as the firing rates , change discontinuously ) . Given this context , constraining the average firing rate to be some small but nonzero value forces the system to be poised in the vicinity of the phase transition . Thus , we believe that the proximity of the system to the transition point simply follows as a consequence of constraining the structured phase to have neurons with low firing rates . In addition , it is difficult to reconcile the notion that the retinal code is poised at a very special , fine-tuned operating point with the observation that similar peaks in the specific heat arise in many circumstances , including very simple models [27] . In this sense , we agree with the conclusion that Nonenmacher and colleagues reached that this notion of fine tuning is not well supported by these wider considerations . In contrast , a first-order phase transition does not imply that something is “special” or “optimal” about the retinal population code . Instead , the hypothesis that the population is in a low temperature state is appealing , because this state can be robustly present without fine tuning . To sum up , our heat capacity analyses do not clearly disprove the hypothesis that the system is poised at a critical point . However , first-order phase transitions are far more common in nature than second-order transitions . They do not require any special tuning of the parameters of the system . Thus , the interpretation that the neural population is in a low temperature state serves as a simple hypothesis that is consistent with all of our data . This interpretation has additional value , because it suggests a connection between the phase transition and the emergence of structure in the probability landscape . And in fact , our analyses have directly confirmed this connection ( Figs 8 and 9 ) . While we believe there are no inconsistencies in our empirical findings and those of other studies that focused on criticality in retinal population codes , our interpretations differ substantially [24 , 25 , 27] . A recent study tested the generality of phase transitions in simulations of retinal ganglion cells [27] , simulating networks of neurons in the retina , and then estimating the specific heat following procedures similar to those presented here . Their results are consistent with ours in that phase transitions were robustly present in different networks of neurons , and the presence of these phase transitions was largely invariant to experimental details . Similar to us , they also found that the sharpness of the peak in the specific heat was systematically enhanced by stronger pairwise correlations . Because Nonnenmacher et al . also found this behavior in very simple models , such as homogeneous neural populations , they concluded that the sharpening peak in the heat capacity does not necessarily provide insight , by itself , into the structure of the population code . We agree . However , because we went on to analyze the structure of the probability landscape for our real , measured neural populations , we could show that the emergence of the peak was related to the clustering of neural activity patterns ( Figs 8 and 9 ) . The relationship between the low temperature phase and clustering is not understood in general . Studies of homogeneous models of neural populations demonstrate that the low temperature phase is not sufficient for clustering [27] , while the current study suggests that the low temperature phase might be necessary . Other factors , such a sufficient heterogeneity of single neuron properties , are presumably also required . In any case , we interpret the robustness of a phase transition in correlated neural populations not as an argument that this property is trivial , but instead as evidence for the generality of clustering in neural population codes . A separate line of work has investigated the presence of Zipf-like relationships in the probability distribution of neural codewords [23–25] . A true Zipf Law is intimately related to a peak in the heat capacity at T = 1 , and Schwab et al . found that a Zipf Law was present under fairly generic circumstances , in which neural activity was determined by a latent variable ( e . g . , an external stimulus ) with a wide range of values [24] . Again , this result is broadly consistent with our finding of great robustness of the low temperature state , and we interpret this a positive evidence for these properties being generically present in correlated neural populations . However , we can’t make more detailed comparisons to [24 , 25] , because we have not chosen to analyze Zipf-like relationships in our experimental data . There are several reasons for this choice: 1 ) we can only sample ∼2 orders of magnitude in rank ( S9 Fig ) , making it difficult to estimate power law exponents; 2 ) we typically observe small deviations from the power law trend , and we are uncertain about how to interpret the importance of these “bumps” . In our study , we have characterized the neural response in a single time bin , ignoring the role of temporal correlations across time bins . One can extend our approach to include temporal correlations by concatenating multiple time bins into each neural codeword . When the number of total time bins was systematically increased in such a manner , the peak in the specific heat sharpened substantially [21] . The authors interpreted these results as further evidence in favor of the critical properties of neural population codes . However , in all cases in both our study and of [21] , the peak was above T = 1 , consistent with our interpretation that neural populations are in a low temperature state . Since increasing the number of time bins this way drastically increases the complexity of the distribution , this treatment of temporal correlations increases the structure of the low temperature state in a manner similar to an increase in the number of neurons analyzed together , N . Using several different analysis methods , neural activity evoked by repeated presentations of the same stimulus has been shown to form clusters in the space of all possible activity patterns . Zero temperature descent in the energy landscape defined by the maximum entropy model mapped a large fraction of all neural activity patterns to non-silent energy basins , which were robustly activated by the same visual stimulus [16 , 68] . Mapping neural activity patterns to latent variables inferred for a hidden Markov model revealed similar robust activation by the stimulus [63 , 65] . Huang et al . found a form of first-order phase transition as a function of the strength of an applied external field , from which they concluded that the energy landscape formed natural clusters of neural activity with no applied field [69] . Ganmor et al . recently uncovered a striking block diagonal organization in the matrix of semantic similarities between neural population codewords [62] , arguing for a clustered organization of neural codewords . All these analyses are likely to be different ways to view the same underlying phenomenon , although a detailed exploration of the correspondences among these methods is a subject for future work . Are our results specific to the retina ? In our approach , the collective state of the retinal population code is entirely determined by the pattern and strength of measured correlation . There is nothing about this pattern of correlation that makes specific reference to the retina . This means that any neural population having similar firing rates and pairwise correlations would also be in a similar collective state . Additionally , the strength of pairwise correlations we report here is smaller than or comparable to those reported in higher order brain areas , such as V1 and MT [57 , 58 , 70 , 71] . This suggests that the collective state of neural activity , which arises due to a clustering of neural activity patterns , could occur throughout higher-order brain regions in population recordings of a suitable size ( N >100 ) .
This study was performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health . The protocol was approved by the Institutional Animal Care and Use Committee ( IACUC ) of Princeton University ( Protocol 1828 ) . We recorded from larval tiger salamander ( Ambystoma tigrinum ) retina using the dense ( 30 μm spacing ) 252-electrode array described in [1] . In Experiment #1 , which probed the adaptational state of the retina at normal and low ambient illumintation levels , the salamander was kept in dim light and the retina was dissected out with the pigment epithelium intact , to help the retina recover post dissection and adjust to the low ambient light levels in the dark condition . The rest of the procedure in Experiment #1 , and the full procedure for Experiment #2 , followed [1] . The chromatic checkerboard stimulus ( CC ) consisted of a random binary sequence per color ( R , G , B ) per checker , allowing 8 unique values for any given checker . Checkers were 66 μm in size , and refreshed at 30 Hz . There were two ( gray scale , 8 bit depth ) natural movies used: grass stalks swaying in a breeze ( M1 , 410 seconds ) and ripples on the water surface near a dam ( M2 , 360 seconds ) . Both were gamma corrected for the CRT , and displayed at 400 by 400 pixel ( 5 . 5 μm per pixel ) resolution , at 60 Hz . In Experiment #1 , after adapting the retina to the absolute dark for 20 minutes , we recorded in the dark condition first ( by placing an absorptive neutral density filter of optical density 3 [Edmund Optics] in the light path ) , stimulating with ( CC ) for 60 minutes , and with ( M1 ) for 90 minutes . The filter was then switched out for the light condition , in which we recorded for an additional 60 minutes of ( M1 ) and another 60 minutes of ( CC ) . To avoid transient light adaptation effects we removed the first 5 minutes of each recording ( 10 minutes from the first checkerboard ) from our analysis . During stimulation with ( M1 ) we sampled 340 sec long segments from ( M1 ) with start times drawn from a uniform distribution in the [0 60] second interval of ( M1 ) . The spike sorting algorithm [1] was run independently on the recordings in response to ( M1 ) at the two light levels , generating separate sets of cell templates at the two light levels , which were then matched across the two conditions , yielding N = 128 ganglion cells . The spike trains were then binned in 20 ms time bins and binarized , giving 2 . 5 ⋅ 105 states in the dark and 1 . 75 ⋅ 105 states in the light . For the recordings from the checkerboard ( in both light conditions ) , the templates from the light recording were used to fit the electrode activity . Across all four stimulus conditions this left us with N = 111 cells for the comparisons of natural movies to checkerboard . The checkerboard in the light condition was binned into N = 145623 states . In Experiment #2 , we alternated stimulation between ( M1 ) and ( M2 ) every 30 seconds , sampling 10 sec segments from both movies . For our analysis here we worked with the statistics of the last 9 . 5 seconds of each 30 second bout , yielding 8 ⋅ 104 states per stimulus condition , for N = 140 cells . Our maximum entropy model inference process implements a modified form of sequential coordinate gradient descent , described in [16 , 47 , 48] , which uses an L1 regularization cost on parameters . For the k-pairwise model , we inferred without a regularization cost on the local fields . Further details are given in the Supplement ( S1 Text ) To measure the heat capacity we simulated an annealing process . Initializing at high temperatures , we monte carlo sampled half a million states per temperature level ( in 100 parallel runs of 5 ⋅ 103 samples each ) , initializing subsequent lower temperature runs with the final states of preceeding higher temperature runs . The heat capacity at a particular temperature was then evaluated as C = ( 〈E2〉 − 〈E〉2 ) /T2 . Our model LN neurons were estimated over the chromatic checkerboard recording in the light condition . For each cell , the three color-dependent linear filters ( the full STA ) were weighted equally before convolution with the stimulus for an estimate of the linear response q . The non-linearity was estimated over the same data by Bayes rule , P ( spike|q ) = P ( q|spike ) P ( spike ) /P ( q ) . Spike trains were simulated from a novel pseudorandom sequence put through the model’s filters and non-linearity , with the non-linearity shifted horizontally to constrain the firing rates of the neurons to be the same as in the experimental recordings . The result was binned and binarized to yield N = 145623 states , for which we inferred the k-pairwise model .
|
Neurons encoding the natural world are correlated in their activities . The structure of this correlation fundamentally changes the population code , and these effects increase in larger neural populations . We experimentally recorded from populations of 100+ retinal ganglion cells and probed the structure of their joint probability distribution with a series of analytical tools inspired by statistical physics . We found a robust ‘collective state’ in the neural population that resembles the low temperature state of a disordered magnet . This state generically emerges at sufficient correlation strength , where the energy landscape develops an attractor-like structure that naturally clusters neural activity .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"action",
"potentials",
"membrane",
"potential",
"ocular",
"anatomy",
"electrophysiology",
"neuroscience",
"probability",
"distribution",
"mathematics",
"ganglion",
"cells",
"systems",
"science",
"neuronal",
"tuning",
"computer",
"and",
"information",
"sciences",
"animal",
"cells",
"probability",
"theory",
"cellular",
"neuroscience",
"retina",
"cell",
"biology",
"anatomy",
"neurophysiology",
"physiology",
"neurons",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"ocular",
"system",
"physical",
"sciences",
"afferent",
"neurons",
"retinal",
"ganglion",
"cells",
"dwell",
"time"
] |
2017
|
The structured ‘low temperature’ phase of the retinal population code
|
Klotho acts as a co-receptor for and dictates tissue specificity of circulating FGF23 . FGF23 inhibits PTH secretion , and reduced Klotho abundance is considered a pathogenic factor in renal secondary hyperparathyroidism . To dissect the role of parathyroid gland resident Klotho in health and disease , we generated mice with a parathyroid-specific Klotho deletion ( PTH-KL−/− ) . PTH-KL−/− mice had a normal gross phenotype and survival; normal serum PTH and calcium; unaltered expression of the PTH gene in parathyroid tissue; and preserved PTH response and sensitivity to acute changes in serum calcium . Their PTH response to intravenous FGF23 delivery or renal failure did not differ compared to their wild-type littermates despite disrupted FGF23-induced activation of the MAPK/ERK pathway . Importantly , calcineurin-NFAT signaling , defined by increased MCIP1 level and nuclear localization of NFATC2 , was constitutively activated in PTH-KL−/− mice . Treatment with the calcineurin-inhibitor cyclosporine A abolished FGF23-mediated PTH suppression in PTH-KL−/− mice whereas wild-type mice remained responsive . Similar results were observed in thyro-parathyroid explants ex vivo . Collectively , we present genetic and functional evidence for a novel , Klotho-independent , calcineurin-mediated FGF23 signaling pathway in parathyroid glands that mediates suppression of PTH . The presence of Klotho-independent FGF23 effects in a Klotho-expressing target organ represents a paradigm shift in the conceptualization of FGF23 endocrine action .
Calcium plays a pivotal role in many biological processes , such as intra-cellular signaling , cell membrane depolarization and excitation , energy metabolism and skeletal mineralization . Accordingly , a fine-tuned regulation of serum calcium level is a prerequisite for normal cellular and organ function in most organisms . Parathyroid hormone ( PTH ) is the principal hormonal regulator of circulating calcium as it rapidly increases its renal tubular reabsorption and mobilization from bone deposits in response to a decrease in serum calcium [1] . In turn , free calcium ions can efficiently inhibit PTH secretion as part of an endocrine feedback loop mediated by the calcium-sensing receptor ( CaSR ) located on parathyroid chief cells [2] . Type I membrane-bound alpha-Klotho ( Klotho ) defines tissue specificity for the phosphaturic hormone fibroblast growth factor-23 ( FGF23 ) by acting as a permissive co-receptor [3] . Klotho is predominantly expressed in organs requiring abundant calcium transport such as kidneys , parathyroid glands and choroid plexus [4] . In the parathyroids , FGF23 binds to binary complexes of an FGF receptor ( FGFR ) and Klotho to suppress PTH secretion [5] , [6] . Klotho activity on the other hand has been implicated as fundamental for the stimulation of PTH secretion during hypocalcemic conditions [7] , although the underlying mechanism has been challenged [8] . Secondary hyperparathyroidism ( sHPT ) is a common manifestation in chronic kidney disease ( CKD ) despite markedly increased serum FGF23 concentrations . This presumably reflects parathyroid resistance to FGF23 action , which was also supported by lack of response to FGF23 injections in a rat model of CKD [9] , [10] . The proposed mechanism underlying such FGF23 resistance is decreased abundance of parathyroid Klotho and FGFRs [11] , [12] . To dissect the role of parathyroid gland resident Klotho in physiology and in pathophysiological states such as CKD , we generated a novel mouse strain harboring a parathyroid-specific deletion of the Klotho gene . The present study sheds new light on the function of parathyroid Klotho and identifies a novel , Klotho-independent signaling pathway of FGF23 that is involved in the regulation of PTH secretion .
Mice with a parathyroid specific deletion of Klotho ( PTH-KL−/− ) were generated using Cre-LoxP recombination ( Figure S1 ) . Floxed Klotho mice were crossed with mice expressing Cre recombinase driven by the human PTH promoter , which was previously shown to have Cre activity exclusively in the parathyroid glands [13] . Successful deletion of parathyroid Klotho protein was confirmed with immunohistochemical staining of thyro-parathyroid tissue ( Figure 1 ) . Overall efficiency of deletion varied , and was up to >90% in investigated samples . Subanalyses of mice with the most efficient deletion showed similar results to the full analyses . Adult PTH-KL−/− mice were viable , fertile and did not display any gross physical or behavioral abnormalities . Survival was similar to wild-type littermates during the study with no mortality up to 6 months of age . Female PTH-KL−/− mice had reduced body weight and crown-rump length compared to wild-type littermates ( p<0 . 05; Figure 1 ) , whereas no such differences were found in male PTH-KL−/− mice . Serum biochemistries in 8-week-old mice are shown in Figure 1 . Interestingly , the serum levels of 1 , 25-dihydroxy vitamin D3 ( 1 , 25 ( OH ) 2D ) were doubled in PTH-KL−/− mice compared to their wild-type littermates ( p<0 . 05 ) , while PTH , calcium and FGF23 remained normal . Serum phosphorous ( wild-type vs PTH-KL−/−; 9 . 54±0 . 38 vs 9 . 39±0 . 32 , p>0 . 05 ) , creatinine ( 0 . 31±0 . 03 vs 0 . 26±0 . 02 , p>0 . 05 ) , urinary concentrations of calcium/creatinine ( 0 . 53±0 . 19 vs 0 . 53±0 . 19 , p>0 . 05 ) and phosphorous/creatinine ( 19 . 1±3 . 4 vs 19 . 4±3 . 0 , p>0 . 05 ) were also unaltered in PTH-KL−/− mice . To exclude the possibility of early onset changes in mineral metabolism , we analyzed serum from 3-week-old animals . No differences were seen for serum calcium ( wild-type vs PTH-KL−/−; 10 . 00±0 . 22 vs 9 . 86±0 . 20 , p>0 . 05 ) or PTH ( 118 . 0±13 . 4 vs 133 . 0±21 . 8 , p>0 . 05 ) . We investigated potential secondary effects that the parathyroid-specific deletion of the Klotho gene might have on the main target organs of PTH signaling , namely bone and kidney . There were no significant histological changes in bone from 6-week old PTH-KL−/− mice and bone mineral density was unaltered compared to wild-type mice ( Figure S2 ) . Renal histology was also normal and no transcriptional changes were found for Klotho , VDR , Cyp27b1 , Cyp24a1 , Npt2a , TRPV5 or CaSR in kidneys from PTH-KL−/− mice compared to wild-type mice ( Table S1 ) . Parathyroid size , histology and proliferation index , defined as Ki67 positive cells/total number of cells , was not affected by the Klotho gene deletion ( wild-type vs PTH-KL−/− , 2 . 9% vs . 2 . 8% , p>0 . 05 ) ( Figure S3A ) . Immunofluorescence staining showed no significant changes in protein expression for PTH , CaSR , VDR ( Figure S3B ) , FGFR1 or Cyp27b1 ( data not shown ) suggesting that Klotho does not regulate the expression of these proteins in an autocrine or paracrine fashion under physiological conditions . The expression of approximately 90 genes critical for parathyroid function was examined using a nanostring array . The data are compiled in Table S2 . In addition to Klotho , changes in expression level were observed for genes important for transcriptional and metabolic control , such as Cfd ( Entrez Gene: 11537 ) , Fabp4 ( 11770 ) and Smad4 ( 17128 ) ( Figure 2 ) . Gli3 ( 14634 ) , Pin1 ( 23988 ) , sFRP3 ( 20379 ) and FGF20 ( 80857 ) were also expressed at different levels in the parathyroids of PTH-KL−/− mice compared to wild-type mice , although at a low absolute level . Notably , the expression of some of the genes that play a key role in the function of the parathyroid gland , such as PTH , Gata3 , CaSR and VDR , or in FGF23 physiology , such as FGFR1 and the vitamin D regulatory enzymes Cyp27b1 and Cyp24a1 , were not significantly affected by the knockdown of parathyroid Klotho gene expression . The serum calcium level of PTH-KL−/− and wild-type mice was either rapidly decreased or increased by an intraperitoneal injection of EGTA or calcium-gluconate , respectively [14] . The parathyroid response to alterations in serum calcium , as measured by serum PTH level , did not differ between PTH-KL−/− mice and wild-type mice ( wild-type vs PTH-KL−/− , R2 = 0 . 77 , p<0 . 0001 vs R2 = 0 . 74 , p<0 . 0001 ) ( Figure S4 ) , supporting that parathyroid Klotho is not essential for parathyroid sensitivity to acute changes in serum calcium , and its response to them with altered PTH secretion . Renal insufficiency is associated with the development of sHPT . To test whether parathyroid function is different in PTH-KL−/− mice under conditions of renal insufficiency , we induced renal failure using an adenine-based protocol [15] . Four weeks after induction of renal insufficiency , PTH-KL−/− mice exhibited sHPT just like their wild-type littermates did . Its severity as well as changes in other markers of mineral metabolism was similar among PTH-KL−/− and wild-type mice ( Figure 3A and Table S3 ) . Notably , serum FGF23 levels were increased by approximately 50-fold in both PTH-KL−/− and wild-type mice with renal failure compared to mice with preserved renal function . Based on our findings that the parathyroid glands of PTH-KL−/− mice retained calcium sensitivity and that sHPT developed similarly as in wild-type mice upon induction of renal insufficiency , we reasoned that the parathyroid response to FGF23 might also be preserved in PTH-KL−/− mice . To test this , we injected mice intravenously with a single dose of recombinant FGF23 and measured serum concentrations of PTH before and 15 min after the injection . As shown in Figure 3B , FGF23 caused a decrease in serum levels of PTH in PTH-KL−/− mice of similar magnitude as in wild-type mice . We next explored which signaling pathway might mediate the inhibition by FGF23 on PTH secretion in PTH-KL−/− mice . Activation of the MAPK cascade was first examined using immunofluorescence staining . As shown in Figure 3C , there was significantly less immunostaining for phosphorylated ERK1/2 in parathyroid tissue from PTH-KL−/− mice , and a complete co-localization with residual Klotho protein . These data indicate that activation of the MAP kinase pathway by FGF23 is markedly suppressed or absent in Klotho-deficient parathyroids . We next tested whether the calcineurin-nuclear factor of activated T cells ( NFAT ) pathway , another proposed downstream signaling pathway of FGFR activation [16] , [17] , [18] , was activated in the parathyroids of PTH-KL−/− mice in response to FGF23 . To this end , we analyzed gene expression of calcineurin and NFAT in parathyroid tissue from wild-type mice . Indeed , all subunits of calcineurin and all four calcium-regulated members of the NFAT family ( NFATC1–NFATC4 ) were expressed in the parathyroid of wild-type mice ( Figure 4A ) . We then examined by immunofluorescence microscopy the subcellular distribution of the NFAT proteins in parathyroid tissue from PTH-KL−/− mice that had been treated with FGF23 . While parathyroid tissue from FGF23 treated wild-type mice showed cytoplasmic immunostaining for NFATC2 , the parathyroid tissue of FGF23 treated PTH-KL−/− mice also showed nuclear localization of NFATC2 ( Figure 4B ) . No clear difference was seen for the other NFATs . In addition , immunostaining analysis for the modulatory calcineurin interacting protein 1 ( MCIP1 ) , a facilitator of calcineurin activity , revealed that MCIP1 expression was markedly upregulated in Klotho-deficient parathyroid tissue compared to wild-type tissue ( Figure 4B ) [19] . Together , the data suggest that the calcineurin-NFAT pathway is activated in the parathyroid of PTH-KL−/− mice , and responds to treatment with FGF23 . To provide in vivo evidence for the role of the calcineurin-NFAT pathway in mediating the suppression by FGF23 of PTH secretion , PTH-KL−/− and wild-type mice were pre-treated with cyclosporine A ( CsA ) , a calcineurin-inhibitor , prior to injection of FGF23 . As shown in Figures 4C and D , the CsA pre-treatment nearly abolished the effect of FGF23 on PTH secretion in PTH-KL−/− mice , whereas it did not impact this FGF23 action in wild-type mice . To investigate whether the parathyroid calcineurin-NFAT pathway acts independent of serum factors such as soluble Klotho , thyro-parathyroid explants from wild-type and PTH-KL−/− mice were treated with FGF23 with or without pre-treatment of CsA [20] . In explants without pretreatment of CsA , 2 h treatment with FGF23 ( 10 ng/mL ) decreased PTH secretion similarly in wild-type and PTH-KL−/− mice ( Figure 4E ) . Conversely , in explants pre-treated with CsA ( 0 . 83 µM ) for 2 h the response to FGF23 treatment was unaltered in wild-type mice but completely blunted in PTH-KL−/− mice ( ΔPTH relative to baseline; 0 . 81 vs 1 . 17 , p<0 . 01 ) .
Regulation of PTH secretion is a fundamental process in calcium homeostasis , yet our current models are constantly revised due to the complex and multi-dimensional control of PTH . Recent data suggest that , in addition to calcium-sensing receptor and vitamin D receptor activation , the FGF23-Klotho endocrine axis negatively regulates PTH secretion . However , previous discordant results of FGF23-Klotho function in physiology and in pathophysiological states such as renal sHPT prompted us to further investigate the concerted parathyroid action of FGF23-Klotho in vivo . We addressed this issue by generating a novel mouse model harboring a parathyroid-specific deletion of Klotho and our principal findings are that i ) absence of parathyroid Klotho does not alter the acute parathyroid response of PTH secretion and calcium sensitivity as previously suggested; ii ) absence of parathyroid Klotho alone does not contribute to the development of renal sHPT; iii ) in the absence of parathyroid Klotho , the calcineurin-NFAT pathway is constitutively active and mediates the suppression by FGF23 of PTH secretion . The activation of this compensatory mechanism in Klotho-deficient parathyroids might explain the lack of a hyperparathyroid phenotype in PTH-KL−/− mice . We postulated the existence of a Klotho-independent signaling pathway of FGF23 in the parathyroids based on the findings that PTH-KL−/− mice had unaltered serum PTH levels , a intact parathyroid response to FGF23 injections and that hyperparathyroidism was not aggravated in the setting of renal failure despite an effective ablation of parathyroid Klotho and an apparent lack of MAPK activation in response to FGF23 . Because FGF23 was recently shown to promote pathological growth of the myocardium and proposed as a mediator of left ventricular hypertrophy through calcineurin-dependent mechanisms [16] , we speculated that this pathway might also be activated by FGF23 in the parathyroids . Indeed , there was a marked upregulation of the calcineurin-modulator MCIP1 and a more prominent nuclear localization of NFATC2 in PTH-KL−/− mice treated with FGF23 , strongly supporting a compensatory activation of the calcineurin-NFAT pathway . The functional importance of this pathway was confirmed in vivo and ex vivo where pretreatment with the calcineurin inhibitor CsA abolished the PTH response to FGF23 treatment in PTH-KL−/− mice but not in wild-type mice . Importantly , our data are also consistent with a previous report demonstrating increased PTH transcript levels in calcineurin Aβ null mice and posttranslational regulation of PTH in vitro by CsA [21] . The blunted PTH response to FGF23 injection in PTH-KL−/− mice indicates that parathyroid FGF23 signaling is dependent on two principal pathways , Klotho-FGFR activation and calcineurin activation respectively . However , the exact intra-cellular mechanisms mediating FGF23 suppression of PTH remain to be defined . Importantly , the expression of FGFR1 , which has been implicated as the relevant FGFR for parathyroid FGF23 signaling , was unaltered in PTH-KL−/− mice . This supports that their FGF23-driven calcineurin activation occurs without a compensatory increase in its target FGFR . The identification of a novel FGF23-calcineurin signaling pathway in the parathyroid glands has several implications . First , it challenges the current concepts of FGF23-mediated PTH regulation and secretion . Second , it raises the possibility of previously neglected Klotho-independent down-stream effects of FGF23 . This should be put in the context that abnormally high levels of FGF23 in humans are associated with adverse clinical outcomes including cardiovascular disease , CKD progression rate and mortality [22] , [23] , [24] , [25] , [26] , [27] , [28] . Indeed , FGF23 may contribute to pathological left ventricular hypertrophy in a Klotho-independent manner [16] . Future studies are warranted to explore Klotho-independent FGF23 signaling at the systemic level . Although FGF23 appears to signal independently of Klotho in both parathyroid glands and cardiomyocytes [16] there is a fundamental difference in its signaling in these tissues . In the heart , FGF23 signaling is an ‘off-target’ effect in a non-classical , non-Klotho expressing organ . In contrast , in parathyroid glands it may rather be viewed as an ‘on-target’ effect in a Klotho-expressing target organ which represents a paradigm shift in the conceptual framework of FGF23 endocrine action . SHPT is an inevitable complication in CKD patients with advanced renal failure . Serum FGF23 concentrations rise in parallel with a decline in glomerular filtration rate and are markedly elevated at later stages of CKD [29] . Because PTH stimulates FGF23 synthesis in bone and FGF23 signals back to the parathyroids and suppresses PTH secretion [30] , a reasonable assumption is that FGF23 rises in CKD in part to counteract sHPT . Indeed , rats subjected to parathyroidectomy prior to induction of renal failure did not respond with increased FGF23 levels [30] . However , the unabated development of sHPT in the face of extreme FGF23 elevations poses a dilemma: either the inhibitory action of FGF23 on PTH is not relevant in vivo or the parathyroids may lack responsiveness to FGF23 in CKD . The latter option is supported by studies in CKD rat models demonstrating an attenuated or abolished parathyroid response to FGF23 [9] , [10] . Parathyroid FGF23 resistance corroborates data showing a reduction in both Klotho and FGFRs in animal models and in human-derived surgically resected hyperplastic parathyroid glands [11] , [12] . Our data unequivocally support that parathyroid FGF23 resistance is not primarily induced by Klotho insufficiency . Speculatively , reduced expression of the cognate FGFR ( s ) could be the principal mechanism underlying this phenomenon . Another potentially relevant implication of our findings is that a large portion of CKD patients receiving a renal allograft suffer from persistent sHPT unproportional to their residual kidney function in the face of high systemic FGF23 [31] . We speculate that this might be due to a dual blocking of FGF23 signaling , both through the endogenously decreased abundance of parathyroid FGFRs/Klotho and the superimposed inhibition of the calcineurin-NFAT pathway by calcineurin inhibitors ( e . g . cyclosporine and tacrolimus ) commonly used as immunosuppressive agents . This hypothesis and the proposed role of parathyroid resident Klotho is modeled in Figure 5 . The role of parathyroid Klotho remains controversial due to previous conflicting data . Klotho was reported to modulate parathyroid Na+/K+-ATPase activity via direct interaction with its α1-subunit causing an increased abundance of plasma membrane Na+/K+-ATPase , which in turn promotes PTH secretion through an increased electrochemical gradient [7] . However , this mechanism was only functional at low extra-cellular calcium concentrations and another study demonstrated that blocking of the Na+/K+-ATPase with the digitalis glycoside ouabain did not affect the PTH secretory response to acute hypocalcemia [8] . This study underscores that parathyroid Klotho does not alter parathyroid sensitivity to acute fluctuations in serum calcium . Given the apparent normal phenotype of PTH-KL−/− mice and their intact response to FGF23 it remains uncertain what other functions parathyroid Klotho may have . The cause of increased serum 1 , 25 ( OH ) 2D in PTH-KL−/− mice is unknown since expression levels of Cyp27b1 and Cyp24a1 are unaltered in kidneys and parathyroid glands . Yet , it is conceivable that parathyroid Klotho has a functional impact on vitamin D metabolism in PTH-KL−/− mice and that their elevated 1 , 25 ( OH ) 2D level contributes to normalizing PTH and counteracts development of hyperparathyroidism . Further , we speculate that Klotho is an important local transcriptional regulator of genes involved in sustained challenge of parathyroid function or having a long-term impact on parathyroid function . In support of this idea , PTH-KL−/− mice had altered transcript levels of several transcriptional regulators including Cfd , Smad4 , Fabp4 and Pin1 . The increased level of Pin1 in PTH-KL−/− mice is intriguing given its role in destabilizing PTH mRNA [32] and may represent yet another protective intracellular mechanism against hyperparathyroidism when FGF23-Klotho signaling is chronically disrupted . In sum , we dissected the physiological and pathophysiological role of parathyroid Klotho and elucidated several critical and novel aspects in this regard . Most importantly , our study uncovered a Klotho-independent FGF23 signaling mechanism in parathyroid glands with potential implications for multiple disorders of mineral metabolism , especially in CKD .
Mice with a parathyroid-specific Klotho deletion were generated using Cre-Lox recombination as previously described [33] . Briefly , loxP sequences were introduced in the flanking regions of exon 2 of the Klotho gene . Floxed mice were crossed with mice expressing Cre recombinase under the human PTH promotor ( 129;FVB-Tg ( PTH-cre ) 4167Slib/J; Jackson laboratory , US ) and the offspring subsequently analyzed . Floxed littermates not expressing Cre were used as wild-type controls . Total DNA was extracted from tail biopsies and genotyping performed with standard PCR techniques . Mice were fed a standard chow ( RM1 , SDS , UK ) containing 0 . 73% calcium and 0 . 52% phosphorous . All animals had free access to food and drinking water . Blood sampling was performed by tail vein incision at intermediate time points and through the axillary artery at sacrifice . Animals were fasted 4 hour prior to blood sampling . All experiments were conducted in compliance with the guidelines of animal experiments at Karolinska Institutet and approved by the regional ethical board ( Stockholm South ethical committee ) . Calcium , phosphorous and creatinine were measured using quantitative colorimetric assay kits ( BioChain , US ) . Urine values were multiplied by 1000 . Serum PTH was measured using a Mouse Intact PTH ELISA kit ( Immutopics , US ) . For the FGF23 injection experiments the newer Mouse 1–84 PTH ELISA kit ( Immutopics ) and plasma samples were used . Serum 1 , 25-dihydroxy vitamin D was measured using a RIA kit ( IDS , US ) . Intact FGF23 was measured using an FGF23 ELISA kit ( Kainos Laboratories , Japan ) . Parathyroid glands and surrounding adhesive thyroid tissue were carefully removed in conjunction with sacrifice and immediately frozen in OCT . The parathyroid tissue was then dissected using laser capture microdissection . Thirty µm thick cryo-sections were cut and mounted on PEN foil–coated slides ( Leica , Germany ) . Sections were stained with the Arcturus histogene kit ( Applied Biosystems , US ) and microdissected with a Leica ASLMD microscope ( Leica ) . Total RNA was extracted by the Arcturus picopure RNA isolation kit ( Applied Biosystems ) . The nanostring nCounter system ( Nanostring technologies , US ) was used to obtain mRNA expression profiles . Data were processed using several normalization strategies , including quantile normalization and 6 house-keeping genes . Thyro-parathyroid tissue were dissected , fixed in 4% paraformaldehyde and embedded in paraffin . 4 µm sections were immersed in 3% H2O2 in methanol , treated with 4% normal serum and blocked with Avidin and Biotin ( Vector Laboratories , US ) . Sections were incubated with primary antibodies at 4°C overnight . For immunohistochemistry slides were incubated with biotinylated secondary antibodies followed by Vector ABC Reagent and developed with DAB substrate ( Vector Laboratories ) . For immunofluorescence , Alexa Fluor conjugated secondary antibodies were used for visualization ( Invitrogen , US ) . The primary antibodies used were rat monoclonal anti-Klotho ( KM2076 , TransGenic Inc . Japan ) mouse monoclonal anti-CaSR ( NB120-19347 , Novus Biologicals , US ) , mouse monoclonal anti-VDR ( sc-13133 , Santa Cruz Biotechnology , US ) , rabbit monoclonal anti-Ki67 ( SP6 , Thermo Scientific , US ) , rabbit monoclonal anti-Erk1/2 and anti-phospho-Erk1/2 ( Cell Signaling , US ) , mouse monoclonal anti-NFATc2 ( sc-7295 , Santa Cruz Biotechnology ) and rabbit polyclonal anti-MCIP1 ( a kind gift from Dr . Christian Faul ) [34] . Processing of undecalcified bone specimens and cancellous bone histology in the distal femoral metaphysis were performed according to standard protocols . The area within 0 . 25 mm from the growth plate was excluded from the measurements . µCT were performed using a Scanco Medical µCT 35 system ( Scanco , Switzerland ) . Investigation of the PTH-calcium relationship was performed by intraperitoneal injections with calcium-gluconate ( 300 µmol/kg ) or EGTA ( 450 µmol/kg ) dissolved in sterile saline as described elsewhere [14] . Blood samples were collected from tail vein for serum PTH and calcium measurements after 30 minutes . We employed an adenine-based model of renal failure in which mice were given adenine mixed in a casein containing diet . Mice were exposed to 0 . 3% adenine during the first 7 days followed by 0 . 2% adenine for 11 days , and finally 0 . 1% adenine for the rest of the study . Adenine was purchased from Sigma-Aldrich ( US ) , and the powdered casein-based diet from Special Diets Services . Recombinant human FGF23 protein ( A28 to I251 ) was produced as previously described [35] , [36] . In all experiments , FGF23 was injected intravenously through the tail vein after a 4-hour fasting period at the dose 0 . 15 mg/kg dissolved in 300 µl saline . Blood samples were drawn 15 min after injection of FGF23 , and the animals were then sacrificed and thyro-parathyroid tissue immediately excised and fixed in 4% formalin for further analysis . Mice were gavage fed with CsA at the dose 60–150 mg/kg ( Sigma-Aldrich ) two hours prior to FGF23 injection . During anesthesia the trachea was severed superiorly and inferiorly to the thyroid gland . The trachea-thyro-parathyroid complex was excised and placed in 1 ml serum-free medium ( DMEM/F-12 , HEPES supplemented with insulin , transferrin and BSA ) as previously described [20] . Medium was replaced daily for 4 days before the experiment started , in accordance with the protocol by Ritter et al . On day 4 fresh medium was added and the tissue incubated for 2 h ( = baseline ) . The medium was then replaced with regular medium ( = control ) or medium containing 10 ng/mL FGF23 and incubated for 2 h . One group was pre-treated with medium containing CsA ( 0 . 83 µM ) for 2 h prior to incubation with FGF23 . All values were calculated as relative change compared to baseline . GraphPad Prism 5 . 0 ( GraphPad Software Inc , US ) was used for statistical analysis . Variables were tested with either two-tailed t-test or Mann-Whitney test . Correlations between serum PTH and calcium levels were tested with linear regression analysis .
|
Inorganic calcium is a critical element for a diverse range of cellular processes ranging from cell signaling to energy metabolism , and its extracellular concentration is controlled by parathyroid hormone ( PTH ) . Klotho is expressed in parathyroid chief cells and reported to facilitate PTH secretion during hypocalcemia and mediate FGF23 suppression of PTH synthesis and secretion . To dissect the role of parathyroid Klotho in health and disease , we generated parathyroid-specific Klotho knockout mice . The mutant mice had normal serum levels of PTH and calcium . Further , their parathyroid sensitivity to acute fluctuations in serum calcium and response to FGF23 treatment were preserved , and mutant mice developed secondary hyperparathyroidism of similar magnitude as wild-type mice when challenged with renal failure . A previously unknown parathyroid FGF23 signaling pathway involving calcineurin was constitutively activated in the mutant mice , and blocking this pathway abolished FGF23-induced suppression of PTH secretion . Our data challenges the concepts of Klotho as a mandatory factor for the acute hypocalcemic PTH response and decreased Klotho abundance as a pathogenic factor in secondary hyperparathyroidism . Finally , the presence of Klotho-independent FGF23 effects in a Klotho-expressing target organ represents a paradigm shift in the conceptualization of FGF23 endocrine action .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2013
|
Parathyroid-Specific Deletion of Klotho Unravels a Novel Calcineurin-Dependent FGF23 Signaling Pathway That Regulates PTH Secretion
|
Osteoporotic fracture is a major cause of morbidity and mortality worldwide . Low bone mineral density ( BMD ) is a major predisposing factor to fracture and is known to be highly heritable . Site- , gender- , and age-specific genetic effects on BMD are thought to be significant , but have largely not been considered in the design of genome-wide association studies ( GWAS ) of BMD to date . We report here a GWAS using a novel study design focusing on women of a specific age ( postmenopausal women , age 55–85 years ) , with either extreme high or low hip BMD ( age- and gender-adjusted BMD z-scores of +1 . 5 to +4 . 0 , n = 1055 , or −4 . 0 to −1 . 5 , n = 900 ) , with replication in cohorts of women drawn from the general population ( n = 20 , 898 ) . The study replicates 21 of 26 known BMD–associated genes . Additionally , we report suggestive association of a further six new genetic associations in or around the genes CLCN7 , GALNT3 , IBSP , LTBP3 , RSPO3 , and SOX4 , with replication in two independent datasets . A novel mouse model with a loss-of-function mutation in GALNT3 is also reported , which has high bone mass , supporting the involvement of this gene in BMD determination . In addition to identifying further genes associated with BMD , this study confirms the efficiency of extreme-truncate selection designs for quantitative trait association studies .
Osteoporotic fracture is a leading cause of morbidity and mortality in the community , particularly amongst the elderly . In 2004 ten million Americans were estimated to have osteoporosis , resulting in 1 . 5 million fractures per annum [1] . Hip fracture is associated with a one year mortality rate of 36% in men and 21% in women [2]; and the burden of disease of osteoporotic fractures overall is similar to that of colorectal cancer and greater than that of hypertension and breast cancer [3] . Bone mineral density ( BMD ) is strongly correlated with bone strength and fracture risk , and its measurement is widely used as a diagnostic tool in the assessment of fracture risk [4]–[6] . BMD is known to be highly heritable , with heritability assessed in both young and elderly twins , and in families , to be 60–90% [7]–[14] . Although the extent of covariance between BMD and fracture risk is uncertain , of the 26 genes associated with BMD at genome-wide significant levels to date , nine have been associated with fracture risk ( reviewed in [15] ) , supporting the use of BMD as an intermediate phenotype in the search for genes associated with fracture risk . There is considerable evidence from genetic studies in humans [12] , [16] , [17] , and in mice [18] , indicating that the genes that influence BMD at different sites , and in the different genders , overlap but are not identical . Thus far all genome-wide association studies ( GWAS ) of BMD have studied cohorts of a wide age range , and with one exception have included both men and women; when only women have been studied , both pre- and postmenopausal women have been included . Therefore , to identify genes involved in osteoporosis in the demographic at highest risk of osteoporotic fracture we have performed a GWAS in postmenopausal women selected on the basis of their hip BMD , and replicated the GWAS findings in a large cohort of adult women drawn from the general population .
SNPs at chromosome 2q24 , in and around GALNT3 , achieved near genome-wide significance in our discovery cohort ( peak P-value rs1863196 , total hip ( TH ) P = 2 . 3×10−5; LS P = 0 . 037 ) ( Figure 1A ) . This SNP was not typed or imputed by either the Rotterdam or the TwinsUK cohorts , but a nearby SNP showed strong association in both AOGC and the combined replication cohorts ( rs6710518; AOGC discovery , TH P = 6 . 9×10−5; combined replication sets , FN P = 2 . 7×10−7 ) . In the combined datasets the finding achieved genome-wide significance at the FN ( P = 1 . 7×10−10 ) . Strong association was also seen with this SNP at LS ( P = 7 . 5×10−5 ) . Another marker within GALNT3 , rs4667492 , was also associated with fracture risk , including vertebral fractures ( OR = 0 . 89; 95%CI = 0 . 80–0 . 99; P = 0 . 032 ) and overall low trauma fractures ( OR = 0 . 92; 95%CI = 0 . 85–0 . 99; P = 0 . 024 ) . We have recently identified a mouse with an N-ethyl-N-nitrosourea induced loss-of-function GALNT3 mutation ( Trp589Arg ) , that develops hyperphosphataemia with extraskeletal calcium deposition , and hence represents a model for FTC [35] . To establish further the association of GALNT3 and BMD , we determined BMD in these GALNT3 mutant mice . This revealed that homozygous ( −/− ) GALNT3 mutant male and female adult mice had a higher areal BMD than their wild-type ( +/+ ) litter mates , with heterozygous ( +/− ) mice having intermediate BMD ( Figure 2 ) . This loss-of-function GALNT3 mutation is predicted to lead to a reduced glycosylation of FGF23 , which increases its breakdown and leads to reduced serum FGF23 concentrations [35] . A novel genome-wide significant association was also seen at markers on chromosome 6q22-23 ( Figure 1B ) . In the combined dataset , marker rs13204965 achieved genome-wide significance at this locus at the FN ( P = 2 . 2×10−9 ) , with strong support in both the AOGC discovery set , and the combined replication sets ( AOGC-discovery , TH P = 2 . 1×10−4; combined replication P = 3 . 5×10−5 ) . Strong association was also seen with LS BMD ( rs13204965 P = 0 . 00067 ) . The peak of association at this locus lies within a cDNA fragment , AK127472 . The nearest gene , RSPO3 ( R-spondin-3 ) , is 275 kb telomeric of the strongest associated SNP , but is within the associated linkage disequilibrium region ( Figure 1B ) . Association was observed at chromosome 16p13 with SNPs in and around CLCN7 , which encodes a Cl−/H+ antiporter expressed primarily in osteoclasts , and critical to lysosomal acidification , an essential process in bone resorption . Peak association at this locus was seen with SNP rs13336428 in the discovery set ( TH P = 7 . 0×10−4; LS P = 0 . 028 ) ( Figure S3A ) , which was confirmed in the replication set ( FN P = 3 . 6×10−5; LS P = 0 . 00012 ) , achieving P = 1 . 7×10−6 at the FN and 1 . 2×10−5 at LS in the overall cohort . Association has previously been reported between two SNPs in exon 15 of CLCN7 ( rs12926089 , rs12926669 ) and FN BMD ( P = 0 . 001–0 . 003 ) [36]; no association was seen with either of those SNPs in the current study ( P>0 . 4 at FN and LS ) . Association was observed with SNPs in IBSP ( integrin-binding bone sialoprotein ) ( Figure S3B ) , encoded at chromosome 4q22 , a gene which has previously had suggestive association reported with BMD in two studies ( rs1054627 , Styrkarrsdottir et al P = 4 . 6×10−5 [22]; Koller et al P = 1 . 5×10−4 [37] ) . In the current study , moderate association was observed in the discovery set with the same SNP as previously reported ( rs1054627 , AOGC discovery TH , P = 6 . 6×10−5 ) , with support in the replication set and strong association overall ( FN combined replication P = 9 . 2×10−5; FN overall association P = 7 . 6×10−7 ) . Nominal association was observed at LS ( rs1054627 , P = 0 . 019 ) . Association with BMD was also seen at chromosome 11p13 , with SNP rs1152620 achieving P = 4 . 4×10−5 ( TH ) in the discovery set , P = 0 . 0051 ( FN ) in the replication set , and P = 3 . 6×10−4 overall ( Figure S3C ) . This SNP was also nominally associated with LS BMD in the discovery set ( P = 0 . 041 ) . The nearest gene to this locus is LTBP3 ( latent transforming growth factor beta binding protein 3 ) , which is located 292 kb q-telomeric of rs1152620 . At chromosome 6p22 , SNPs in and around SOX4 ( Sex determining region Y box 4 ) were moderately associated with BMD in our discovery set ( most significant association rs9466056 , TH P = 5 . 3×10−4; LS P = 0 . 0036 ) ( Figure S3D ) , with support at the hip and LS in the replication set ( FN P = 0 . 00013 , LS P = 0 . 013 ) , achieving association overall with P = 2 . 6×10−7 ( FN ) and P = 0 . 00081 ( LS ) .
This study demonstrates convincing evidence of association with six genes with BMD variation , GALNT3 , RSPO3 , CLCN7 , IBSP , LTBP3 and SOX4 . Using a moderate sample size , the use of a novel study design also led to the confirmation of 21 of 26 known BMD-associations . This study thus demonstrates the power of extreme-truncate selection designs for association studies of quantitative traits . GALNT3 encodes N-acetylgalactosaminyltransferase 3 , an enzyme involved in 0-glycosylation of serine and threonine residues . Mutations of GALNT3 are known to cause familial tumoral calcinosis ( FTC , OMIM 2111900 ) [38] and hyperostosis-hyperphosphataemia syndrome ( HOHP , OMIM 610233 ) [39] . FTC is characterised by hyperphosphataemia in association with the deposition of calcium phosphate crystals in extraskeletal tissues; whereas in HOHP , hyperphosphataemia is associated with recurrent painful long bone swelling and radiographic evidence of periosteal reaction and cortical hyperostosis . FGF23 mutations associated with FTC cause hyperphosphataemia through effects on expression of the sodium-phosphate co-transporter in the kidney and small intestine , and through increased activation of vitamin D due to increased renal expression of CYP27B1 ( 25-hydroxyvitamin-D 1 alpha hydroxylase ) [40] . It is unclear whether FGF23 has direct effects on the skeleton or if its effects are mediated through its effects on serum phosphate and vitamin D levels . FGF23 signals via a complex of an FGF receptor ( FGFR1 ( IIIc ) ) and Klotho [41]; mice with a loss-of-function mutation in Klotho develop osteoporosis amongst other abnormalities , and modest evidence of association of Klotho with BMD has been reported in several studies [42] , [43] , [44] , [45] . We saw no association with polymorphisms in Klotho and BMD in the current study ( P>0 . 05 for all SNPs in and surrounding Klotho ) . To our knowledge , this finding is the first demonstration in humans that genetic variants in the FGF23 pathway are associated with any common human disease . RSPO3 is one of four members of the R-spondin family ( R-spondin-1 to −4 ) , which are known to activate the Wnt pathway , particularly through effects on LRP6 , itself previously reported to be BMD-associated [46] , [47] . LRP6 is inhibited by the proteins Kremen and DKK1 , which combine to induce endocytosis of LRP6 , reducing its cell surface levels . R-spondin family members have been shown to disrupt DKK1-dependent association of LRP6 and Kremen , thereby releasing LRP6 from this inhibitory pathway [48] . R-spondin-4 mutations cause anonychia ( absence or severe hypoplasia of all fingernails and toenails , OMIM 206800 ) [49] . No human disease has been associated with R-spondin-3 , and knockout of R-spondin-3 in mice is embryonically lethal due to defective placental development [50] . Mutations of CLCN7 cause a family of osteopetroses of differing age of presentation and severity , including infantile malignant CLCN7-related recessive osteopetrosis ( ARO ) , intermediate autosomal osteopetrosis ( IAO ) , and autosomal dominant osteopetrosis type II ( ADOII , Albers-Schoenberg disease ) . These conditions are characterized by expanded , dense bones , with markedly reduced bone resorption . Our data support associations of polymorphisms at this locus with BMD variation in the population . IBSP is a major non-collagenous bone matrix protein involved in calcium and hydroxyapatite binding , and is thought to play a role in cell-matrix interactions through RGD motifs in its amino acid sequence . IBSP is expressed in all major bone cells including osteoblasts , osteocytes and osteoclasts; and its expression is upregulated in osteoporotic bone [51] . IBSP knockout mice have low cortical but high trabecular bone volume , with impaired bone formation , resorption , and mineralization [52] . IBSP lies within a cluster of genes including DMP1 , MEPE , and SPP1 , all of which have known roles in bone and are strong candidate genes for association with BMD . MEPE has previously been associated with BMD at genome-wide significance [17] . In the current study the strongest association was seen with an SNP in IBSP , rs1054627 , as was the case with two previous studies [22] , [37] . Linkage disequilibrium between this SNP , and the previously reported BMD-associated SNP rs1471403 in MEPE , is modest ( r2 = 0 . 16 ) . Whilst out study supports the association of common variants in IBSP in particular with BMD , further studies will be required to determine if more than one of these genes is BMD-associated . Recessive mutations of LTBP3 have been identified as the cause of dental agenesis in a consanguineous Pakistani family ( OMIM 613097 ) [53] . Affected family members had base of skull thickening , and elevated axial but not hip BMD . LTBP3−/− mice develop axial osteosclerosis with increased trabecular bone thickness , as well as craniosynostosis [54] . LTBP3 is known to bind TGFβ1 , -β2 and -β3 , and may influence chondrocyte maturation and enchondral ossification by effects on their bioavailability [54] . Our study also confirms the previously reported association of another TGF pathway gene , TGFBR3 , encoded at chromosome 1p22 , with BMD [33] ( Figure S3E ) . In that study , association was observed in four independent datasets , but overall the findings did not achieve genome-wide significance at any individual SNP ( most significant SNP rs17131547 , P = 1 . 5×10−6 ) . In our discovery set , peak association was seen at this locus with SNP rs7550034 ( TH P = 1 . 5×10−4 ) , which lies 154 kb q-telomeric of rs17131547 , but still within TGFBR3 ( rs17131547 was not typed or imputed in our dataset ) ( Figure S3E ) . This supports TGFBR3 as a true BMD-associated gene . This study also demonstrated that SOX4 polymorphisms are associated with BMD variation . Both SOX4 and SOX6 are cartilage-expressed transcription factors known to play essential roles in chondrocyte differentiation and cartilage formation , and hence endochondral bone formation . SOX6 has previously been reported to be BMD-associated at genome-wide significant levels [17] . Whilst SOX4−/− mice develop severe cardiac abnormalities and are non-viable , SOX4+/− mice have osteopaenia with reduced bone formation but normal resorption rates , and diminished cortical and trabecular bone volume [55] . Our data suggest that SOX4 polymorphisms contribute to the variation in BMD in humans . This study has a unique design amongst GWAS of BMD reported to date , using an extreme-truncate ascertainment scheme , focusing on a specific skeletal site ( TH ) , and with recruitment of a narrow age- and gender-group ( post-menopausal women age 55–85 years ) . Our goal in employing this scheme was to increase the study power by reducing heterogeneity due to age- , gender- and skeletal site-specific effects . Whilst osteoporotic fracture can occur at a wide range of skeletal sites , hip fracture in postmenopausal women is the major cause of morbidity and mortality due to osteoporosis . To date , with only one exception , all GWAS of BMD have studied cohorts unselected for BMD [28] , and no study has restricted its participants to postmenopausal women ascertained purely on the basis of hip BMD . Assuming marker-disease-associated allele linkage disequilibrium of r2 = 0 . 9 , for alpha = 5×10−8 our study has 80% power to detect variants contributing 0 . 3% of the additive genetic variance of BMD . An equivalent-powered cohort study would require ∼16 , 000 unselected cases . Considering the 26 known genes ( or genomic areas ) associated with BMD , P-values less than <0 . 05 were seen in our discovery for 21 of the BMD-associated SNPs . Of the 26 known BMD genes , 16 would have been included in our replication study on the basis of the strength of their BMD association in our discovery cohort , but were not further genotyped as they were known already to be BMD-associated . Had these 16 genes replicated , 22 genes would have been identified in this single study , demonstrating the power of the design of the current study . A potential criticism of studies of highly selected cohorts , such as the AOGC-discovery cohort , is that the associations identified may not be relevant in the general population . However , the confirmation of our findings in replication cohorts of women unselected for BMD confirms that our findings are of broad relevance . In summary , our study design therefore represents a highly efficient model for future studies of quantitative traits and is one of the first reported studies using an extreme truncate design in any disease . We have identified two new BMD loci at genome-wide significance ( GALNT3 , RSPO3 ) , with GALNT3 SNPs also associated with fracture . Strong evidence was also demonstrated for four novel loci ( CLCN7 , IBSP , LTBP3 , SOX4 ) . Further support was also provided that TGFBR3 is a true BMD-associated locus . Our discovery cohort replicated 21 of 26 previously identified BMD-associated loci . Our novel findings further advance our understanding of the aetiopathogenesis of osteoporosis , and highlight new genes and pathways not previously considered important in BMD variation and fracture risk in the general population . Our study also provides strong support that the use of extreme truncate selection is an efficient and powerful approach for the study of quantitative traits .
All participants gave written , informed consent , and the study was approved by the relevant research ethics authorities at each participating centre . The discovery sample population included 1128 Australian , 74 New Zealand and 753 British women , between 55–85 years of age , five or more years postmenopausal , with either high BMD ( age- and gender-adjusted BMD z-scores of +1 . 5 to +4 . 0 , n = 1055 ) or low BMD ( age- and gender-adjusted BMD z-scores of −4 . 0 to −1 . 5 , n = 900 ) ( Tables S1 and S2 ) . BMD z-scores were determined according to the Geelong Osteoporosis Study normative range [19] . Low BMD cases were excluded if they had secondary causes of osteoporosis , including corticosteroid usage at doses equivalent to prednisolone ≥7 . 5 mg/day for ≥6 months , past or current anticonvulsant usage , previous strontium usage , premature menopause ( <45 years ) , alcohol excess ( >28 units/week ) , chronic renal or liver disease , Cushing's syndrome , hyperparathyroidism , thyrotoxicosis , anorexia nervosa , malabsorption , coeliac disease , rheumatoid arthritis , ankylosing spondylitis , inflammatory bowel disease , osteomalacia , and neoplasia ( cancer , other than skin cancer ) . Screening blood tests ( including creatinine ( adjusted for weight ) , alkaline phosphatase , gamma-glutamyl transferase , 25-hydroxyvitamin D and PTH ) were checked in 776 cases , and no differences were found between the high and low BMD groups . Therefore no further screening tests were done of the remaining cases . Fracture data were analysed comparing individuals who had never reported a fracture after the age of 50 years , with individuals who had had a low or non-high trauma ( low trauma fracture = fracture from a fall from standing height or less ) osteoporotic fracture ( excluding skull , nose , digits , hand , foot , ankle , patella ) after the age of 50 years . Vertebral , hip and non-vertebral fractures were considered both independently and combined . All participants were of self-reported white European ancestry . DNA was obtained from peripheral venous blood from all cases except those recruited from New Zealand , for whom DNA was obtained from salivary samples using Oragene kits ( DNA Genotek , Ontario , Canada ) . We have previously demonstrated that DNA from these two sources have equivalent genotyping characteristics [20] . After quality control checks including assessment of cryptic relatedness , ethnicity and genotyping quality , 900 individuals with low TH BMD and 1055 individuals with high TH BMD were available for analysis . The replication cohort consisted of 8928 samples drawn from nine cohort studies , outlined in Tables S3 and S4 ( ‘AOGC replication cohort’ ) which were directly genotyped , These replication cases were adult women ( age 20–95 years ) , unselected with regard to BMD , and who were not screened for secondary causes of osteoporosis . Replication was also performed in silico in 11 , 970 adult women from the TwinsUK and Rotterdam , and deCODE Genetics GWASs [21] , [22] , [23] , in which association data were available at LS and FN . High and low BMD ascertainment was defined according to the TH score , because this has better measurement precision than FN BMD [24] . However , neither TwinsUK nor the Rotterdam Study had TH BMD on the majority of their datasets and therefore were analysed using the FN measurement for which data were available on the whole cohort . All replication findings at the hip are reported therefore for FN BMD . TH and FN BMD are closely correlated ( r = 0 . 882 in the AOGC dataset ) , with FN BMD one of the components of the TH BMD measurement . Genotyping of the discovery cohort ( n = 2036 ) was performed using Illumina Infinium II HumHap300 ( n = 140 ) , 370CNVDuo ( n = 4 ) , 370CNVQuad ( n = 1882 ) and 610Quad ( n = 10 ) chips at the University of Queensland Diamantina Institute , Brisbane , Australia . Genotype clustering was performed using Illumina's BeadStudio software; all SNPs with quality scores <0 . 15 and all individuals with <98% genotyping success were excluded . 289499 SNPs were shared across all chip types . Cluster plots from the 500 most strongly associated loci , were manually inspected and poorly clustering SNPs excluded from analysis . Following imputation using the HapMap Phase 2 data , 2 , 543 , 887 SNPs were tested for association with TH and LS BMD ( Manhattan plot of association findings , Figure S1 ) . After data cleaning , minimal evidence of inflation of test statistics was observed , with a genomic inflation factor ( λ ) of 1 . 0282 ( qq plot , Figure S2 ) . A total of 124 SNPs were successfully genotyped in the AOGC replication cohort . These replication study SNPs were selected from the findings of the discovery cohort , either based on the strength of association ( P-value ) or following analysis with GRAIL ( n = 45 ) [25] , using as seed data all SNPs previously reported to be associated with BMD at GWAS significant levels ( results for all replication SNPs presented in Table S5 ) . GRAIL is a bioinformatic program that assesses the strength of relationships between genes in regions surrounding input SNPs ( usually derived from genetic association studies ) and other SNPs or genes associated with the trait of interest , by assessing their co-occurrence in PubMed abstracts . Where genes surrounding input SNPs occur more frequently in abstracts with known associated genes , these SNPs are more likely themselves also to be associated , and can thus be prioritized for inclusion in replication studies . For the replication study , genotyping was performed either by Applied Biosystems OpenArray ( n = 113 ) or Taqman technology ( n = 11 ) ( Applied Biosystems , Foster City , CA , USA ) , according to the manufacturer's protocol . Eleven individuals were removed because of abnormal X-chromosome homozygosity ( X-chromosome homozygosity either <−0 . 14 , or >+0 . 14 ) . Outliers with regard to autosomal heterozygosity ( either <0 . 34225 or >0 . 357 , n = 40 ) and missingness ( >3% , n = 4 ) were removed . Using an IBS/IBD analysis in PLINK to detect cryptic relatedness , one individual from 35 pairs of individuals with pi-hat >0 . 12 ( equivalent to being 3rd degree relatives or closer ) were removed . SNPs with minor allele frequency <1% ( n = 561 ) , and those not in Hardy-Weinberg equilibrium ( P<10−7 , n = 170 ) were then removed , leaving 288 , 768 SNPs in total . Nine replication SNPs were removed because of excess missingness ( >10% ) or because they failed tests of Hardy-Weinberg equilibrium ( P<0 . 001 ) . To detect and correct for population stratification EIGENSTRAT software was used . We first excluded the 24 regions of long range LD including the MHC identified in Price et al . before running the principal components analysis , as suggested by the authors [26] . Sixteen individuals were removed as ethnic outliers , leaving 1955 individuals in the final discovery dataset . Imputation analyses were carried out using Markov Chain Haplotyping software ( MaCH; http://www . sph . umich . edu/csg/abecasis/MACH/ ) using phased data from CEU individuals from release 22 of the HapMap project as the reference set of haplotypes . We only analyzed SNPs surrounding disease-associated SNPs that were either genotyped or could be imputed with relatively high confidence ( R2≥0 . 3 ) . For TH measurements , a case-control association analysis of imputed SNPs was performed assuming an underlying additive model and including four EIGENSTRAT eigenvectors as covariates , using the software package MACH2DAT [27] which accounts for uncertainty in prediction of the imputed data by weighting genotypes by their posterior probabilities . For FN and LS BMD analyses , Z-transformed residual BMD scores ( in g/cm2 ) were generated for the entire AOGC cohort after adjusting for the covariates age , age2 , and weight , and for centre of BMD measurement . Because the regression coefficient for BMD on genotype would be biased by selection for extremes , we adopted the approach detailed in Kung et al ( 2009 ) [28] . Specifically , the regression coefficient of genotype on BMD was estimated , and subsequently transformed to the regression coefficient of BMD on genotype through knowledge of the population variance of the phenotype and the allele frequencies . For fracture data , analysis was by logistic regression . Only SNPs achieving GWAS significance were tested for fracture association . The SNPs used for replication from the Rotterdam Study were analyzed using MACH2QTL implemented in GRIMP [29] . Data from the discovery and replication cohorts were combined using the inverse variance approach as implemented in the program METAL [30] . SNPs associated with BMD were also tested for association with fracture in the AOGC discovery and replication cohorts ( hip , vertebral , nonvertebral , and all low trauma fractures , age ≥50 years , as defined above ) , by logistic regression . Study power was calculated using the ‘Genetic Power Calculator’ [31] . All animal studies were approved by the MRC Harwell Unit Ethical Review Committee and are licensed under the Animal ( Scientific Procedures ) Act 1986 , issued by the UK Government Home Office Department . Dual-energy X-ray absorptiometry ( DEXA ) was performed using a Lunar Piximus densitometer ( GE Medical Systems ) and analysed using the Piximus software . Data related to this study will be available to research projects approved by a Data Access Committee including representatives of the University of Queensland Research Ethics Committee . For enquiries regarding access please contact the corresponding author , MAB ( matt . brown@uq . edu . au ) .
|
Osteoporotic fracture is a major cause of early mortality and morbidity in the community . To identify genes associated with osteoporosis , we have performed a genome-wide association study . In order to improve study power and to address the demographic group of highest risk from osteoporotic fracture , we have used a unique study design , studying 1 , 955 postmenopausal women with either extreme high or low hip bone mineral density . We then confirmed our findings in 20 , 898 women from the general population . Our study replicated 21 of 26 known osteoporosis genes , and it identified a further six novel loci ( in or nearby CLCN7 , GALNT3 , IBSP , LTBP3 , RSPO3 , and SOX4 ) . For one of these loci , GALTN3 , we demonstrate in a mouse model that a loss-of-function genetic mutation in GALNT3 causes high bone mass . These findings report novel mechanisms by which osteoporosis can arise , and they significantly add to our understanding of the aetiology of the disease .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genetics",
"and",
"genomics/genetics",
"of",
"disease",
"diabetes",
"and",
"endocrinology/bone",
"and",
"mineral",
"metabolism",
"rheumatology/bone",
"and",
"mineral",
"metabolism"
] |
2011
|
Genome-Wide Association Study Using Extreme Truncate Selection
Identifies Novel Genes Affecting Bone Mineral Density and Fracture
Risk
|
Genome-wide association studies ( GWAS ) have recently identified KIF1B as susceptibility locus for hepatitis B virus ( HBV ) –related hepatocellular carcinoma ( HCC ) . To further identify novel susceptibility loci associated with HBV–related HCC and replicate the previously reported association , we performed a large three-stage GWAS in the Han Chinese population . 523 , 663 autosomal SNPs in 1 , 538 HBV–positive HCC patients and 1 , 465 chronic HBV carriers were genotyped for the discovery stage . Top candidate SNPs were genotyped in the initial validation samples of 2 , 112 HBV–positive HCC cases and 2 , 208 HBV carriers and then in the second validation samples of 1 , 021 cases and 1 , 491 HBV carriers . We discovered two novel associations at rs9272105 ( HLA-DQA1/DRB1 ) on 6p21 . 32 ( OR = 1 . 30 , P = 1 . 13×10−19 ) and rs455804 ( GRIK1 ) on 21q21 . 3 ( OR = 0 . 84 , P = 1 . 86×10−8 ) , which were further replicated in the fourth independent sample of 1 , 298 cases and 1 , 026 controls ( rs9272105: OR = 1 . 25 , P = 1 . 71×10−4; rs455804: OR = 0 . 84 , P = 6 . 92×10−3 ) . We also revealed the associations of HLA-DRB1*0405 and 0901*0602 , which could partially account for the association at rs9272105 . The association at rs455804 implicates GRIK1 as a novel susceptibility gene for HBV–related HCC , suggesting the involvement of glutamate signaling in the development of HBV–related HCC .
Hepatocellular carcinoma ( HCC ) is the sixth common cancer and the third common cause of cancer mortality worldwide [1] . The incidence rate of HCC varies considerably in the world , with the highest in East , Southeast Asia and Sub-Saharan Africa , and China alone accounts for approximately half of HCC malignancies [1] , [2] . Major risk factors for HCC are chronic infections with the hepatitis B or C viruses , and exposure to dietary aflatoxin B1 . Hepatitis B virus ( HBV ) infection is particular important , because of its coherent distribution with the HCC prevalence [1] , [2] . However , it is known that only a minority of chronic carriers of HBV develop HCC [3] , and the chronic HBV carriers with a family history of HCC have a two-fold risk for HCC than those without the family history [4] , strongly suggesting the importance of genetic susceptibility for HBV-related HCC . A number of candidate genes were investigated by genetic association studies to evaluate their roles in the susceptibility to HCC [5] . However , the findings from these studies are inconclusive due to moderate evidence and lack of independent validation . Recently , a genome-wide association study ( GWAS ) of HBV-related HCC was performed [6] , in which 355 HBV–positive HCC patients and 360 chronic HBV carriers were used for the genome-wide discovery analysis , and the top 45 SNPs from the discovery analysis were further evaluated in additional 1 , 962 HBV–positive HCC patients and 1 , 430 controls ( both chronic HBV carriers and population controls ) as well as 159 trios . The study identified KIF1B as a novel susceptibility locus ( top SNP rs17401966 ) on 1p36 . 22 . Further study with better design and bigger sample size was recommended for identifying additional susceptibility loci for HCC [7] , [8] . These motivate us to carry out a GWAS with a large sample size in Chinese population to discover novel susceptibility loci for HCC .
We performed a genome-wide discovery analysis by analyzing 523 , 663 common autosomal SNPs in two independent cohorts of the Han Chinese: 480 cases and 484 controls from central China and 1058 cases and 981 controls from southern China ( Table S1 and Figure S1 ) . The principal component analysis ( PCA ) confirmed all the samples to be Chinese , but indicated moderate genetic mismatch between the cases and controls in the cohort of southern China ( Figure S2 ) . To minimize the effect of population stratification , we performed the genome-wide association analysis using PCA-based correction for population stratification . After the adjustment by the first principal component , the λgc of the genome-wide association results is 1 . 013 for the cohort of central China , 1 . 003 for the cohort of southern China and 1 . 012 for the combined samples . Furthermore , for all the three genome-wide analyses of central , southern and combined samples , the quantile-quantile ( QQ ) plot of the observed P values revealed a good overall fit with the null distribution ( Figure S3 ) . Taken together , these results clearly indicate that the final association results from our genome-wide discovery analysis are free of inflation effect due to population stratification . The genome-wide discovery analysis revealed multiple suggestive associations ( P<10−5 ) on 2q22 . 1 , 6p21 . 32 , 11p15 . 1 and 20q12 ( Figure S4 and Table S2 ) . To validate these findings , 39 SNPs were selected according to their overall association evidence in three GWAS analyses as well as their consistencies of association between the two independent GWAS samples ( Central and Southern China ) ( see the Methods for the selection criteria ) . The 39 SNPs were genotyped in additional 2 , 112 HBV–positive HCC cases and 2 , 208 HBV carriers ( Phase I validation ) ( Table S1 ) . Of the 39 SNPs , only 3 ( rs9272105 on 6p21 . 32 , rs11148740 on 13q21 . 32 and rs455804 on 21q21 . 3 ) were validated , showing consistent association between the GWAS discovery and Phase I validation samples ( Table S3 ) . These 3 SNPs were then genotyped in additional 1 , 021 HBV–positive HCC cases and 1 , 491 HBV carriers ( Phase II validation ) . The Phase II validation analysis ( Table 1 ) confirmed the associations at rs9272105 on 6p21 . 32 ( OR = 1 . 41 , P = 7 . 63×10−9 ) and rs455804 on 21q21 . 3 ( OR = 0 . 83 , P = 3 . 63×10−3 ) , but not the association at rs11148740 on 13q21 . 32 ( Table S3 ) . For both rs9272105 and rs455804 , no heterogeneity of associations were observed among the GWAS and validation samples ( P>0 . 05 ) , and the associations in the combined GWAS and validation samples achieved genome-wide significance ( P<5 . 0×10−8 ) ( rs9272105: OR = 1 . 30 , P = 1 . 13×10−19 and rs455804: OR = 0 . 84 , P = 1 . 86×10−8 ) ( Table 1 ) . As a replication , these two SNPs were genotyped in the fourth independent samples of 1 , 298 cases and 1 , 026 controls from central China , which further confirmed the associations at rs9272105 ( OR = 1 . 25 , P = 1 . 71×10−4 ) and rs11148740 ( OR = 0 . 84 , P = 6 . 92×10−3 ) ( Table 1 ) . When combining all the five groups of samples , the two SNPs resulted in a 28% increased , and a 16% decreased risk for HCC development ( rs9272105: OR = 1 . 28 , P = 5 . 24×10−22 and rs455804: OR = 0 . 84 , P = 5 . 24×10−10 ) ( Table 1 ) , respectively . The associations at the two SNPs remained genome-wide significant after adjusting for age , gender , smoking and drinking ( Table S4A ) . Furthermore , stratification analysis by age , gender , smoking and drinking status revealed similar ORs for rs9272105 and rs455804 among subgroups , except that the association at rs9272105 showed a stronger effect in the non-smoking group than the smoking one ( OR = 1 . 38 vs . 1 . 19 , P for heterogeneity = 0 . 004 ) ( Table S4B ) . Pair-wise interaction analysis among these two SNPs , smoking and drinking status did not reveal any significant interaction ( data not shown ) . The samples used in the GWAS , validation and replication analyses are summarized in Table S1 , and the multi-stage design of the whole study is shown in Figure S5 . We further investigated the association of HLA alleles in our GWAS samples through imputation . After QC filtering ( see the Methods ) , 37 HLA alleles were successfully imputed , and 5 alleles showed nominal association ( P<0 . 05 ) ( Table S5 and Table 2 ) . Further stepwise conditional analysis revealed that only two DRB1 alleles showed independent associations ( DRB1*0405: OR = 0 . 69 , P = 6 . 18×10−4; DRB1*0901: OR = 0 . 82 , P = 3 . 62×10−3 ) ( Table 2 ) . Conditioning on rs9272105 could abolish the associations of the DRB1 alleles , and conditioning on the DRB1 alleles could weaken , but not eliminate , the association at rs9272105 ( Table 2 ) . The haplotype analysis of rs9272105 and the two DRB1 alleles revealed consistent result , showing that both the DRB1 alleles sit on the haplotypes carrying the protective G allele of rs9272105 ( Table S6 ) . Taking together , there seems to be additional risk effect beyond the ones carried by the DRB1 alleles . We further explored whether the SNPs rs9272105 and rs455804 play any role in HBV infection . First , we compared the frequencies of these 2 SNPs between 408 non-symptomatic HBV carriers and 521 symptomatic chronic HBV patients from southern China ( GWA scanned ) . The analysis revealed a protective effect at rs9272105 ( OR = 0 . 80 , P = 1 . 67×10−2 ) on the development of symptomatic chronic hepatitis B , but no association at rs455804 ( Table S7A ) . Furthermore , we genotyped these 2 SNPs in 1 , 344 individuals with HBV nature clearance and compared their frequencies with those in 4 , 183 asymptomatic HBV carriers ( all from the Central China ) . The analysis also revealed a protective association at rs9272105 for HBV chronic infection ( OR = 0 . 88 , P = 3 . 78×10−3 ) ( Table S7B ) .
SNP rs9272105 is located between HLA-DQA1 and HLA-DRB1 on 6p21 . 32 ( Figure 1A ) . SNP imputation in the GWAS discovery samples revealed additional SNPs showing association , but rs9272105 remained to be the top SNP within the region ( Figure 1A ) . The residual association at rs9272105 after conditioning the association effects of the HLA alleles DRB1*0405 and *0901 suggests that there may be additional risk effect beyond the DRB1 alleles in Chinese population . The associations of the DRB1 alleles revealed by this study are consistent with the previous reports that HLA-DQ/DR alleles associated with HCC risk [9] , [10] . In addition , we investigated the previously reported HBV infection-associated SNPs rs3077 , rs9277535 , rs7453920 , and rs2856718 within the HLA DP/DQ region [11] , [12] with HCC development in our GWAS samples . By imputation , we found the evidence of the association at rs9277535 with HCC ( rs9277535: OR = 0 . 85 , P = 7 . 9×10−3 ) . However , there is no linkage disequilibrium ( LD ) between rs9277535 and our SNP rs9272105 ( r2 = 0 . 016 according the HapMap CHB+JPT samples ) , suggesting that the associations at rs9277535 and rs9272105 may be independent . The HLA-DQ locus has also been shown to be associated with HCV-related HCC in a Japanese GWAS ( rs9275572 , OR = 1 . 30 , P = 9 . 38×10−9 ) [13] . SNPs rs9275572 and rs9272105 are 79 kb away from each other and in weak LD ( D′ = 0 . 43 , r2 = 0 . 08 in the HapMap CHB samples ) . The SNP rs9275572 did not show any association with HBV-related HCC in our GWAS discovery samples ( OR = 0 . 93 , P = 0 . 24 ) ( Table S8 and Figure S6B ) . In addition to HLA-DQ , MICA ( rs2596542 ) on 6p21 . 33 and DEPDC5 ( rs1012068 ) on 22q12 . 3 were also identified as independent susceptibility loci for HCV-related HCC in Japanese population [13] , [14] . But , our GWAS discovery analysis did not reveal any supportive evidence for these two loci ( rs2596542: OR = 1 . 06 , P = 0 . 36; and rs1012068: OR = 1 . 06 , P = 0 . 37 ) ( Table S8 and Figure S6C and S6D ) . We also evaluated the power of our GWAS discovery samples and found that our samples should have sufficient power for detecting the previously reported associations at rs9275572 ( power = 94% ) , rs2596542 ( power = 92% ) and rs1012068 ( power = 94% ) . Taken together , the disparity of associations may suggest the different genetic background of the susceptibilities for HCV- and HBV-related HCC . Further studies will be required to confirm the genetic heterogeneity of HCV- and HBV-related HCC . The association of rs9272105 ( HLA-DQA1/DRB1 ) with HBV infection is consistent with the extensive reports on the association of HLA-DRB1 with HBV infection where both protective and risk DRB1 alleles for HBV infection and outcome were identified [11] , [12] , [15]–[19] . Intriguingly , our study has revealed that the variant allele of rs9272105 showed a protective effect for HBV infection ( OR = 0 . 88 ) and the progression to chronic symptomatic hepatitis B , but a risk effect for the development of HCC ( OR = 1 . 30 ) . Further studies will be needed to demonstrate whether the opposite associations of HBV infection and HBV-related HCC progression at rs9272105 are due to different causal variants within the HLA class II region . SNP rs455804 is located within the first intron of GRIK1 that is the only gene within the LD region of the association ( Figure 1B ) , strongly implicating GRIK1 as a novel susceptibility gene for HBV-related HCC . SNP imputation of the region did not reveal any SNPs that showed stronger association than rs455804 . GRIK1 encodes CLUR5 , which is involved in the glutamate signaling , as one of the ionotropic glutamate receptor , kainite 1 protein ( GLUR5 ) , a subunit of ligand-activated channels and involved in glutamate signaling . Our discovery of the association of GRIK1 with HCC has enhanced the emerging evidences for the important role of glutamate signaling pathway in cancer development . Glutamate has been shown to play a central role in the malignant phenotype of gliomas through multiple molecular mechanisms [20] . Inhibition of glutamate release and/or glutamate receptor activity can inhibit the proliferation and/or invasion of tumor cells in breast cancer [21] , laryngeal cancer [22] , and pancreatic cancer [23] , and ionotrpic glutamate receptor ( GLUR6 ) was also suggested to play a tumor-suppressor role in gastric cancer [24] . Recently , the exome sequencing analysis revealed that GRIN2A ( encoding the ionotrpic glutamate receptor ( N-methyl D-aspartate ) subunit 2A ) was mutated in 33% of melanoma tumors , clearly indicating the involvement of glutamate signaling in melanoma development . Finally , SNPs within GRIK1 have also been found significantly associated with paclitaxel response in NCI60 cancer cell lines , and may play a role in the cellular response to paclitaxel treatment in cancer [25] . Consistent with the previous observations , our discovery of GRIK1 as a HBV-related HCC susceptibility gene has suggested the importance of glutamate signaling in HBV-related HCC development , and , although still speculative , has highlighted the glutamate signaling pathway as a potentially novel target for the treatment of HCC . We also assessed the previously reported susceptibility locus KIF1B on 1p36 . 22 ( rs17401966 ) for HBV-related HCC [6] . Our GWAS discovery analysis did reveal the consistent result for the association at rs17401966 , but the strength of association in our GWAS discovery sample ( OR = 0 . 90 ) is much weaker than the previously reported one ( OR = 0 . 61 ) ( Table S8 ) . SNP imputation in our GWAS discovery samples did not reveal any stronger association than the association at rs17401966 within the LD region surrounding the 1p36 . 22 locus ( Figure S6A ) . Previous studies have clearly shown the existence of subpopulation structure of Chinese Han population along the north-south axis , and further demonstrated that geographic matching can be used as a good surrogate for genetic matching , and PCA-based correction is very effective in controlling the inflation effect of population stratification [26] . In the current study , all the cases and controls were matched by their geographic origin of residence . Moreover , the GWAS discovery samples were from central and southern China , while all the validation and replication samples were from central China . Our PCA analysis indicates that while there was mild population stratification in the sample of southern China , the cases and controls from central China were well matched without any indication of population stratification . In our study , the PCA-based correction was used in the GWAS analysis , and all the validation and replication analyses were from central China . Therefore , our findings should be free of adverse effect of population stratification in Chinese population . In conclusion , the current GWAS identified two biologically plausible , novel loci on 6p21 . 32 and 21q21 for HBV-related HCC . These findings highlight the importance of HLA-DQ/DR molecules and glutamate signaling in the development of HBV-related HCC .
The genome-wide discovery analysis was performed by genotyping 731 , 442 SNPs in 1 , 575 HBV positive HCC patients and 1 , 490 HBV positive controls derived from two independent case-control cohorts of 500 cases and 500 controls from Central China ( Shanghai ) and 1 , 075 cases and 990 controls from Southern China ( Guangdong ) . The first stage validation samples included 2 , 112 HBV–positive cases and 2 , 208 HBV–positive controls recruited from Jiangsu . The second stage validation samples consisted of 1 , 021 HBV–positive cases and 1 , 491HBV carriers recruited from Shanghai . The replication samples of 1 , 298 HBV–positive cases and 1 , 026 HBV carriers were recruited from Central China ( Shanghai and Jiangsu ) . ( Table S1 and Figure 1 ) All the samples are Han Chinese and partially participated in the previously published studies [27] , [28] . The diagnosis of HCC was confirmed by a pathological examination and/or α-fetoprotein elevation ( >400 ng/ml ) combined with imaging examination ( Magnetic resonance imaging , MRI and/or computerized tomography , CT ) . Because HCV infection is rare in Chinese , we excluded HCC with HCV infection . Cancer-free HBV+ control subjects from central China were recruited from those receiving routine physical examinations in local hospitals or those participating in the community-based screening for the HBV/HCV markers and frequency-matched for age , gender , and geographic regions to each set of the HCC patients . Almost all these community-based controls are asymptomatic HBV carriers . Similarly , cancer-free control subjects from southern China are all HBV+ , and 408 of them were asymptomatic HBV carriers and 521 were symptomatic chronic hepatitis B patients . All the HBV+ controls were positive for both HBsAg and antibody to hepatitis B core antigen ( anti-HBc ) , and negative for anti-HCV . We also recruited a HBV natural clearance cohort form Jiangsu Province ( Zhangjiagang and Changzhou cities ) through a population based screening for the HBV/HCV markers in 2004 and 2009 , respectively ( 58 , 142 persons ) . Subjects with HBV natural clearance were negative for HBsAg and anti-HCV , positive for both antibody to hepatitis B surface antigen ( anti-HBs ) and anti-HBc . About 9 , 610 subjects with HBV natural clearance were identified . No history of hepatitis B vaccination was reported for these people . Then , we randomly selected 1 , 344 HBV natural clearance people without self-reported history of cancer in the current study . The age for the 1 , 344 people were 52 . 6±10 . 2 years , and 217 ( 16 . 2% ) were females . We collected smoking and drinking information through interviews . Those who had smoked an average of less than 1 cigarette per day and less than 1 year in their lifetime were defined as nonsmokers; otherwise , they were considered as smokers . Individuals were classified as alcohol drinkers if they drank at least twice a week and continuously for one year during their lifetime; otherwise , they were defined as nondrinkers . At recruitment , the informed consent was obtained from each subject , and this study was approved by the Institutional Review Boards of each participating institution . We performed standard quality control on the raw genotyping data to filter both unqualified samples and SNPs . The samples with overall genotype completion rates <95% were excluded from further analysis ( 26 subjects ) . Eight subjects were excluded as they showed discrepancy between the recorded and genetically inferred genders . An additional 21 duplicates or probable familial relatives were excluded based on the IBD analysis implemented in PLINK ( all PI_HAT>0 . 25 ) . SNPs were excluded when they fit the following criteria: ( i ) not mapped on autosomal chromosomes; ( ii ) had a call rate <95% in all GWA samples or in either of Central cohort study or Southern study samples; ( iii ) had minor allele frequency ( MAF ) <0 . 05 in either of Central cohort study or Southern study samples; and ( iv ) genotype distributions deviated from those expected by Hardy-Weinberg equilibrium ( P<1×10−5 in either of Central cohort study or Southern study samples ) . We detected population outliers and stratification using a principal component analysis ( PCA ) based method . After removing MHC SNPs on chromosome 6 from 25–37 Mb , PCA was performed by using common autosomal SNPs with low LD ( r2<0 . 2 ) in the reference samples of the HapMap project ( YRI ( n = 90 ) , CEU ( n = 90 ) , CHB ( n = 45 ) and JPT ( n = 44 ) ) as the internal controls and our 3 , 010 participants of the GWAS discovery samples ( after removal of samples with low call rates , ambiguous gender , and familial relationships ) . Projection onto the two multidimensional scaling axes is shown in Figure S2A . 7 outliers ( more than 6 standard deviations ) were identified and excluded . Finally , 523 , 663 autosomal SNPs in 1 , 538 cases and 1 , 465 controls , consisting of 480 cases and 484 controls from Central China and 1 , 058 cases and 981 controls from Southern China , were retained for association testing ( Table S1 ) . SNPs for the first stage validation were selected based on the following criteria: ( i ) SNP had P joint≤1 . 0×10−4 in the analysis of the combined GWA samples or either the Central China sample or the Southern China sample , and had a consistent association in the two participant studies , meaning that the ORs from the two samples are both either above or below 1; ( ii ) only SNP with the lowest P value was selected when multiple SNPs showed a strong LD ( r2≥0 . 8 ) . As a result , a total of 39 SNPs were included in the first stage validation . 3 SNPs that were significantly associated with HCC risk in the first validation stage were further genotyped in the second stage validation samples . Genotyping in the two validation samples were done by using the iPLEX platform ( Sequenom ) or the TaqMan assays ( Applied Biosystems ) . The primers and probes were available upon request ( Table S9 ) . Laboratory technicians who performed genotyping experiments were blinded to case/control status . For TaqMan assay , ten percent of random samples were repeated , and the reproducibility was 100% . The 2 validated SNPs were genotyped in another independent replication using the same method . Population structure was evaluated by the PCA in the software package EIGENSTRAT 3 . 0 [26] . PCA revealed one significant ( P<0 . 05 ) eigenvector which was included in the logistic regression with other covariates of age , gender , smoking and drinking status for both the genome-wide discovery analysis and the joint analysis of the combined discovery and replication samples . Ancestral origin checking by PCA confirmed all the samples to be Han Chinese and further demonstrated moderate genetic stratification between the cases and the controls of the Southern cohort ( Figure S2 ) . The genome-wide association analysis was therefore performed in logistic regression using PCA-based correction for population stratification and by treating the samples of two cohorts as independent studies . The genomic-control inflation factor ( λgc ) after adjustment by the first PC was calculated for the Central cohort samples ( λgc = 1 . 013 ) , the Southern cohort samples ( λgc = 1 . 003 ) and the combined GWAS discovery samples ( λgc = 1 . 012 ) . Consistently , the QQ plot of the observed P values also showed a minimal inflation of genome-wide association results due to population stratification ( Figure S3 ) . Statistical analyses were performed by using PLINK 1 . 07 [29] and R 2 . 11 . 1 . The Manhattan plot of −log10P was generated using Haploview ( v4 . 1 ) [30] . Untyped genotypes were imputed in the GWAS discovery samples by using IMPUTE2 [31] and the haplotype information from the 1000 Genomes Project ( ASN samples as the reference set ) and HapMap3 ( CHB and JPT samples as the reference samples ) . The regional plot of association was created by using an online tool , LocusZoom 1 . 1 . P value was two-sided , and OR presented in the manuscript was estimated by using additive model and logistic regression analyses if not specified . To impute classical HLA alleles , we used 180 phased haplotypes from the HapMap CHB and JPT samples as our reference panel . This panel comprised dense SNP data and HLA allele types at 4-digit resolution for the HLA class I ( HLA-A , B , C ) and II ( DQA1 , DQB1 and DRB1 ) genes as previously described [32] . Genotypes , probability and allelic dosages were then imputed separately in the two discovery samples of Central and Southern Chinese using the BEAGLE program . Association testing was performed by using a logistic regression model on the best-guessed genotypes and allelic dosages . The results were checked for consistency between the two methods , and the results from best-guessed genotypes were presented .
|
Previous studies strongly suggest the importance of genetic susceptibility for hepatocellular carcinoma ( HCC ) . However , the studies about genetic etiology on HBV–related HCC were limited . Our genome-wide association study included 523 , 663 autosomal SNPs in 1 , 538 HBV–positive HCC patients and 1 , 465 chronic HBV carriers for the discovery analysis . 2 , 112 HBV–positive HCC cases and 2 , 208 HBV carriers ( the initial validation ) , and 1 , 021 cases and 1 , 491 HBV carriers ( the second validation ) , were then analyzed for validation . The fourth independent samples of 1 , 298 cases and 1 , 026 controls were analyzed as replication . We discovered two novel associations at rs9272105 ( HLA-DQA1/DRB1 ) on 6p21 . 32 and rs455804 ( GRIK1 ) on 21q21 . 3 . HLA-DRB1 molecules play an important role in chronic HBV infection and progression to HCC . The association at rs455804 implicates GRIK1 as a novel susceptibility gene for HBV–related HCC , suggesting the involvement of glutamate signaling in the development of HBV–related HCC .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"genome-wide",
"association",
"studies",
"cancer",
"genetics",
"genetics",
"biology",
"genetics",
"and",
"genomics"
] |
2012
|
GWAS Identifies Novel Susceptibility Loci on 6p21.32 and 21q21.3 for Hepatocellular Carcinoma in Chronic Hepatitis B Virus Carriers
|
Quantifying heterogeneity in gene expression among single cells can reveal information inaccessible to cell-population averaged measurements . However , the expression level of many genes in single cells fall below the detection limit of even the most sensitive technologies currently available . One proposed approach to overcome this challenge is to measure random pools of k cells ( e . g . , 10 ) to increase sensitivity , followed by computational “deconvolution” of cellular heterogeneity parameters ( CHPs ) , such as the biological variance of single-cell expression levels . Existing approaches infer CHPs using either single-cell or k-cell data alone , and typically within a single population of cells . However , integrating both single- and k-cell data may reap additional benefits , and quantifying differences in CHPs across cell populations or conditions could reveal novel biological information . Here we present a Bayesian approach that can utilize single-cell , k-cell , or both simultaneously to infer CHPs within a single condition or their differences across two conditions . Using simulated as well as experimentally generated single- and k-cell data , we found situations where each data type would offer advantages , but using both together can improve precision and better reconcile CHP information contained in single- and k-cell data . We illustrate the utility of our approach by applying it to jointly generated single- and k-cell data to reveal CHP differences in several key inflammatory genes between resting and inflammatory cytokine-activated human macrophages , delineating differences in the distribution of ‘ON’ versus ‘OFF’ cells and in continuous variation of expression level among cells . Our approach thus offers a practical and robust framework to assess and compare cellular heterogeneity within and across biological conditions using modern multiplexed technologies .
Transcriptomic profiling is widely used in biomedical research , but until recently it often relies on measuring mRNAs pooled from thousands to millions of cells , thus obscuring the well-appreciated biological variation that exists among individual cells of the profiled population . Quantifying variation in gene expression across single cells could help address fundamental biological questions and empower new applications previously not possible using cell-population based measurements . Such new applications include de novo assessment of tissue composition without a priori knowledge on cell-type defining markers [1 , 2] and inferring biologically relevant changes in cell-to-cell variations . Despite rapid technological advances , accurate measurement of single-cell expression is a major challenge , particularly because many mRNAs are expressed at levels close to or below the detection limit of current profiling technologies [3 , 4] . For example , the estimated rate of capturing individual mRNA molecules ranges from ~10% to ~20% using state-of-the-art single-cell RNA-Seq protocols [4 , 5] . Indeed , typical single-cell gene-expression data obtained by quantitative PCR ( qPCR ) or RNA-Seq contain a substantial number of zero or non-detected measurements ( “non-detects” ) , which cannot be entirely attributable to cells expressing zero transcripts . For example , some non-detects may arise from technical factors such as measurement noise , and missed capture or amplification of mRNA transcripts at or near the detection limit , as revealed by recent studies using measurements of spike-in standards and statistical inference methods [6–12] . An alternative approach to direct single-cell profiling , called “stochastic profiling” [13] , has been proposed to mitigate detection issues: measure the expression of random pools of a small number of cells ( k ) ( e . g . , k = 10 ) , followed by computationally deconvolving these pooled-cell measurements to infer the underlying cell-to-cell variation parameters . This approach offers more robust detection due to the increased amount of input mRNA and has been used to , for example , assess whether expression distributions across cells are bimodal [13–15] . Each approach can offer advantages , e . g . , single-cell for its direct interpretability and k-cell for improved sensitivity and therefore better quantitative estimates of certain cell-to-cell variation parameters . In principle they can also be complementary , and when both data types are obtained from a cell population , utilizing them together could potentially provide richer information for assessing cellular heterogeneity than using either one alone; however , in practice , no approach has been developed to take advantage of both data types simultaneously . To utilize both data types jointly and also allow the flexibility of using either one alone , here we present a Bayesian approach ( called QVARKS ) that quantifies the degree and the statistical uncertainty of expression variation across cells by using k- and/or single-cell data , after accounting for technical detection limits . A key contribution of our approach includes a newly developed statistical model and associated Bayesian inference and model assessment procedures that can handle single-cell , k-cell , or both data types jointly to infer cellular heterogeneity parameters ( CHPs ) , including the fraction of cells in the population expressing the gene ( “ON” cells ) or variation in expression level among “ON” cells . Both types of cellular heterogeneity can reflect meaningful biology , for example , the former , or “discrete” heterogeneity , may capture the frequency of functionally distinct cell subsets as classically defined by marker gene expression , while the latter , or “continuous” heterogeneity may capture the spread of the expression phenotype that could ultimately influence the overall population-level response to a perturbation [16] . Another feature of QVARKS is that it can model data jointly from two distinct cell populations to quantitatively assess differences in CHPs ( or “differential heterogeneity” , DH ) between the two conditions . While assessing differences in mean expression between two conditions is widely applied , the biology of differences in cell-to-cell expression variations ( CEV ) has been underexplored . Given that CEV can play functional roles and can be under genetic regulation [17] , QVARKS can be used to help reveal gene expression heterogeneity among cell populations , such as those exposed to different environments or from distinct developmental lineages . QVARKS thus complements existing single-cell data analysis approaches that either focus on identifying differential expressed ( DE ) genes [7 , 11 , 18] , or aim to find genes with high overall variability but do not deconvolve the overall variability into discrete vs . continuous components [6 , 12 , 10 , 19] . We systematically assessed the performance of QVARKS using both simulations and joint single- and k-cell data obtained from two biological conditions . We took advantage of QVARKS’ flexibility to handle different input data types to study the relative performance of using single-cell , k-cell or both data types to infer CHPs in a single condition or compare them between two conditions . We found scenarios where different input data types ( single-cell , k-cell , or both ) offer advantages . However , integrating both single- and k-cell data often offers the advantages of both . We also evaluated whether single-cell data would lead to inferred parameters consistent with k-cell data and vice versa , and found many situations where single- or k-cell analysis on its own led to significantly different results . Thus , this argues for proper integration of the two data types for robust parameter estimation and cross-checking them for consistencies when possible . We illustrate the practical biological utility of our approach by applying it to compare CEV in resting versus inflammatory-activated human macrophages , an important immune cell type known to function in diverse tissues and biological processes , including chronic inflammation associated with numerous common human diseases and aging [20] . QVARKS revealed significant differences in the CEV of key genes ( e . g . , RELA , a component of NFκB ) upon inflammatory activation , potentially reflecting condition-dependent regulation of cellular heterogeneity . QVARKS is provided as an R package with detailed documentation ( see Data Availability for download URL ) , and thus offers a practical , robust approach to quantitatively assess and compare CEV within and among biological states or conditions .
We focus on assessing two aspects of CEV for a given gene . First , “discrete” heterogeneity , arises due to the presence of cells with zero ( OFF cells ) vs . non-zero expression ( ON cells ) of the gene . Biologically , this type of CEV can originate from differences in the transcriptional status or activity at the gene locus among single cells , e . g . , some cells have actively transcribing promoters while others have inactive promoters . Second , “continuous” patterns of CEV among ON cells reflecting , for example , that some cells have higher levels of upstream transcriptional regulators than other cells; inherent stochasticity in biochemical reactions , such as transcript and protein production , can also contribute substantially to continuous variability among single cells [17] . QVARKS mathematically models these two CEV patterns for each gene and the Bayesian procedure described below infers the value of these parameters using single-cell ( SC ) data alone , k-cell ( KC ) data alone , or both ( SCKC ) . The overall framework is depicted in Fig 1 . The output of QVARKS includes an estimate of the CEV parameters ( or their differences between conditions when run in two-condition mode ) and the statistical uncertainty around each of the parameters ( all computed from the posterior distribution inferred by the Bayesian procedure; see description below and in Methods ) . We first sought to assess the relative performance of the three input data modes of QVARKS ( SC , KC and SCKC ) in inferring CHPs ( π , μ and σ ) across a range of scenarios involving either a single condition/cell population , or two conditions/cell populations aimed at comparing the inferred CHPs between the two conditions . Here , the unique capability of QVARKS for handling all three input data types served as a common inference procedure to help evaluate their relative performance . We simulated single- and k-cell data ( using k = 10 and 50 ) subjected to measurement by an assay that would suffer from increasingly missed detection as the input mRNA level is lowered and additional measurement noise under a range of scenarios , and used the resulting single-cell data ( SC ) , k-cell data ( KC ) , or both ( SCKC ) to infer the value and statistical uncertainty of the CHPs ( using a posterior surface scanning procedure; see Methods ) . The number of samples was fixed at n = 1000 across all three approaches–i . e . , n single-cell , n k-cell , or n/2 single-cell and n/2 k-cell samples were used . Data were simulated under scenarios reflecting low , medium and high levels of difficulty , corresponding to , respectively , high π and low σ ( i . e . , high ON fraction and low cell-to-cell variation among ON cells ) , medium π and σ , and low π and high σ ( i . e . , low ON fraction and high cell-to-cell variation among ON cells ) ( see S1 Fig and Methods ) , as well as assays with varying detection efficiencies ( bad , medium and good sensitivity assays corresponding to , respectively , 18% , 50% and 82% average detection of single-cell samples and nearly perfect detection of all k-cell samples; see Methods ) . The simulated data was further subjected to known , realistic sources of experimental noise including sampling , amplification and efficiency noise [12] using five distinct noise configurations ( see S1 Fig and Methods ) . Assessing such diverse scenarios is important and informative since a wide range of possibilities is expected across the tens to thousands of genes targeted by multiplexed techniques such as microfluidic qPCR and RNA-Seq . Performance across the input data types was evaluated via the error ( differences between inferred and true values ) and statistical uncertainty of the parameter estimates ( Fig 2 and S2 and S3 Figs ) . Our single-condition simulations revealed that 1 ) as expected , SC’s performance on inferring both discrete and continuous heterogeneity ( π and σ ) was worst in medium- or high-difficulty scenarios using assays suffering from sensitivity issues ( Fig 2 and S2 Fig ) ; 2 ) SCKC was comparable to KC across a range of scenarios , and tended to be better than KC when k is larger using assays of medium to high sensitivity for inferring π and μ—these were similar to the scenarios under which SC is better than KC when high sensitivity assays are used ( see Fig 2 and S2 Fig ) . Note that SCKC’s inferred values for π and μ ( and to a lesser extent , SC’s ) tended to center more tightly around the true values than those inferred from KC , because a given k-cell expression distribution for a larger k could be explained by scenarios involving different combinations of π and μ ( e . g . , more ON cells because of higher π but lower mean expression μ , or alternatively , less ON cells but higher μ ) and these possibilities could be better disentangled by having single-cell data using a sensitive assay; under this scenario in the case of SCKC , the samples allocated for single-cell measurements provide unique information not well-captured by k-cell data . Depending on the expression-level distribution of the target genes and the assay detection properties , a larger k may sometimes be warranted to improve detection and therefore SCKC could be advantageous in such situations as suggested by these simulation results ( see Discussion ) . Note that these conclusions held largely across the different measurement noise configurations tested except for a few cases as shown in S2 Fig . A key aim of QVARKS is to enable a quantitative , comparative assessment of heterogeneity parameters of each gene between two biological conditions . Thus we augmented the single-condition assessments with two-condition simulations . We simulated each gene under two conditions with different ON cell fractions ( e . g . , 80% ON cells in one condition vs . 50% in the other ) and subjected them to the same logistic assay and measurement noise settings as in the single-condition scenarios ( see Methods ) . While the relative performance trend of the three methods in inferring parameter differences between the two simulated conditions tends to be similar to that when inferring single-condition parameters ( S3 Fig ) , the performance difference among methods , particularly that between SCKC and KC , was less pronounced in the comparative setting . Again , as above , the measurement noise setting tends to have little effect on the relative performance of different methods except for a few cases as shown in S3 Fig . Examining single- and two-condition simulation results together revealed differences in how measurement noise affects inference outcomes . As expected , as more noise is added , all three methods suffered from increased error in their estimates within individual conditions because none of the methods explicitly model experimental noise . However , such noise-induced error appeared largely mitigated when comparing across two conditions ( S4 Fig ) , likely because such errors tend to cancel out in comparative analyses between two conditions . However , under some comparative scenarios , such as when both the mean expression and cell-to-cell variation are different between the two conditions , estimates of differences could still be error-prone due to the dependence of technical noise on average/mean expression . This error/bias can potentially be handled by explicitly checking the mean-variance relationship to assess whether the observed difference in heterogeneity of a given gene can be accounted for by changes in the average expression alone between two conditions ( as illustrated in the biological application below; see also S11 Fig ) . In summary , our simulation results revealed scenarios where different input data types can each offer advantages . We confirm that when assay sensitivity is high , SC can be desirable particularly for providing directly observable estimates of discrete heterogeneity ( π ) . As the assay sensitivity lowers and the inference difficulty increases , the advantages of KC become apparent , at the cost of masking biological heterogeneity at larger values of k . Our simulations revealed that SCKC tends to offer the best of both SC and KC under many scenarios , thus suggesting that simultaneous generation and integration of the two data types can be a robust , valuable approach , particularly under multiplexed settings where different genes would fall under different inference difficulty and detection scenarios as simulated above . Since certain features of real data cannot be fully captured by simulated data , we next sought to assess the relative performance of the three input data modes using single- and 10-cell data ( i . e . , k = 10 ) we had jointly generated for studying cell-to-cell expression variation of human macrophages in resting conditions ( control , CNT ) vs . those exposed to inflammatory cytokines for 24 hours ( IFNγ together with TNFα , hereafter referred to as IFNT ) ( see full description of this data below and in Methods ) . We performed several analyses to assess relative precision , or conversely statistical uncertainty , using appropriately downsampled data so that the sample sizes were the same across SC , KC and SCKC ( see Methods ) . Here we can only assess precision rather than error ( the difference between the true and the estimated values ) , due to the lack of ground truth about parameter values in real data . We first assessed the precision of estimated parameters as quantified by the credible interval ( CrI ) width reflecting the amount of statistical uncertainty about the true value of the parameter . In particular , we focused on genes with similar inferred values across the three input modes to avoid confounding from potential correlation between precision and error/bias . This analysis revealed that SCKC tends to provide more precise estimates for π and σ than either SC or KC alone across a larger fraction of genes ( Fig 3A and S5 Fig ) . By using a t-like fold-change statistic to assess statistical power for detecting changes in heterogeneity ( DH ) between two conditions , this improved precision of SCKC in comparison to SC or KC also translated to mild increases in sensitivity for detecting DH , especially for the π and μ parameters ( Fig 3B ) . We next evaluated whether using single-cell data alone would lead to inferred parameters consistent with k-cell data and vice versa . In a perfect world where the single-cell assays have 100% detection and the value of k is sufficiently small so that masking of biological cell-to-cell heterogeneity due to convolution of k cells is minimal , single- and k-cell data would reveal similar information about cell-to-cell heterogeneity . In reality , these two data types may provide different parameter estimates due to a variety of reasons , including their differences in susceptibility to detection limit problems , and thus it would be informative using simultaneously generated single- and k-cell data to formally assess their consistency . Towards this end , we performed a cross-validation analysis by randomly dividing our data ( 84 single- and randomly down-sampled 84 ten-cell samples per condition ( IFNT and CNT ) ) in 2:1 ratio into training and testing sets ( Fig 4A ) . We used the training data to infer models separately using SC or KC and the testing fraction to assess how well the inferred model fits using our AD-test model assessment criteria ( see Fig 1B and Methods ) . This analysis showed SC-inferred models tend to explain only the corresponding single-cell test data well but not the k-cell test data , and similarly but to a lesser degree , the KC-inferred models are more aligned with the k- than single-cell test data ( Fig 4B ) . However when both training data types were used ( SCKC ) , the inferred models fit both the single and k-cell test data substantially better than using the SC or KC inferred models alone ( Fig 4B ) . To an extent , these observations are to be expected given that SCKC uses both single- and k-cell training data ( half of each to keep the same sample size across methods ) . Yet , importantly , our results revealed that single- or k-cell analysis on their own could lead to models irreconcilable with the other data type , each of which contains important information about the expression distribution of a gene across single cells . Given the highly multiplexed nature of modern gene expression assays , a sizable number of genes could fall under such “incompatible” scenarios , thus providing support for the generation , cross-checking and proper integration of both data types when possible . While the focus of QVARKS is to infer CHPs within a single condition or compare CHPs between two conditions , several existing single-cell approaches focus on assessing differential expression between conditions using single-cell data . QVARKS can also be used to assess DE using one of the three possible input data modes . Assessing DE affords us another way to check our model assumptions and the overall method because QVARKS computes DE using the posterior distribution of its model parameters and thus achieving reasonable DE results require that our inferred posterior distributions be accurate . Here following previous studies [11 , 18] , we view the DE estimates ( log2-fold-change ( log2 ( FC ) ) of the average expression of a gene between two conditions ) computed from bulk , cell population-level RNA-Seq data as “ground truths” . Our bulk RNA-Seq data was generated in a related study in the same macrophage conditions as our single- and k-cell data . In addition to comparing against bulk RNA-seq , we also assessed QVARKS’ performance in recapitulating bulk DE relative to that of two other single-cell based DE methods: MAST [18] and SCDE [11] ( S6 Fig ) . Specifically , we applied MAST and SCDE to our macrophage qPCR data and compared with QVARKS running in SC , KC or SCKC modes using relative ( Pearson correlation shown in S6A and S6B Fig ) and absolute ( average squared error shown in S6C and S6D Fig ) measures of recovering ground-truth fold-change values . Evaluation based on the relative , correlation based measure revealed that QVARKS SCKC and KC nicely capture bulk level DE and perform comparably to MAST and SCDE , and QVARKS SC performs comparably once genes with large CrI are removed ( S6A Fig ) . QVARKS SC in general yields more parameter estimates with large uncertainty ( large CrI ) as has also been observed above ( Fig 3A ) , thus removal of these genes led to improved performance . Evaluation based on the absolute measure showed that DE estimates from all QVARKS input modes were well-calibrated with respect to the ground-truth values , and thus all had relatively low absolute error ( S6C Fig ) . Thus , the overall DE performance of QVARKS suggests that our model assumptions and inference procedures are reasonable and capture the data well . To assess and illustrate the practical applicability and utility of QVARKS , we next applied QVARKS to the full single/10-cell macrophage data set introduced above . Here we focused our analysis using SCKC as the input data so that we can use all available data; SCKC also tends to provide the most robust results as suggested by the evaluations above . The data was obtained in a related study using fluorescence activated cell sorting ( FACS ) of single and ten cells followed by microfluidic qPCR ( measuring 93 transcripts and 3 spike-in control RNAs; see Methods ) . Macrophages are immune cells that exhibit diverse environment-dependent phenotypes and hence are good models to study how the environment shapes CEV in transcript levels , such as changes in the fraction of ON cells ( π ) . Environment- or signal-induced changes in gene expression have largely been assessed by measuring alterations in average expression using a population of cells , but measuring the average alone could miss changes in CEV . By taking advantage of the posterior distributions containing measures of statistical uncertainty around the inferred heterogeneity parameters , we can begin to quantitatively assess changes in CEV in resting vs . IFNT-activated ( inflammatory ) macrophages . Our Bayesian procedure successfully modeled and obtained posterior distributions of model parameters for 53 and 60 genes in CNT and IFNT , respectively . Of these , 41 are shared between CNT and IFNT , and hence their CEV can be quantitatively compared across these two conditions ( see below ) . A majority of the remaining genes failed model assessment due to scarcity of detected cells ( observed ON cells less than 5 in single-cell data for 23 failed genes in CNT and 17 failed genes in IFNT ) . Since these genes are inherently difficult to work with for inferring CHPs ( e . g . inferring ON-cell variance from 5 or fewer cells is challenging ) , they could essentially be filtered out a priori . Under such a filter , the model assessment rate increases to 71–75% ( 47 out of 66 filtered genes in CNT , and 51 out of 68 in IFNT were successfully modeled ) . Thus , consistent with our DE assessment above , our model assessment indicates that for many genes , the data is captured well by our models , suggesting that our model assumptions , including distribution choices , are also robustly supported by the data . Despite the relatively high sensitivity of qPCR assays , signs of imperfect detection were apparent given that the average expression derived from single-cell data tended to be consistently lower than that obtained from 10-cell data ( divided by 10 ) for transcripts expressed at medium or low levels ( S7 Fig ) . Thus , some apparently “OFF” cells likely had some non-zero level of expression that simply escaped detection . Indeed , the inferred fraction of OFF cells ( 1-π ) for a majority of genes in both conditions was consistently lower , albeit only slightly in most cases , than that observed on single-cell data alone ( Fig 5A ) . As expected , genes with a higher observed fraction of OFF cells also tended to have larger uncertainty for π because it is more difficult to narrow down model parameters ( including ones associated with detection ) using information from just a few ON cells ( analogous to the “high difficulty” simulation scenario; see Fig 2 ) . However , even for such “difficult” genes , not all of the OFF cells can be attributed to escaped detection–the inferred fraction of OFF cells is at least 25% for a majority of these genes even after accounting for statistical uncertainty ( based on counting genes whose CrI is above 0 . 25 in Fig 5A ) . Thus , our Bayesian analysis integrating single- and 10-cell data helped estimate the proportion of OFF cells attributable to detection issues and thereby helped obtain better estimates of the expression distribution among single cells for a majority of the genes we measured . Bulk differential expression ( DE ) between two conditions can be attributed to a combination of changes in the fraction of ON cells and in the mean expression level among ON cells [7] . Here QVARKS provides a rigorous framework to explore the relationship between bulk differential expression ( DE ) and alterations in heterogeneity parameters π and μ [7] upon inflammatory activation . We assessed the status of these parameters for 23 DE genes between IFNT and CNT conditions ( adjusted P-value or adjP < 0 . 05 ) determined using the posterior distribution of the average mRNA level ( reflected by the mean μ of ON-cells weighted by π; see Methods ) –these genes were largely similar to DE genes obtained directly from 10-cell data ( see Methods and S8 Fig ) . Eight of the DE genes had significant changes in π but not μ ( e . g . , RELA ) , while four had the opposite behavior ( e . g . , FTL ) , and the rest either had significant changes in both ( e . g . , CD274 ) or neither ( Fig 5B ) . Note that many genes in the last category lacked significant change in π and μ not necessarily because the mean fold changes are small , but more because the statistical uncertainty around their inferred fold changes is large . Taken together , our results show that in macrophages adapting to inflammatory stimulation , both “digital” ( altering π ) and “fine-tuning” ( altering μ ) modes of regulating gene expression are prevalent . A key application of QVARKS is to assess changes in heterogeneity parameters ( or DH ) , such as asking whether environment shapes cellular heterogeneity . Thus , we next inferred differences between IFNT and CNT in terms of 1 ) the standard deviation of expression in ON cells ( σ ) reflecting continuous heterogeneity and 2 ) the ON-fraction ( π ) reflecting discrete heterogeneity ( Fig 6 ) . The inferred differences in these CHPs for successfully modeled genes were also verified to be robust against well-to-well variation in detection efficiency quantitated using spiked-in ERCC control mRNAs ( see S9 Fig; note that detection efficiency was a major source of technical noise assessed in [12] ) . In addition to π and σ , comparative assessment of heterogeneity across conditions can also be extended to other notions of heterogeneity , such as the Shannon entropy function computed using π , by taking advantage of our Bayesian approach’s ability to handle inference of any function of the CHPs ( see S10 Fig and Methods ) . Our analyses revealed both shared and condition-specific patterns of CEV ( Fig 6A and 6D ) . Genes such as FTL had similar continuous and discrete heterogeneity across conditions . However , RELA , a component of the NFκB complex , for example , had a higher fraction of ON cells ( adjP = 2 . 15e-05 , calculated as described in Methods ) as well as higher variability among ON cells in IFNT than CNT ( adjP = 0 . 012 , Fig 6C ) . Such changes in continuous variability could not be solely attributed to differences in average expression among ON cells ( μ ) or fraction of ON cells ( π ) between the conditions ( S11 Fig ) . The inferred change in σ across the two conditions for genes such as RELA was also robust against sampling noise when random ( 50% ) downsamplings of the data were assessed ( S12 Fig ) , and is thus potentially reflective of changes in biological cell-to-cell variation . Since NFκB is a key transcription factor mediating responses to diverse inflammatory signals in macrophages [25] , an increase in the heterogeneity of RELA expression upon inflammatory activation by IFNT suggests that elevating the response diversity among cells in , for example , infected tissues may play an important functional role , such as to counteract bacterial targeting of activated cells or to prevent detrimental over-inflammation . CLEC7A also showed changes in both types of heterogeneity , albeit in opposite directions , i . e . , up in continuous ( adjP = 0 . 002 ) and down in discrete ( adjP = 0 . 0006 ) variations . Differences in σ could , for instance , reflect changes in stochastic dynamics during transcription and mRNA degradation , or in negative feedback regulation . Thus , changes in both heterogeneity parameters in inflammatory vs . resting macrophages can be robustly detected by our approach . It will be interesting in the future to study the underlying mechanisms and function of such distinct gene regulation modes revealed by the differential heterogeneity and differential expression tests enabled by QVARKS .
Environment- or signal-induced changes in gene expression have largely been assessed by measuring alterations in average expression using a population of cells . However , measuring the average alone could miss changes in cell-to-cell heterogeneity . Assessing such changes in a statistically rigorous manner has been challenging , in part due to difficulties in disentangling technical vs . biological variations in single-cell measurements . To help overcome these challenges , we have developed a flexible computational method for inferring CHPs within a cell population or across two cell populations using single-cell , k-cell , or both data types simultaneously , by obtaining measures of statistical uncertainty around the inferred heterogeneity parameters so that the significance of an observed change can be evaluated . Here using inflammatory activation of human macrophages as a first biological application , we have uncovered a number of significant changes in the fraction ( π ) of and/or variance ( σ ) among ON cells . Furthermore , we have shown that changing π alone ( “digital” gene expression ) can contribute to changes in mean expression at the bulk level . Our approach can also be used to explore whether cellular heterogeneity can change without accompanying changes in average expression , e . g . , increasing π and σ while reducing μ such that the population average is unaltered . Indeed , it is increasingly recognized that cell-to-cell variations could themselves be regulated both genetically and environmentally without affecting the average expression level to , for example , reduce the chance of noise-induced cellular activation by lowering heterogeneity or hedge against environmental fluctuations by ensuring sufficient variation across cells in a population [17] . However , it is worth noting that despite our modeling of technical factors such as misdetection , interpreting continuous CEV parameters ( the variation among ON cells , or σ ) can be more challenging in general than interpreting discrete CEV parameters ( the fraction of ON cells , or π ) . For example , other sources of technical noise are still possible and thus the biological and technical sources may not be fully disentangled , and additional checks , such as those in S11 Fig based on linear regression , do not account for non-linear relationship between mean and variance , and could violate some linear regression assumptions ( such as error-free measurement of independent variables and homoscedasticity ) . One way to help disentangle biological from technical variation in the future is to utilize dilution-series experiments to obtain an estimate of the technical variation of all gene assays at different average expression levels , although sampling/pipetting noise also need to be considered properly at lower ends of the concentration scale . The unique ability of QVARKS in handling the three input data types ( SC , KC and SCKC ) allowed us to assess their relative performance for the first time under a common Bayesian analysis framework . While analyses of both simulated and real data suggest that there are situations under which each of single- or k-cell input data types can offer advantages , in general combining single- and k-cell data—when both are generated simultaneously—could yield more robust estimates of CHPs compared to using single- or k-cell data alone across a range of scenarios , particularly for genes with moderate to poor detection . A further practical appeal of using single- and k-cell data jointly , particularly in multiplexed settings such as microfluidic qPCR and RNA-Seq where tens to thousands of genes are measured , is that the single-cell data for genes with good detection ( e . g . , highly expressed genes ) can be directly analyzed to enable applications that require measurements of multiple genes within individual cells , such as identifying novel cell subsets . However , our approach is flexible and therefore can also be applied to single-cell data for assessing CHPs within a single condition or comparatively between two conditions . Our methodological framework can be extended to incorporate additional parameters or use alternative model parameterizations to account for features such as multimodality within ON cells , e . g . , in samples comprising mixed cell types or subsets . But the inference accuracy would be constrained by sample size . For example , each additional mode would require three extra parameters per biological condition in order to capture its frequency , mean and variance ( assuming using a log-normal distribution ) . Thus , it would only become feasible to learn the extra parameters for such models when larger sample sizes are available , which will likely be the case in the near future as the experimental cost of profiling continues to drop with the introduction of new technologies [5] . Here we chose to use the unimodal distribution for ON cells for several reasons . First , it is simple , yet captures our single- and k-cell data well—as discussed above , our model assessment criteria indicated that this parameterization was sufficient to successfully fit a majority of genes . Among the remaining genes , the predominant reason for the lack of fit appeared to be the scarcity of ON cells rather than bimodality among them: there are only two GCCs with more than 10% ON cells that appeared to be multi-modal ( TIGD6 in CNT and PTGS1 in IFNT according to Hartigan’s Dip test for unimodality using a relaxed P < = 0 . 2 cutoff ) . While additional multimodal genes may be present in our study , they were not evident at our current sample size . The model assessment procedure used in this study , the default in the R package implementing our QVARKS approach , leans on the stringent/conservative side ( i . e . , allows less genes to pass model assessment ) , in that we required most data sets generated from each of the posterior parameter draws to be statistically indistinguishable from the observed data ( see Methods ) . Users of the R package can choose other forms of model assessment , including a variant of the one used here where data simulated from different posterior parameter draws are concatenated into one dataset , and a single AD test is carried out to assess the concordance of this dataset against the observed data . This approach can work well for genes with just a few ON cells since concatenating all simulated datasets into one provides a larger sample size to enable a more robust comparison against observed data . However , one advantage of the default approach is that it matches the sample sizes of the simulated and observed data when performing each of the multiple AD tests . Another model assessment criterion that our R package supports is based on the well-accepted posterior predictive p values ( ppp; [23 , 24] ) , wherein we empirically test if a particular aspect of the observed data , such as mean or variance of observed ON cells , is captured well by simulated datasets generated from the posterior parameter draws . However , the ppp requires the user to choose which aspect of the data ( e . g . , mean and/or variance ) to use , and does not use information from the whole data distribution the way the AD test does . The vignette/manual of the QVARKS R package illustrates and compares these model assessment options . Having multiple model assessment options allows the users to have more means to determine whether to trust a particular gene’s fitted model , and depending on the goals of the analysis , proper tradeoffs can be made between ensuring that model assumptions are comprehensively met vs . discarding expensive single- or k-cell data . For example , consider a gene that is expressed in a high number of ON cells in condition A but in very few in condition B . If the goal were to compare the continuous heterogeneity of genes between the two conditions , it would be challenging to analyze this gene as the very few ON cells in condition B would yield unreliable estimates for σ in that condition . However , changes in the discrete heterogeneity ( π ) or average differential expression can be more reliably studied for this gene between the two conditions . When the k-cell approach is used , the optimal value of k could depend on the cell type and assayed genes , thus empirical assessment is needed for choosing k . For example , a dilution series of bulk mRNA can be profiled to estimate the minimum value of k that would provide reasonable detection for most genes . Our simulation assessment suggests a tradeoff between the value of k and sensitivity: increasing k would better mitigate detection issues at the cost of masking biologically relevant heterogeneity from convolving single-cell samples . Indeed , in general k should be large enough to avoid detection issues , but sufficiently small to avoid such masking effects where the relative contribution of biological cell-to-cell variations would become increasingly small and thus undetectable due to technical noise . Similarly , the optimal proportion of single- vs . k-cell samples to profile is context-dependent and best chosen using an empirical approach . When no prior information is available , as in this study , one could start with an equal proportion of both data types ( e . g . , 100 single- and 100 k-cell samples ) . One can then use this data to estimate the relative contributions of each data type to the model likelihoods and use this information to help determine the relative proportion of single- and k-cell samples to profile in the next round of experiments . This process can be iterated to fine-tune the proportion until a fixed profiling budget is reached . However , until the cost and time of sample preparation and sequencing drop to negligibly low levels , this iterative strategy can be prohibitive; thus , in practice , it is best to simply profile an equal number of single- and k-cells and only generate more data of one of the data types if it is quantitatively clear that additional data would help with answering the biological question of interest . The vignette document distributed along with our QVARKS R package describes the functionality of our package using actual data examples and discusses issues related to experimental design ( e . g . , the dilution-series experiments discussed above when k-cell data is desired ) as well as statistical considerations ( e . g . , data quality assessment , preprocessing , transformation steps , and model assessment ) and computational considerations ( e . g . , scalability of the inference procedure to process a large number of genes via parallelization ) . Though QVARKS was primarily developed to analyze single-cell qPCR datasets , as a proof of concept , the vignette also illustrates how QVARKS can be applied to an externally pre-processed single-cell RNA-Seq dataset using a parallel computing cluster . Our approach thus provides a principled and practical method to explore cellular heterogeneity in diverse settings .
Peripheral blood was obtained by leukapheresis from a de-identified healthy donor by the NIH Blood Bank , Department of Transfusion Medicine , National Institutes of Health Clinical Center under IRB-approved protocol 99-CC-0168 . We motivated and described the parametric model of gene expression in single- and k-cell samples from a cell population in the main text . Here , we formally specify the model along with the likelihood function of model parameters . As shown below , the choices of specific distributions in our model are inspired by earlier single-cell modeling studies ( eg . bimodal mixtures of ON/OFF cells and log-normally distributed ON-cell transcript levels assayed using qPCR [7] or RNA-Seq [18 , 21] , and logistic detection behavior of qPCR [8 , 22] or RNA-Seq [11] assays ) . We extend the existing single-cell models by integrating both single- and k-cell data , explicitly modeling imperfect detection assays and handling more than one environmental condition under which a gene is profiled . Multiple environmental conditions are handled by keeping the detection behavior parameters the same across conditions ( which is reasonable as the same gene assay is used across conditions ) and letting the CHPs ( π , μ and σ ) be condition-specific . We present the single-condition model first and then point out what changes need to be made to handle the two conditions focused in this study . An interesting future direction would be to extend these single/two-condition models to handle more complex experimental designs , e . g . , those involving more than two conditions and handling covariates such as donor information , batch effects , and cell size . We extend a model of “underlying” or “true” single-cell expression in an earlier study [7] to obtain a model of “measured” single-cell expression that handles imperfect detection as follows . Let G denote the expression of a gene transcript ( i . e . , counts of transcript copies without any log transformation ) in a random cell in the population . Then the bimodal mixture model for G could be written in random variable notation as G = I X , where I ∼ Bernoulli ( π ) is an indicator variable that is 1 for an ON-cell of the gene with probability π and 0 otherwise , and log2X ∼ Normal ( μ , σ ) . Note that the log of average overall expression is given by log2 ( E ( G ) ) = log2 ( E ( I ) E ( X ) ) = log2 π + μ + σ2 loge ( 2 ) /2 ( in line with the intuition that both fraction and mean expression of ON cells determines the overall expression output , with σ2 contribution coming from cells in the tail of a log-normal distribution ) . The measured expression H ( e . g . , 2Et values of the gene across single-cell samples in a given condition , where Et = 40-Ct is a unit of measurement for qPCR assays as described below ) is either detected at the true underlying measurement G with a certain detection probability DP ( G ) ( which ranges from 0 to 1 and increases with expression strength in a logistic manner ) or not detected ( 2Et = 0 ) otherwise . So H = J ( G ) G , where J ( G ) ∼ Bernoulli ( DP ( G ) ) is an indicator variable that is 1 for detected measurements and 0 for non-detects , and DP ( G ) is a logistic function with intercept c and slope m , i . e . , DPc , m ( G ) =1/ ( 1+exp ( −log2 ( G ) −cm ) ) . This parameterization reflects the standard way of modeling the probability of a binary response , such as non-detect vs . detects in expression profiles . We can now write the likelihood function of the model parameters given single-cell expression measurements h1 , h2 , … , hn of the gene in n cells for a given condition as below . Here , Pμ , σ ( X = x ) denotes the log-normal probability density function evaluated at x or equivalently the normal density function with mean μ and variance σ2 evaluated at log2 ( x ) . Since each k-cell measurement is modeled as aggregating mRNAs from k randomly chosen independent cells and subjecting it to the same imperfect detection as single-cell data , we can write “true” k-cell expression V=∑i=1kGi ( with the Gis being independent and identically distributed as G above , and hence fully defined by the single-cell expression parameters π , μ and σ ) and the k-cell measurement W = J ( V ) V ( in the same way that H is obtained from G using the c , m-parameter logistic “assay” ) . Since the true single-cell distribution G is bimodal , the true k-cell distribution V is multimodal depending on how many ON cells constitute a particular k-cell pool ( denoted l ) . The likelihood function of this multimodal distribution can be approximated either empirically or analytically [15]: The empirical approximation becomes more accurate by increasing the number of repeated samples but at the burden of increased computational time , and the log-normal sum approximation we use , called the Fenton-Wilkinson approximation , is accurate when the standard deviation σ is not too large [26] . Both approaches yielded similar likelihood values for the parameter configurations we checked ( involving σ values ranging from 0 . 25 to 2 , and using 10 , 000 samples to build each KDE ) –so we eventually chose the analytical approach in this study for computational efficiency reasons . The likelihood function of the model parameters given k-cell measurements w1 , w2 , … , wn of a gene in n k-cell pools for a given condition can either be directly read off from the KDE for each measurement and multiplied together in the empirical approach , or given by the formula below in the analytical approach . Here , Pμ , σ ( X ( l ) = x ) denotes the Fenton-Wilkinson approximated probability density function of the sum of l ON-cell transcripts evaluated at x , and Pπ ( #ONcells=l ) = ( kl ) πl ( 1−π ) k−l denotes the probability of observing l ON-cells in a random k-cell pool . This likelihood function agrees with intuition that pools of small number of cells retain information on single-cell variations , but that of large number of cells lead to “bulk or masked-out” effect that makes the π and μ parameters inseparable , as also noted in the main text when presenting simulation results . Because as k increases , the binomially distributed number of ON cells in a random k-cell pool becomes more tightly concentrated around the value kπ , and so the k-cell mixture asymptotically collapses from ( k + 1 ) components corresponding to different numbers of ON cells to a single component whose distribution is the sum of kπ i . i . d . log-normal variables . Note that the sum of kπ i . i . d . log-normal variables can be ( Fenton-Wilkinson ) approximated by another log-normal with log-space parameters μ’ = ln ( kπeμ ) + ( σ2 − σ’2 ) /2 and σ’2=ln ( ( eσ2–1 ) / ( kπ ) +1 ) . Such a k-cell mixture essentially makes the three heterogeneity parameters inseparable ( or non-identifiable ) using k-cell data alone , since the same k-cell distribution ( given by the same μ’ , σ’ ) can be explained equally well by different combinations of π , μ and σ . Note that though in reality the k-cell mixture model collapses to more than a single component for large k values , the log-space parameters of these different components would be so close to each other that it would be difficult to separate them out using reasonable sample sizes . To integrate single/k-cell data and obtain overall log likelihood of the five model parameters given both single/k-cell data , we simply add the log of the k-cell likelihood and log of the single-cell likelihood specified above . To handle two environmental conditions , we add three new parameters π″ , μ″ and σ″ for the new condition and use the same likelihood function specified above to compute the likelihood of all eight model parameters given each condition’s single/k-cell measurements separately and finally add the log-likelihood corresponding to all measurements together . Since we use a non-informative prior , the joint posterior distribution is proportional to the likelihood function . In case of an informative prior , the posterior is proportional to the product of the prior probability and the likelihood function ( i . e . , P ( parameters | data ) ∝ P ( parameters ) P ( data | parameters ) ) . In other words , what you know about the parameters after the data arrive is what you knew before , and what the data told you [27] . To infer the joint posterior distribution of all parameters , we use a random walk Metropolis ( RWM ) Markov Chain Monte Carlo ( MCMC ) procedure with an adaptive tuning phase [27] . The proposal distribution or proposed jump for a RWM MCMC is an independent Gaussian variable in each direction ( each of the five parameters for single-condition or eight parameters for two-condition gene expression ) with mean 0 and a certain fixed proposal variance . This would work for a posterior with unbounded support , but would be too computationally inefficient for our posterior with bounded support ( due to MCMC steps being wasted in non-permissible values of the parameter , which do not satisfy parameter constraints such as 0 ≤ π ≤ 1 and σ ≥ 0 ) . Therefore we use a truncated Gaussian proposal for each parameter , with the truncation ensuring that each new proposed state of the MCMC is always within the constrained parameter range . Note that the pdf ( probability density function ) of a Gaussian distribution truncated to the interval [a , b] is 0 outside this interval , and same as the normal Gaussian pdf otherwise , except for a uniform scaling of pdf values within the interval by a constant so that the integral is unity . In the case of qPCR expression data in Et = 40-Ct units [7] ( also see below for our qPCR data description ) , we imposed additional constraints to further improve computational efficiency including 0 ≤ μ , σ , c ≤ 40 and 0 ≤ m ≤ 5 . These constraints can also work for other data types such as RNA-Seq data , once the RNA-Seq-based read counts have underwent additional pre-processing steps ( such as expression noise thresholding and log-transformation [18 , 21] ) that are required to reveal the bimodal mixture of OFF/ON cells and log-normal distribution of ON-cells for conforming genes . Our implementation is in the R statistical environment and extends the RWM MCMC implementation in the R package mcmc that works for any unbounded continuous distribution on Rd to the case of a continuous distribution with bounded support ( which in our case is the posterior or likelihood function for non-informative priors ) . The whole MCMC-based inference procedure involves three phases: a ) an adaptive phase where the proposal variance is tuned to achieve good mixing–we use a noisy gradient algorithm as implemented in the R package JAGS for this phase using a maximum of 400 , 000 iterations , b ) a burn-in phase where the tuned proposal variance is fixed and the chain is allowed to mix for 20 , 000 iterations and all resulting samples discarded , and c ) a final sampling phase where the actual samples are collected over 200 , 000 iterations . These iteration counts for single-condition inference are multiplied by two for a two-condition inference , and were determined based on pilot MCMC runs and inspection of convergence diagnostics . Convergence diagnostics such as MCMC trace plots and autocorrelation times are also reported alongside each gene’s inference , and could be used to filter out genes with poor convergence; however we rely instead on model assessment “fit” criteria of the parameter posteriors as described below to filter out poorly fit genes . The 90% credible interval ( shortest interval containing 90% mass ) is constructed from the empirical cumulative density of the posterior samples of each parameter using the R package coda . For model assessment , i . e . , to assess whether the data is captured well by our inferred models and satisfies our model assumptions including specific choices of distributions , we used concepts from the posterior predictive checking framework [23] . We specifically check the agreement between the distribution of observed data and of data generated from ( “predicted by” ) 100 inferred models , each of which was specified by parameters independently drawn from the posterior distribution , following the framework of Graphical posterior predictive checks but replacing the graphical check with a more quantitative AD-test check ( Section 6 . 4 of [23] ) . The specific model assessment criterion we used required more than 75% of the models drawn to be capable of generating data samples statistically indistinguishable from the observed single- and 10-cell data ( i . e . , AD-test P ≥ 0 . 1 ) . The Anderson-Darling or AD test is similar to the KS-test in that it tests if the same distribution could have yielded two sets of data samples ( which in our case are the observed single- or k-cell samples and the same number of samples simulated from the inferred single- or k-cell model respectively ) . We use the ad . test implementation in the R package kSamples . In addition to filtering out genes that do not pass this AD-test criteria , we also excluded genes with extremely high variance of ON cells ( posterior mean of σ greater than 5 in any of the two conditions ) as part of model assessment . Genes whose inferred models pass these model assessment criteria in both conditions are referred to as successfully modeled genes , and genes that pass model assessment in a given condition are referred to as successfully modeled gene-condition models or gene-condition combinations ( GCCs ) . Note that besides the default model assessment approach just described , our R package QVARKS also supports other model assessment options , including a single AD-test approach and posterior predictive pvalues ( see Discussion above and vignette/manual provided with our R package ) . When comparing two conditions for changes in an inferred quantity ( either the parameter such as σ or a function of the parameters such as average overall expression of single-cells specified above as the log mean of G ) , we reduced the joint posterior distribution of the quantity in both conditions into a P-value . We chose to report P-values , instead of other hypotheses comparison measures such as the Bayes factor , to facilitate interpretation by researchers more familiar with classical “frequentist” statistical notions and to permit us to adjust P-values of all tested genes to account for multiple testing ( using the Benjamini-Hochberg FDR procedure ) . We specifically converted the posterior probability that a parameter/variable in one condition is different than that in another condition to a P-value using a half-space approach [28]; this approach requires specification or estimation of the proportion of null hypotheses among all tested genes and we specify it conservatively at 100% . Several strengths of MCMC-based Bayesian inference are at play here in comparative assessment of heterogeneity in different conditions: CD14+CD16- human peripheral blood monocytes were purified to >98% purity by magnetic bead negative selection ( Dynabeads Monocyte selection kit , Invitrogen , Carlsbad , CA ) . These monocytes were aliquoted and frozen in 10% DMSO , 40% human serum , and 50% X-vivo 15 . Frozen monocytes were thawed and differentiated into macrophages by in vitro culture on tissue culture-treated plastic dishes in X-Vivo-15 media ( Lonza , Walkersville , MD ) supplemented with 100 ng/mL macrophage-colony stimulating factor ( M-CSF , R&D Systems , Minneapolis , MN ) treatment over a period of 6 days with a full media change and M-CSF re-addition at day 3 and day 6 . Fresh media on day 6 was supplemented with M-CSF + 100 ng/mL Interferon ( IFN ) γ and 100 ng/mL Tumor necrosis factor α ( IFNγ+TNFα or IFNT; both from Peprotech , Rocky Hill , NJ ) , or M-CSF alone ( Control or CNT ) . Cells were harvested for analysis 24 hours after treatment ( day 7 of differentiation ) by scraping in cold media . We used fluorescence activated cell sorting ( FACS ) on a BD FACS-Aria II ( Becton-Dickinson , Mountain View , CA ) to sort single and ten cells into individual wells of 96-well low-profile PCR plates ( Bio-rad , Richmond , CA ) followed by reverse transcription and 18 cycles of specific-target pre-amplification using Celldirect one-step RT-PCR kit ( Invitrogen ) . Preamplified cDNA was quantified using microfluidic qPCR ( Taqman probe-based qPCR assays ( Life Technologies , Carlsbad , CA ) targeting 93 gene transcripts ( selected based on their relevance in macrophage activation and core cellular processes ( e . g . , metabolism , RNA processing , core transcriptional and translational regulators ) as well as representative genes from modules of co-expressed transcripts having myeloid-enriched expression patterns obtained using published gene expression data [29] ) and three spike-in artificial control RNAs from the ERCC spike-in set ( Life Technologies ) ) to comparatively assess cell-to-cell expression heterogeneity of human macrophages between CNT and IFNT conditions . The single-cell and 10-cell samples in each condition were profiled in four Fluidigm 96 . 96 plates ( Fluidigm Corporation , South San Francisco , CA ) . To reduce technical confounding when comparing single cell responses between conditions , we FACS sorted cells from each of the conditions ( CNT and IFNT conditions , as well as two other conditions from a related study ) using a balanced distribution across multiple 96-well plates , followed by qPCR profiling . QPCR was performed on a Fluidigm Biomark instrument using the normal speed cycling gene expression protocol . Data was exported from Fluidigm Real-time PCR Analysis software version 3 . 1 . 3 , using Linear ( Derivative ) Baseline method , a global threshold of 0 . 01 , and a 0 . 65 quality threshold , parameters which were found to exclude non-specific amplification and reduce plate-to-plate variation . We converted gene expression data exported from Fluidigm Real-time PCR Analysis software from Ct ( Cycle threshold ) units to the more convenient Et = 40-Ct units , as in previous studies [7] , so that transcript copy counts can be approximated by 2Et values upto a scaling factor as assumed in our model . Even if this copy count assumption does not hold exactly ( or even approximately ) for certain gene assays due to less than 100% qPCR efficiency , the model assessments performed after our MCMC-based Bayesian inference should automatically exclude such problematic assays . In the future , we could recover such gene assays if we know their qPCR efficiency ( e . g . , based on independent titration standards experiments ) and use it to derive the copy counts instead of the 100% efficiency that is often assumed . We assigned samples where a gene is not detected an Et of–Infinity , and called them as non-detected samples ( non-detects ) . After excluding samples with fewer than 10% of all assayed genes detected ( as they may reflect wells that failed to amplify or receive a sorted cell ) , we had 84 single-cell and 88 10-cell samples available per condition for downstream analysis . Since the single and 10-cell pool samples in CNT and IFNT conditions ( along with samples in the two other conditions from a related study ) had to be spread across 8 Fluidigm plates , all of which could not be run at the exact same time , we used spiked-in ERCC control RNA levels to track and correct for potential batch/plate effects . The ERCC control RNA were added at the same concentration to all the wells of each 96-well sorting plate prior to cell sorting , and we normalized all detected ( i . e . , Et > 0 ) gene measurements using the plate median of the ERCC spike-in with the highest Et value ( ERCC-0003 ) . In detail , we shifted all detected gene measurements in a plate by a plate-specific global factor , which is chosen such that the median ERCC-0003 level of the shifted data across all ( non-standard ) samples in the plate becomes the same across all plates . After removing plate effects using this plate-level normalization , the different plates’ data are concatenated to obtain one single/10-cell dataset per experiment . In certain analyses where we had to compare the precision of SCKC against SC or KC ( Fig 3 ) , to ensure the same sample size ( n = 80 ) across all three methods , we first randomly downsample the 84 single- and 88 ten-cell samples available per condition to 80 samples each per condition , and then run SCKC method on a random half of the single-cell and a random half of the 10-cell samples ( repeating this random halving ten times to inspect run-to-run variation ) , and SC or KC method on all single- or all 10-cell samples respectively . Note that to obtain the bulk population-level log2 ( FC ) values , data from a related study was used . Similar experimental setup as described for generating our Fluidigm qPCR data was used to obtain and differentiate cells from donors , but instead of using qPCR on single- and 10-cell samples in CNT or IFNT ( 24 hour post stimulation ) conditions , we analyzed bulk RNA material in CNT or IFNT ( 18 hours post stimulation ) conditions from three different human donors using the Illumina TruSeq Ribozero RNA-Seq protocol with 500ng of total RNA according to manufacturer’s instructions . Standard RNA-Seq data analysis methods were used: Tophat2 [30] for splice-aware mapping , featureCounts [31] for counting reads mapped to gene exons , and DESeq2 [32] to perform the DE test and generate the log2 ( FC ) values using the design”gene expression ~ Donor + Treatment” in R’s formula notation . We identified DE genes using the posterior probabilities converted to P-values as specified above . As described in the main text , we could also identify DE genes in a model-free or posterior-free fashion by directly testing for changes in k-cell expression mean in both conditions . The idea behind this test is that k-cell data is closer to bulk population-level data than single-cell data , and that bulk data when analyzed using t test or similar tests often is considered to yield more accurate ( “ground-truth” ) estimates of changes in overall expression of a gene ( DE ) than single-cell data ( though this is not the case for estimating CHPs , where single- and k-cell data are valuable ) . But this posterior-free method treats all non-detects in the k-cell data as truly zero expression , which may not be valid for some lowly expressed genes or low values of k . The posterior-free method uses a linear regression model of each gene’s k-cell measurement across two conditions against these dependent variables: treatment to model fold change between the two conditions , ERCC expression to adjust for potential well effects , and plate variable to adjust for potential plate effects . Non-detects are assumed to be have zero expression . In R notation , the linear model is “k-cell expression of a given gene ~ treatment + ERCC-0002-expression + ERCC-0003-expression + ERCC-0044-expression + plate” . We extract the P-values of the significance of the treatment coefficients ( log fold change between two conditions such as IFNT vs . CNT ) being different from zero , and adjust them for multiple testing using the Benjamini-Hochberg procedure . Only genes at adjusted P or adjP < 0 . 05 are declared as DE hits . We simulated datasets under different levels of biological and technical variation , thereby translating to different levels of difficulty when inferring the true parameters . We first explain this procedure for single-condition simulations and then outline changes to the procedure for comparative two-condition simulations . The mean of ON-cells μ is set at 10 Et units in all scenarios , whereas the other parameters are set at These three biological variation scenarios are illustrated in S1A Fig . The detection efficiency of the gene assay as given by the logistic function is also set at three levels . Specifically we set the logistic function’s parameters at: These settings of the logistic “assay” leads to negligible loss of k-cell samples on average , since the mean k-cell expression is k-fold higher than mean single-cell expression and the slope m of the logistic function is set above such that the detection probability changes from 0 to 100% in a steep fashion as a function of the increasing expression strength . Note that this likely favors the k-cell only approach when comparing its performance against our single/k-cell strategy ( since assay drop-offs may occur for certain genes in real k-cell data when k is smaller ) . Larger values of m that lead to a more gradual increase in detection probability with expression increase ( and hence drop-off or loss in both single/k-cell samples ) could also be tested for simulations in future , if dilution standards experiments of typical gene assays have this detection behavior . Besides noise from imperfect detection of gene transcripts , there could be additional experimental noise when measuring gene expression [12] . So we ran our simulations under four realistic configurations of measurement noise and a reference configuration of no additional noise . The four realistic configurations were based on quantitative measures of total technical noise in our experimental setup ( encompassing all noise sources such as efficiency , amplification and sampling noise [12] ) derived from a dilution series experiment conducted alongside our macrophage single/k-cell experiments . This titration data exhibited similar patterns of dependency between the level of total noise and mean expression ( S1B Fig ) as previously observed [12] , and thereby helped select these noise settings: We simulate single-cell/k-cell data ( SC/KC ) under any of the above “3 biological variation x 3 assay sensitivity x 5 noise configuration” scenarios = 45 settings using our model , and infer back the true values using only the simulated data . We could have used our MCMC procedure for inference , but wanted to rule out potential differences in the MCMC convergence behavior of the SC , KC or SCKC approaches from confounding the intrinsic differences among these methods . So we opted for a simpler posterior surface scanning procedure that involved computing the posterior in each combination of the parameter values listed below and using this coarse grid of likelihood values to approximate the posterior distribution and calculate its mode and 90% CrI ( credible interval ) reported in Fig 2 and S2 and S3 Figs discussed in the main text . The posterior surface is scanned at the grid points defined by these parameter values: Note that the posterior ( derived from single-cell or Fenton-Wilkinson approximated k-cell likelihoods ) is a function of just the five parameters above and not of any additional measurement noise configuration parameter , since our model is unaware of the additional measurement noise added to simulated data . We performed the comparative two-condition simulations similar to the single-condition simulations described above ( e . g . , under the same 3 assay sensitivity and 5 noise configuration settings ) , but with the following changes . Since we need to assess the performance of different methods in identifying changes in heterogeneity parameters between two conditions ( as opposed to inferring the parameter values within one condition as in the three single-condition biological variation scenarios ) , we now tested three comparative scenarios . A gene has In all these scenarios , the ON cells have a mean ( μ ) of 10 Et and a standard deviation ( σ ) of 1 Et in both conditions , so that both conditions are simulated using an assay with the same logistic detection behavior ( note that the logistic parameters in our simulations are derived from μ , σ as described above ) . Grid-based posterior scan is now run separately on both conditions of a simulated gene to approximate the posterior distribution of a parameter in each condition , and the two resulting distributions jointly analyzed to approximate the posterior of the difference in the parameter between the conditions . Relevant code snippets implementing the single/two-condition simulations are provided ( see Data Availability for download URL ) .
|
Different cells can make different amounts of biomolecules such as RNA transcripts of genes . New technologies are emerging to measure the transcript level of many genes in single cells . However , accurate quantification of the biological variation from cell to cell can be challenging due to the low transcript level of many genes and the presence of substantial measurement noise . Here we present a flexible , novel computational approach to quantify biological cell-to-cell variation that can use different types of data , namely measurements directly obtained from single cells , and/or those from random pools of k-cells ( e . g . , k = 10 ) . Assessment of these different inputs using simulated and real data revealed that each data type can offer advantages under different scenarios , but combining both single- and k-cell measurements tend to offer the best of both . Application of our approach to single- and k-cell data obtained from resting and inflammatory macrophages , an important type of immune cells implicated in diverse diseases , revealed interesting changes in cell-to-cell variation in transcript levels upon inflammatory stimulation , thus suggesting that inflammation can shape not only the average expression level of a gene but also the gene’s degree of expression variation among single cells .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"blood",
"cells",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"pathology",
"and",
"laboratory",
"medicine",
"engineering",
"and",
"technology",
"signal",
"processing",
"immunology",
"population",
"genetics",
"gene",
"pool",
"simulation",
"and",
"modeling",
"mathematics",
"signs",
"and",
"symptoms",
"statistics",
"(mathematics)",
"white",
"noise",
"population",
"biology",
"research",
"and",
"analysis",
"methods",
"statistical",
"distributions",
"white",
"blood",
"cells",
"inflammation",
"animal",
"cells",
"gene",
"expression",
"probability",
"theory",
"immune",
"response",
"diagnostic",
"medicine",
"cell",
"biology",
"genetics",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"physical",
"sciences",
"macrophages",
"evolutionary",
"biology",
"statistical",
"data"
] |
2016
|
Robust Inference of Cell-to-Cell Expression Variations from Single- and K-Cell Profiling
|
The proper temporal and spatial expression of genes during plant development is governed , in part , by the regulatory activities of various types of small RNAs produced by the different RNAi pathways . Here we report that transgenic Arabidopsis plants constitutively expressing the rapeseed SB1 SINE retroposon exhibit developmental defects resembling those observed in some RNAi mutants . We show that SB1 RNA interacts with HYL1 ( DRB1 ) , a double-stranded RNA-binding protein ( dsRBP ) that associates with the Dicer homologue DCL1 to produce microRNAs . RNase V1 protection assays mapped the binding site of HYL1 to a SB1 region that mimics the hairpin structure of microRNA precursors . We also show that HYL1 , upon binding to RNA substrates , induces conformational changes that force single-stranded RNA regions to adopt a structured helix-like conformation . Xenopus laevis ADAR1 , but not Arabidopsis DRB4 , binds SB1 RNA in the same region as HYL1 , suggesting that SINE RNAs bind only a subset of dsRBPs . Consistently , DCL4-DRB4-dependent miRNA accumulation was unchanged in SB1 transgenic Arabidopsis , whereas DCL1-HYL1-dependent miRNA and DCL1-HYL1-DCL4-DRB4-dependent tasiRNA accumulation was decreased . We propose that SINE RNA can modulate the activity of the RNAi pathways in plants and possibly in other eukaryotes .
Short Interspersed Elements ( SINEs ) are repetitive sequences , present in the genome of most eukaryotes and ancestrally derived from small functional RNAs ( tRNAs , 7SL RNAs or 5S RNAs ) [1] . SINEs can be transcribed by the RNA polymerase III ( polIII ) machinery [2] . They propagate in genomes following reverse transcription and integration due to their capacity to interact efficiently with the translation products of Long Interspersed Elements ( LINEs ) [3] , [4] , a family of active retrotransposons . SINEs copy number usually ranges from several hundred to several thousand in most eukaryotic species , except in mammals where tens of thousands up to millions of copies can be found [1] . Evaluating SINE impact on genome structure and gene expression has been the subject of numerous investigations in the past 20 years ( reviewed in [1] , [5]–[7] ) . Most of these studies have been conducted at the DNA level , by evaluating how SINE copies affect chromatin structure , DNA recombination , replication and transcription . The effect of SINE sequences in mRNAs and the corresponding impacts on splicing , editing , degradation and translation processes have also been evaluated . Recently , several SINE polIII-specific transcripts were shown to act as noncoding riboregulators of basic cellular processes , including transcription and translation , in stress situations or in specific tissues . In rodents , following heat shock , several members of the SINE B2 family are actively transcribed [8] , [9] . The B2 SINE RNA was shown to interact with and inhibit the RNA polymerase II complex , leading to a general repression of gene transcription in this stress situation [8] , [9] . The polIII-specific transcription of human Alu , rodent B1 and silkworm Bm1 SINEs can also be activated by several biotic and abiotic stresses [10]–[15] . Alu RNA was proposed to regulate translation either by modulating the activity of the Protein Kinase R ( PKR ) , a double-stranded RNA binding protein ( dsRBP ) that down-regulates translation in stress situations [10] , or by a PKR-independent process [16] , [17] . Recently , human Alu RNA was also shown to act as a modular transacting repressor of mRNA transcription during heat shock [18] . The rodent BC1 and human BC200 SINE-related elements are transcribed specifically in neurons where they regulate translation . BC1 and BC200 RNAs could potentially act as guides for the RNA-binding FMRP protein and regulate the translation of a small subset of neuron mRNAs [19] , [20] although this mode of action was recently contested [21] . These RNAs can also have a more general impact on neuron translation by trapping essential translation factors such as eIF4B and PABPs [22] , [23] . These different examples reveal that certain SINE loci produce non-coding regulatory RNA molecules that act on basic cellular functions . In this respect , SINE RNAs are similar to other polIII-transcribed riboregulators such as the cellular 7SK RNA regulating transcription elongation [24] and the viral VA1 and EBER1 RNAs regulating translation by interacting with PKR [25] . The Arabidopsis thaliana genome possesses six different SINE families representing a total of 334 repeated copies [26] , [27] . In a previous study , we introduced a single copy of a Brassica napus SINE founder locus ( SB1 ) under the control of its natural promoter in Arabidopsis and followed SINE RNA production and maturation in two independent transgenic lines [28] . Here we present evidence that the constitutive production of SINE RNA in these Arabidopsis lines can induce severe developmental defects . The SINE-induced phenotypes are similar to several RNAi mutant phenotypes . We show that SINE RNAs interact with a subset of highly divergent dsRBPs and affect the production of different families of small RNAs and the accumulation of their corresponding mRNA targets . Our results suggest that SINE RNAs influence the activity of a subset of dsRBPs and consequently , influence a variety of basic cellular processes including RNAi .
Fourteen Arabidopsis thaliana transgenic lines transformed with the founder SB1 . 7 ( na7 ) locus from Brassica napus [2] , [29] , [30] were generated from two independent transformation experiments . We observed that most T2 individuals from nine of the fourteen transgenic lines displayed an apparent and similar developmental phenotype . To further characterize this phenotype , two transgenic lines ( Col0-SB1 . 7 ( 4 ) and Col0-SB1 . 7 ( 18 ) ) , one for each independent transformation experiment , were selected . Both lines contained a single integration locus and were established at the homozygous state ( data not shown ) . The SB1 . 7 locus contains transcriptional cis-enhancer motifs that allow the SINE to partially escape transcriptional repression in its natural host [2] . SINE SB1 primary transcripts and maturation products were detected in the two transgenic lines by Northern hybridization followed by a 18 to 48 hour exposure time [28] . The level of SB1 RNAs in these lines is therefore much lower compared to other endogenous polymerase III products such as U6 RNA , which only require a few minutes of exposure after hybridization under identical conditions ( see Figure 1D ) . The global severity of the developmental defects was variable between the two lines . Also , the penetrance of the phenotype was variable within each transgenic line , as plants with relatively mild to severe developmental defects were observed in each population ( see Figure 1 for examples ) . Selfing plants with severe developmental defects gave progenies composed again of a mixture of plants with mild to severe developmental defects . The same result was observed when plants with mild defects were selfed , suggesting that the severity of the phenotype is somehow determined by a stochastic process during development . In Arabidopsis , SB1 transcription is associated with delayed growth and flowering time , abortive siliques , partial sterility , reduction of leaf and root size , leaf serration associated with a downward curvature , and partial loss of apical dominance ( Figure 1 ) . Several of these defects resemble those observed in hyl1 and drb4 mutants , which are impaired in the two dsRBPs required for miRNA and trans-acting small interfering RNA ( tasiRNA ) pathways , respectively ( see Figure 1B and 1C ) , suggesting that SB1 RNA could interact with RNA-binding proteins of the miRNA or tasiRNA pathways . To explain the observed similarity between SB1 expressing lines and RNAi mutants , we hypothesized that if SB1 RNA mimicked the structure of natural mi/tasiRNA substrates , it could interact with and titrate proteins involved in the biogenesis of these small RNAs ( Figure 1 ) . While SINEs derived from 7SL RNA ( including mammalian Alu and B1 ) conserve the RNA folding of the ancestral molecule [31] , [32] , this is usually not the case for tRNA-derived SINEs like SB1 [33] . Indeed , using enzymatic and chemical probing approaches , we recently confirmed that SB1 RNA do not conserve the ancestral tRNA folding pattern but instead adopt a structure consisting of three stem-loops with bulges and mismatches [33] . This SB1 RNA secondary structure raises the possibility that it could interact with dsRBPs given that the recognition of dsRNA by dsRBPs generally does not involve sequence specificity and several structured RNAs forming stem-loops with bulges or mismatches were shown to bind efficiently to dsRBPs [34] , [35] . The Arabidopsis genome has 19 dsRBPs , many of which are involved directly in RNAi [36] . These proteins include the four DICER-LIKE proteins ( DCL1 to 4 ) , the five dsRNA-BINDING PROTEINS ( HYL1 and DRB2 to 5 ) and the HUA ENHANCER1 ( HEN1 ) protein . Because the production of SINE RNA induces development defects that are similar to those of hyl1-2 and drb4-1 null mutants ( Figure 1 ) , we tested the capacity of SINE RNA to bind to HYL1 and DRB4 . HYL1 is part of the DCL1 complex and is involved in processing miRNA primary transcripts ( pri-miRNAs ) and short precursors ( pre-miRNAs ) [37]–[40] . In gel retardation experiments , we observed that SB1 SINE RNA , but not DNA or single-stranded RNA fragments of a similar size , associate with a recombinant GST-HYL1 fusion protein ( Figure 2A ) . Although a perfect RNA duplex also could bind HYL1 , this association was less efficient compared to SB1 RNA , suggesting that HYL1 prefers dsRNA substrates containing unpaired nucleotide bulges and/or distal loops ( Figure 2A ) . Using an RNase V1 protection assay , we defined more precisely the SB1 RNA binding sites of HYL1 ( Figure 3 ) . We observed that HYL1 binds mainly to the first and longest SB1 stem-loop , which corresponds to the region of SB1 RNA that adopts a fold similar to pre-miRNAs . Indeed , the protected region includes an RNA duplex containing mismatches and extends into the single-stranded terminal loop region ( Figure 3C ) . A structurally similar , although weaker HYL1 binding site also is present on the second stem-loop of SB1 RNA ( Figure 3 ) . We also tested the capacity of DRB4 , a dsRBP involved in the production of tasiRNAs [37] to bind SB1 RNA . In this case , the GST-DRB4 fusion protein did not bind significantly to SINE RNA in our in vitro assay , although it did efficiently bind to a perfect RNA duplex , which likely resembles the structure of tasiRNA templates ( Figure 2B ) . This result suggests that SINE RNAs interact with dsRBPs specifically adapted to bind imperfect double-stranded RNA rather than those that bind perfect RNA duplexes . Recently , double-stranded RNA binding domains from two Xenopus laevis proteins , xlADAR1 and xlRPBA , were shown to bind efficiently to short stem-loop RNA structures containing bulges and mismatches [41] . For xlADAR1 , this result is consistent with the observation that ADARs can bind and modify miRNA precursors in vivo [42] . Because most tRNA-derived eukaryotic SINEs can adopt an RNA structure similar to SB1 RNA , [33] it is possible that many SINE RNAs interact with dsRBPs . We performed binding experiments with SB1 RNA and the second double-stranded RNA binding domain of xlADAR1 ( called Dr2 ) and mapped the RNA binding sites . We observed that Dr2 bound SB1 RNA in the same region as HYL1 ( Figure 3 ) , suggesting that SINE RNAs have the potential to interact with a subset of dsRBPs across eukaryotic species , including the ones involved in miRNA production . DRB4 had no impact on RNase V1 cleavage pattern , confirming its inability to bind SB1 . Also , no obvious enhancing ( synergetic ) effect was observed when Dr2 and HYL1 were used in the same binding experiment ( Figure 3 ) . The binding of HYL1 appears to increase the RNase V1 sensitivity of certain regions of the SB1 RNA ( indicated by asterisks on Figure 3A ) . The RNase V1 activity is sensitive to RNA conformation and , although sensitivity does not always imply hydrogen bonding of the bases in a canonical double stranded helix , it does require a structured , helix-like conformation [43] . As such , the increased RNase V1 sensitivity following HYL1 binding suggests that HYL1 is able to force some single-stranded RNA regions to adopt a more structured helix-like conformation , possibly by promoting non Watson-Crick base pairing . To test a chaperon-like activity for HYL1 and to explore its generality , we performed binding experiments using the SELEX clone 11Dr2 ( 7 ) , a short imperfect double-stranded RNA known to bind the Dr2 motif [41] . Following RNase V1 digestion , we confirmed the binding of Dr2 to 11Dr2 ( 7 ) ( Figure 4 ) . We observed that HYL1 is able to bind strongly to 11Dr2 ( 7 ) and generate regions protected from RNase V1 activity ( represented by green lines on Figure 4B ) and regions with increased sensitivity to RNase V1 ( represented by asterisks on Figure 4A and B ) . Again DRB4 was unable to bind 11Dr2 ( 7 ) and no synergetic effect was observed when HYL1 and Dr2 were used together . Similar results were observed when the SB2 Arabidopsis SINE RNA was used as a substrate ( see Figure S1 ) . Our results suggest that , upon binding RNA , HYL1 has the general capacity to force single-stranded regions to adopt a more organized , helix-like configuration . To determine the molecular consequences of SB1 expression on the miRNA and tasiRNA pathways , we analyzed small RNA accumulation in our SB1 expressing lines . DCL1-HYL1-dependent miRNA accumulation was reduced in the two SB1 transgenic lines ( Figure 5 and data not shown ) . Reduced miR171 accumulation coincided with increased accumulation of its target SCL6-III RNA ( Figure 5A ) , suggesting that SB1 RNA could compete with miRNA precursors for HYL1 binding and thus reduce miRNA processing efficiency and miRNA-mediated regulation in planta . Consistent with the inability of SB1 RNA to bind DRB4 in gel retardation experiments ( Figure 2 ) , accumulation of DCL4-DRB4-dependent miRNA was unchanged in SB1 transgenic lines ( see Figure 5C ) . The accumulation of tasiRNA also was reduced in SB1 transgenic lines , presumably because tasiRNA production primarily relies on the action of DCL1-HYL1-dependent miRNA miR173 and miR390 ( Figure 5B ) . Indeed , reduced miR390 accumulation was consistent with reduced TAS3 tasiRNA levels and increased accumulation of TAS3 tasiRNA targets ARF3/ARF4 mRNAs ( Figure 5B ) . No change in HYL1 , DCL1 and HEN1 mRNA accumulation was detected in the Col0-SB1 . 7 ( 18 ) SINE expressing line ( see Figure S2 ) suggesting that the observed reduction in miRNA levels in this line does not result from repression of these miRNA pathway genes , and instead directly results from SINE RNA interaction with HYL1 .
SB1 expressing lines display a diversity of phenotypes , suggesting that many important developmental transition steps are affected in these plants . The variable phenotypic penetrance within each line also suggests a stochastic effect of the RNA on these transition steps . Although we do not know the precise molecular mechanism ( s ) responsible for these phenotypes , our data raise the possibility that an interaction between SB1 SINE RNA and a subset of dsRBPs , some of which are involved in RNAi , could contribute to the developmental defects . Using a double-stranded RNA binding domain from Xenopus laevis ADAR1 , we have shown that SB1 RNA is fit to bind highly divergent dsRBPs , and therefore many dsRBPs could be affected by SINE RNA expression . Not all Arabidopsis dsRBPs are known to be involved in RNAi . For example , FIERY2 is a dsRBP involved in transcriptional regulation [44] , while two other dsRBPs of unknown function ( At1g48650 , at5g04895 ) contain a helicase domain . Therefore SB1 RNA potentially affects basic cellular processes other than RNAi , and this in turn could affect plant development . Based on our in vitro studies , we propose that SB1 RNAs interact in vivo with HYL1 , and consequently modify the steady-state level of several miRNAs and tasiRNAs . In this scenario , tasiRNA accumulation would be indirectly affected because tasiRNA biogenesis relies on miRNA-guided cleavage ( miR173 targets TAS1 and TAS2 and miR390 targets TAS3 ) [45] ( Figure 5B ) . DCL4-DRB4-dependent miRNA accumulation is unaffected in SB1 transgenic lines ( see Figure 5C ) , consistent with the inability of SB1 RNA to bind DRB4 in gel retardation experiments . The SB1/HYL1 interaction and its molecular consequences on the miRNA pathway are unlikely to be solely responsible for the observed SB1-induced phenotype . Indeed , the global reduction of miRNA levels in SINE-expressing lines is generally moderate to low ( from 0 . 5 to 0 . 8 of the initial amount ) and this level of reduction does not always correlate with detectable increases in corresponding mRNA target levels ( for examples see [46] ) . These relatively modest changes might be because several of these miRNAs derive from multigene families and thus potentially arise from several RNA precursors . In such cases , SINE RNAs would need to compete with several differently structured RNA precursors to limit effectively miRNA production . Also , in vivo , such competitions likely are influenced by the varying tissue/spatial distributions of different miRNAs:HYL1 complexes , which would probably effect the capacity of SINE RNA to modulate the production of a given miRNA in a given tissue . In conclusion , we propose that SB1 RNA compete for several dsRBPs , not only HYL1 , and that these competitions likely accounts for the extent and unusual characteristics ( such as the variable penetrance ) of the SINE-induced phenotype . We observed that , in vitro , HYL1 binds efficiently different imperfect double-stranded RNA molecules , including the rapeseed SB1 ( Figures 2 and 3 ) and the Arabidopsis SB2 ( Figure S1 ) SINE RNAs , while DRB4 only binds perfect RNA duplexes ( Figures 2 and 3 ) . These results are fully compatible with the known natural substrates of these two proteins . In vivo , HYL1 is known to interact with pri- and pre-miRNAs , which are organized as stem-loops containing mismatches and bulges ( see the miRBase http://microrna . sanger . ac . uk/ for examples of pre-miRNA structures ) . On the other hand , DRB4 binds perfect linear RNA duplexes formed by the action of RDR6 on a single stranded primary transcript [36] . The fact that HYL1 does not play a major role in double-stranded RNA-induced posttranscriptional gene silencing ( PTGS ) [46] further suggests that , in vivo , HYL1 preferentially interacts with imperfect double-stranded pri- and pre-miRNAs and not perfect double-stranded PTGS precursors . Based on our RNase V1 mapping results , the basis of this selectivity could be the capacity of HYL1 to interact with single-stranded RNA regions ( Figures 3 and 4 ) . Indeed , the binding specificity of other eukaryotic dsRBPs , such as ADARs and Staufen , was shown to depend on their ability to interact with single-stranded RNA loops [47] . We therefore suggest that HYL1 has intrinsic RNA binding specificities distinct from DRB4 , and that these specificities dictate different in vivo binding preferences . We also observed that upon binding RNA HYL1 has the general capacity to force single-stranded regions to adopt a more organized , helix-like configuration ( Figures 3 , 4 and Figure S1 ) . In vivo , HYL1 mainly is involved in promoting processing steps from pri-miRNA to pre-miRNA in association with DCL1 , another dsRBP [38] , [39] . HYL1 also influences the cleavage positioning of DCL1 on the pre-miRNA to generate the mature miRNA [40] . Consequently , in the hyl1-2 null mutant , pri-miRNAs accumulate and misplaced cleavages of pre-miRNAs were observed in some cases [38]–[40] , [46] . However , HYL1 is not fully necessary for plant miRNA processing by DCL1 because the hyl1-2 mutant retains some ability to accumulate wild type miRNAs , although the accumulation level is reduced . Also , this reduction is variable depending on the different miRNAs [46] , [48] . HYL1 may therefore promote , to variable extents , the processing activity of DCL1 . Based on our observations , one way HYL1 could do this is by inducing a conformational change in the RNA structure , forcing key single stranded regions to adopt organized , helix-like , configurations , including non Watson-Crick base pairing . This could in turn be important for promoting the cleavage activity of DCL1 on pri-miRNAs or for helping to precisely define the cleavage site on pre-miRNAs to generate mature products . It remains to be determined whether the chaperoning-like activity of HYL1 is important for miRNA production . We recently observed that related structural motifs are present in most SINE RNAs from mammals , fishes and plants , suggesting common selective constraints imposed at the SINE RNA structural level [33] . Using a double-stranded RNA binding domain from Xenopus laevis ADAR1 , we have shown here that the plant SB1 RNA is fit to bind highly divergent dsRBPs . Therefore , the common trend of structural evolution observed for tRNA-related SINE could result in similar constraints imposed by a subset of dsRBPs across eukaryote species . If true , this predicts that SINE RNAs are under selective pressure to keep intact their capacity to interact with some dsRBPs . This would in turn forge the SINE RNA structure and impose , as observed [33] , a common evolutionary history for most eukaryote tRNA-related SINEs . The reason why SINE RNA/dsRBP interaction would be under positive selective pressure is unclear , but precise and punctual expression of SINEs during a key development step or in a stress situation , could induce genetic and/or epigenetic variations and increase diversity and/or adaptability . It is interesting to note that SINE-specific expression in their natural host is highly regulated at the transcriptional and post-transcriptional levels by complex genetic and epigenetic processes ( reviewed in [1] , [10] , [28] ) . SINEs are non-autonomous in their mobility and need the activation of an autonomous LINE partner to retrotranspose . Therefore , based on our results , we suggest that the main purpose of limiting SINE-specific transcription is not to prevent its mobility ( the control of LINEs is sufficient to achieve this ) but to preserve cell homeostasis by preventing SINE RNA to interact with a large subset of dsRPBs .
The construction of the SB1 expressing transgenic lines [28] and the hyl1 , drb4 mutant phenotypes [48] , [49] were described previously . Plants were cultivated on soil in a greenhouse in standard conditions . For the study of root growth , plants were cultivated on germination medium ( 1 time MS salts; 10 g l−1 sucrose ) plates at 21°C under a 12-h light/12-h dark regime . RNA structures were predicted using RNA/DNA folding and hybridization software Mfold , version 2 . 3 [50] . For the SB1 and SB2 RNAs , the predicted structure was confirmed experimentally using chemical and enzymatic probing [33] . cDNAs encoding HYL1 or DRB4 were amplified by PCR from an Arabidopsis cDNA library ( Stratagene ) using primers designed according to the Arabidopsis sequence database . All PCR amplifications were performed using 5′-primers with a terminal BamHI restriction site in combination with a 3′-primer ending with a XhoI restriction site . After PCR amplification and BamHI/XhoI digestion , the coding sequences of HYL1 or DRB4 were cloned into the pGEX-5X-1 expression vector ( Pharmacia Biotech ) . In the resulting constructs named pGEX-HYL1 or pGEX-DRB4 , HYL1 or DRB4 are fused to the C-terminal end of GST . Prior to expression in bacteria , sequencing was performed to verify the sequence of the cDNAs and the translational fusions . To express the HYL1 or DRB4 recombinant proteins , pGEX-HYL1 or pGEX-DRB4 were transformed into Escherichia coli BL21 ( DE3 ) cells . A single colony of E . coli cells containing a recombinant pGEX plasmid was used to inoculate 50 ml of LB medium containing 100 µg/ml ampicillin . Cells were incubated overnight at 37°C with vigorous shaking . Cultures were diluted 1∶100 into fresh LB medium containing 50 µg/ml carbenicillin and grown at 37°C with shaking until the A600 reaches 0 . 7–1 . Recombinant protein expression was then induced by addition of 0 . 1 mM isopropyl-ß-D-thiogalactopyranoside ( IPTG ) followed by an incubation of 4 to 5 hours at 37°C . Induced cells were harvested by centrifugation at 7700 g for 10 min at 4°C and pellets were frozen at −20°C overnight . Bacterial sonication and batch purification of the fusion proteins using Glutathione Sepharose 4B were performed according to the manufacturer's protocol ( Pharmacia Biotech ) . RNA was in vitro transcribed and radioactively trace-labeled ( for quantification purposes ) from linearized T7 promoter-containing vector using recombinant T7 RNA polymerase and [α-32P] ATP ( GE Healthcare ) . Transcripts were gel-purified . 2 pmol of RNA was then dephosphorylated using calf intestinal phosphatase ( New England Biolabs ) according to manufacturer's protocol . Dephosphorylated RNAs were 5′-end labeled with T4 polynucleotide kinase ( New England Biolabs ) and [γ-32P] ATP ( GE Healthcare ) . Gel-purified 5′-labeled RNAs were subsequently used for nuclease protection assays [41] . RNase V1 recognizes any 4-6-nt segment of polynucleotide backbone with an approximately helical conformation and cleaves leaving 5′-phosphates [43] . For the partial digest with RNase V1 , 20 fmol ( corresponding to 50 , 000 cpm ) of RNA were centrifuged , washed , dried and resuspended in structure buffer ( Ambion: 100 mM Tris at pH 7 , 1 M KCl , 100 mM MgCl2 ) . After annealing , 1 µL of tRNA ( 1 µg/µL , Ambion ) was added , followed by the addition of increasing protein concentrations ( 50 nM , 150 nM and 500 nM ) . To ensure protein binding to RNA , samples were incubated for 15 min at room temperature . Then , 0 . 005 units of RNase V1 ( Ambion ) were added and the reactions were incubated for additional 10 min at room temperature . Reactions were stopped by ethanol/salt precipitation . Samples were loaded together with alkaline hydrolysis ladder and denaturing RNase A ( Ambion ) and RNase T1 ( Boehringer Mannheim ) digests of RNAs on denaturing RNA gels . Total RNA was extracted using inflorescences ( stages 1–12 ) , as described elsewhere [28] . Northen blot analyses of mRNA accumulation were performed as described previously [51] . For the detection of small RNAs , 15 µg of total RNA samples were heat-treated in 1 . 5 volume of standard formamide buffer and loaded on 15% polyacrylamide ( 19∶1 acrylamide:bis-acrylamide ) - 8 M urea - 0 . 5× TBE gel and separated by electrophoresis . The samples were electroblotted to hybond-NX membranes ( GE healthcare ) and fixed following a carbodiimide-mediated cross-linking procedure [52] . Pre-hybridization and hybridization was carried out in 5× SSC , 20 mM Na2HPO4 pH 7 . 2 , 7% SDS , 2× Denhardt solution , 50 mg/ml denaturated hering DNA at 50°C . Filters were washed twice with 3X SSC , 25 mM NaH2PO4 pH 7 . 5 , 5% SDS at 50°C for 10 min , followed by one to two washes with 1× SSC , 1% SDS at 50°C . Signals were visualized using a phosphorimager ( Molecular Imager FX; Bio-Rad ) for quantification . The random-primed 32P-labelled probes used for the detection of the ARF3 , ARF4 and SCL6-III mRNAs have been described previously [48] , [53] . For mi/tasiRNAs detection , DNA oligonucleotides whose sequences are complementary to individual mi/tasiRNAs were 32P-labeled with T4 polynucleotide kinase ( New England Biolabs ) . For U6 detection the following oligonucleotide was used: 5′-AGGGGCCATGCTAATCTTCTC-3′ .
|
Short interspersed elements ( SINEs ) are transposable elements in eukaryotic genomes that mobilize through an RNA intermediate . Recently , mammalian SINE RNAs were shown to have roles as noncoding riboregulators in stress situations or in specific tissues . Mammalian SINE RNAs modulate the level of mRNAs and proteins by interacting with key proteins involved in gene transcription and translation . Here we show that constitutive production of a plant SINE RNA induces developmental defects in Arabidopsis thaliana and that this SINE RNA interacts with HYL1 , a double-stranded RNA-binding protein required for the production of microRNA and trans-acting small interfering ( tasi ) RNA . We mapped the binding site of HYL1 to a SINE RNA region that mimics the hairpin structure of microRNA precursors . We also found that HYL1 induces conformational changes upon binding to RNA substrates . These data suggest that SINE RNAs modulate the activity of RNAi pathways in Arabidopsis .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genetics",
"and",
"genomics/epigenetics",
"genetics",
"and",
"genomics",
"biochemistry/rna",
"structure",
"molecular",
"biology/rna-protein",
"interactions"
] |
2008
|
SINE RNA Induces Severe Developmental Defects in Arabidopsis thaliana and Interacts with HYL1 (DRB1), a Key Member of the DCL1 Complex
|
Geophysical models of climate change are becoming increasingly sophisticated , yet less effort is devoted to modelling the human systems causing climate change and how the two systems are coupled . Here , we develop a simple socio-climate model by coupling an Earth system model to a social dynamics model . We treat social processes endogenously—emerging from rules governing how individuals learn socially and how social norms develop—as well as being influenced by climate change and mitigation costs . Our goal is to gain qualitative insights into scenarios of potential socio-climate dynamics and to illustrate how such models can generate new research questions . We find that the social learning rate is strongly influential , to the point that variation of its value within empirically plausible ranges changes the peak global temperature anomaly by more than 1°C . Conversely , social norms reinforce majority behaviour and therefore may not provide help when we most need it because they suppress the early spread of mitigative behaviour . Finally , exploring the model’s parameter space for mitigation cost and social learning suggests optimal intervention pathways for climate change mitigation . We find that prioritising an increase in social learning as a first step , followed by a reduction in mitigation costs provides the most efficient route to a reduced peak temperature anomaly . We conclude that socio-climate models should be included in the ensemble of models used to project climate change .
According to many ancient myths , humans did not invent fire-making de novo but rather learned it from personalities like Prometheus and subsequently spread the practice amongst themselves . These stories reveal how ancient myth-makers already grasped the fundamental importance of social learning—the process whereby individuals learn new behaviours , values and opinions from others [1] . Social learning is no less relevant in the era of human-environment challenges [2–4] . The importance of social learning and social processes more generally in climate change mitigation and adaptation is well recognised [5–8] . Increasingly sophisticated geophysical climate models are helping us understand the impacts of anthropogenic greenhouse gas ( GHG ) emissions [9–11] , and the importance of these models is hard to understate . However , climate projections depend strongly on the assumed trajectory of GHG emissions [12] . This trajectory is determined by human behaviour and yet climate models generally do not incorporate dynamic social processes relevant to GHG emissions . Rather , GHG emissions are assumed to follow some specified trajectory . These trajectories are constructed with socio-economic factors in mind , ( see Representative Concentration Pathways [12] and Shared Socioeconomic Pathways [13] for instance ) , but are not coupled to climate dynamics and do not capture human responses to climate change in a mechanistic way . Just as human behaviour influences climate trends , climate change in turn influences human behaviour concerning GHG emissions , including both climate change mitigation and adaptation [5 , 6 , 8 , 14 , 15] . Individuals in places with rising average temperatures are more likely to perceive climate change [15] , and social effects are apparent when individuals take steps in response to such shifting perceptions [6 , 8] . There is also an important distinction between social learning and social norms—socially accepted and widely practised modes of conduct [16] . Social norms are known to have a strong influence on human behaviour [17] including aspects relating to climate change [7 , 16 , 18] and therefore play an important role in determining emission trajectories [6] . Multiple studies show a tendency for individuals to conform to emerging norms in support of climate change mitigation [7 , 16] . Moreover , it appears that individuals are often not consciously aware of the importance of social norms in their decision-making and instead falsely ascribe their decisions to other factors [18] . However , it is important to note that social norms do not automatically promote socially beneficial outcomes . They can equally well force conformity to a destructive norm such as political extremism [19] . This also happens in the context of climate change behaviour , where it has been found that individuals may also conform to a norm of non-mitigation , by adjusting their habits to match those of less environmentally friendly neighbours [18 , 20] . Hence , Earth’s climate and human subsystems are part of a single coupled system where social dynamics play a vital role . Yet , models of Earth’s coupled climate-behaviour system remain essentially undeveloped . One such approach [21] couples a sophisticated climate model [11] to a model for individual behavioural change based on the theory of planned behaviour—a dominant paradigm in psychology [22] . The authors find that the sensitivity of global temperature change to human factors such as response to extreme events , social norms and perceived ability to adopt mitigative strategies is of a similar magnitude to its sensitivity to geophysical factors . They deduce that quantifying behavioural uncertainty and physical uncertainty in climate projections deserve equal attention . The model focuses on how individual psychology and behaviour are influenced by extreme weather events . Social effects are modelled phenomenologically ( i . e . , exogenously imposed ) : individuals do not learn behaviour or opinions from one another , and social norms are treated as a fixed effect that does not depend on the population’s current composition of attitudes . Here , we treat social learning and social norms endogenously , by modelling their dynamics as they emerge from rules governing how individuals interact , learn and behave . Our first objective is to develop qualitative insights into how different aspects of the system—endogenous social processes , temperature trends , and mitigation costs—separately and together determine possible dynamics of the larger socio-climate system . Our second objective is to illustrate potential uses of coupled socio-climate models to chart social and economic policy pathways that mitigate climate change as quickly as possible . To meet these objectives , we sought to develop a model that ( 1 ) could capture a range of IPCC climate change scenarios , ranging from 4 degrees of warming by 2100 ( RCP 8 . 5 scenario ) to sub 2 degrees of warming ( RCP 2 . 6 scenario ) , ( 2 ) was simple enough to analyse so that we could learn which mechanisms drive the predicted socio-climate dynamics , ( 3 ) was based on existing approaches for modelling social dynamics and climate dynamics , and ( 4 ) captured the salient features of social and climate systems . Given the model’s simplicity , it is primed for insights as to how social and climate processes interact , though limited in its predictive capacity due to the complexity of the socio-climate system . The development of more complex socio-climate models will be an important research avenue , once the mechanisms of socio-climate dynamics are better understood .
Geophysical models in the climate science literature span a wide range of different complexities depending on the associated research objective . Highly complex models are the state-of-the-art for weather and climate prediction [11 , 23 , 24] , whereas simple models allow us to assess processes and feedbacks , thereby improving our intuition of climate system dynamics [25–29] . Likewise , the behavioural sciences have benefited from a variety of modelling approaches , that address the diverse set of social processes that take place on the individual and societal level [30] . Here , we use minimal models for both social and climate dynamics . Starting simple allows us to build intuition on the effect of socio-climate feedbacks that have yet been considered in the climate change literature . The social model is widespread and , despite its simplicity , captures the salient aspects of social dynamics [2 , 30 , 31] . Moreover , the simple Earth system model that we use [25] accurately follows the projections of the state-of-the-art CMIP5 models when forced with the IPCC emission scenarios ( S1 Fig ) . Over the period from 1800 to 2014 , the socio-climate model is simulated with a fixed social component , forced with historical anthropogenic carbon emissions . Initial conditions for all climate variables are zero since they represent deviations from pre-industrial values . Social dynamics are initiated in 2014 with an initial proportion of mitigators x0 = 0 . 05 . The ensuing dynamics of ϵ ( t ) follow an increasing but saturating trend corresponding to the world’s increasing but saturating population size and energy demands . Specifically ϵ ( t ) = { linear interpolation of historical emissions t ≤ 2014 ϵ 2014 + ( t − 2014 ) ϵ max t − 2014 + s t ≥ 2014 ( 14 ) where ϵmax is the saturating value , and s the half-saturation constant , of ϵ ( t ) . This expression is shown graphically in S7 Fig . The system of ( delay ) ordinary differential equations is simulated using the NDSolve package in Wolfram Mathematica . Historical CO2 emissions were obtained from the CDIAC data repository [35] . Baseline climate parameters are obtained from the original Earth system model [25] where they were fitted to obtain historical trends of temperature and carbon dynamics . Social parameters are more speculative and so are given wide upper and lower bounds . The relative cost of warming ( f ( T ) ) with respect to the net cost of mitigation ( β ) is chosen in accordance with the argument that the costs of preventative action will be far less than the cost implied otherwise by global warming [36] . For sensitivity analyses we draw parameters from triangular distributions that peak at baseline values and extend to upper and lower bounds ( S2 Table ) . Parameters are kept fixed preceding 2014 to retain historical trends in the simulations .
The model demonstrates how the social learning rate can strongly determine temperature trends . We first consider a null hypothesis where adaptive behaviour is removed from the model by forcing the proportion of mitigators in the population to remain constant . In this case of fixed behaviour , emissions saturate and the temperature anomaly increases indefinitely ( S2 Fig ) . However , once social learning is added and the proportion of mitigators is allowed to evolve dynamically as in our baseline model , the predicted average global temperature anomaly can peak anywhere from 2 . 2°C , near the Intergovernmental Panel on Climate Change ( IPCC ) limit [37] ( in the case of very rapid social learning ) to 3 . 5°C ( in the more realistic case where social learning unfolds on a generational timescale ) ( Fig 1a–1c ) . Whether people discuss climate change more or less often can therefore strongly influence temperature trends . Because we model social norms as something that tends to reinforce majority behaviour and attitudes—whatever they might be—one might think that social norms act as a double-edged sword . In fact , they operate more like an unhelpful scimitar , as illustrated by comparing cases of low and high strength of social norms . Because the population starts off from a state of largely non-mitigating behaviour , increasing the strength of social norms suppresses the spread of mitigating behaviour for decades by entrenching non-mitigation as a norm , even when rising temperatures strongly justify an immediate shift ( Fig 1d–1f ) . ( This model dynamic echoes not only current climate norms reinforcing non-mitigation [20] but also past social shifts occurring on decadal timescales , such as evolving social norms about when and where smoking is acceptable . ) However , when mitigating behaviour eventually does become widespread , a higher strength of social norms does not significantly accelerate its spread . Rather , the two curves for cases of high and low social norm strength simply move in parallel to one another because by this time , the utility function that determines behaviour change is dominated by the large temperature anomaly ( Fig 1d ) . In this parameter regime , social norms generate a perverse asymmetry , in contrast to findings from other socio-climate models that assume social norms can only support climate change mitigation [20] . The model also shows how a reduction in net mitigation cost can significantly accelerate the onset of social change . For instance , a 67% reduction in the mitigation cost increases the percentage of mitigators by 2060 from 10% to 90% ( Fig 1g–1i ) . Therefore , policies that reduce the cost of mitigation ( through e . g . subsidies , tax cuts ) will benefit from the accelerating effects of social learning and must be timed correctly . Our baseline model assumes that individuals’ perceived cost of climate change impacts depends on a linear extrapolation of the recent temperature anomaly over the previous ten years ( Methods ) . If individuals instead base their decisions only on the current temperature anomaly , the simulated global temperature anomaly lies well above the 2°C target set by the IPCC , and exhibits wide variation in sensitivity analysis ( Fig 2 ) . This contrasts with our baseline model where the population movement towards mitigative strategies ignites earlier , significantly reducing the global temperature anomaly . This predicted dynamic stems from the multi-decadal lag between GHG emissions and the consequent global temperature rise [38] . Our model predicts medium-term GHG emission trajectories ( Fig 1b , 1e and 1h ) that are qualitatively similar to those often assumed under various future emissions scenarios . This raises the question of how such models can be useful . The socio-climate model enables us to explore how socio-climate dynamics might respond to changes that are under the control of policymakers . For instance , it is possible to compare social and economic policy interventions by considering the effects of simultaneous parameter changes , instead of one at a time . This enables us to chart out the quickest pathways from highest to lowest temperature anomalies . The relative merits of increasing the social learning rate vs . reducing the net cost of mitigation are illustrated with a contour plot where the contours represent peak temperature anomaly as a function of the two parameters ( Fig 3 ) . Increasing the social learning rate ( e . g . through media coverage and public fora devoted to climate change ) is particularly effective when social learning is slow , but has saturating benefits , as indicated by the increasing vertical spacing of contour lines for at higher learning rates . In contrast , reducing the net mitigation cost ( e . g . through tax breaks ) drives a more linear response in peak temperature anomaly . Crucially , it should be noted that both a reduction of net mitigation cost and an increase in the social learning rate are required to achieve the IPCC target . The arrows in Fig 3 show the ‘path of steepest descent’—the most efficient combination of the two measures . Starting from a situation of high projected temperature anomalies , the model predicts that increasing the social learning rate should first be prioritised , followed by a reduction in net mitigation cost once the benefits of social learning begin to saturate . This approach gets us to the region of parameter space corresponding to the IPCC target faster than alternative trajectories . A sensitivity analysis reveals the relative influence of each parameter on the peak temperature anomaly ( Fig 4 ) . The time horizon of individuals’ temperature projection , social learning rate and costs of mitigation are major factors , all of which may be influenced by appropriate intervention . The importance of social parameter uncertainties in determining climate predictions indicated by our model has also been predicted by other socio-climate models [21] . Interestingly , the system is relatively insensitive to the initial proportion of mitigators , suggesting that the mediation of social processes , as opposed to the current social state , is key to guiding the socio-climate system to a trajectory of reduced emissions . Sensitivity analyses such as these can help investigators determine priorities for data collection: the parameters exhibiting the greatest influence on predictions should be targeted for data collection so we can best reduce model uncertainty . A striking feature revealed by the sensitivity analysis is the asymmetry in many of the parameter dependencies . Consider the three parameters with highest impact on the peak temperature anomaly ( concerning forecast horizon , learning rate and global warming costs ) . A decrease in these parameters is more detrimental than an increase is beneficial . For example , a forecast horizon 10 years above baseline value results in a 0 . 6 degree decrease in peak temperature anomaly , whereas a forecast horizon 10 years below baseline value results in a 1 degree increase . This imbalance is a manifestation of the nonlinear interactions between and within each of the social and climate system . The sensitivity analysis also reveals non-monotonic relationships between the peak temperature anomaly and the parameters . For example , both an increase and a decrease in solar flux results in a higher peak temperature anomaly . Interestingly , this is not the case if the climate subsystem is considered in isolation . For a fixed emissions scenario , a higher ( lower ) solar flux will always result in a higher ( lower ) peak temperature anomaly , since the solar flux is proportional to the net downward radiation absorbed by the planet’s surface . The coupling to social dynamics fundamentally alters this relationship . In the socio-climate system , a reduced solar flux results in a slower increase in surface temperature . As a consequence , individuals are less incentivised to mitigate , causing the social system to maintain a regime of non-mitigative behaviour . The accompanying high rate of CO2 emissions quickly overcompensates for the reduced solar flux , yielding a higher peak temperature anomaly . Thus seemingly useful interventions to the physical system can actually end up doing more harm than good when there is strong coupling to a social system , as is the case for global warming .
This study has shown how social processes can influence climate dynamics , according to one possible way of modelling social dynamics and norms . However , other frameworks for modelling human behaviour could yield different predictions . For instance , the socio-climate model of Ref . [21] does not include social learning . Individuals respond directly to changes in the climate , and not through interactions with one another . As a consequence , the rate at which individuals adopt mitigative strategies only varies with the current climate situation , and not with current population consensus . Mitigation efforts can therefore be expected to closely follow the severity of climate change in the model . In our model , social learning manifests as a feedback within the social system , resulting in qualitatively different socio-climate trajectories . Mitigative behaviour is initially suppressed–even as temperatures rise to levels that should incentivise mitigation–due to low numbers of mitigating individuals and therefore little turnover of behaviour in the population . However , social learning creates a positive feedback loop once there is a net positive utility to mitigate , and so as the numbers of mitigators increases , so too does the rate at which non-mitigators switch to being mitigators . This results in a sharp non-linear increase in mitigators , as a combined outcome of both the social and the climate system dynamics . We note that , all else being equal , adding social learning to a model has the effect of slowing down behaviour change in the human population ( since a process takes time , by definition ) , and therefore the mitigation response of human populations . Conversely , the case of very rapid social learning recovers a ‘best response’ model similar to those assumed in classical economics , where individuals immediately adopt the highest payoff strategy without learning the behaviour from others . Whether or not this assumption can approximate behaviour in real human populations hinges upon how fast social learning occurs—individuals would need to sample others rapidly enough to enable complete population behaviour change within 5 years for this approximation to work in our model , which seems implausible ( Fig 1a–1c ) . In a different vein , we assumed a homogeneous population with respect to mixing and individual utilities . The model’s social dynamics capture interactions at the individual level , though there are many different scales of social organisation that the model does not consider , from families/neighbourhoods to cities/states and up to interacting countries . Future models could include this more hierarchical social structure . Similarly , these models could include different types of individual with correspondingly different utilities . For instance , the model could include industrial corporations with utilities biased toward shareholder profit , and social institutions ( such as laws , taxes , the education system ) that reflect the current governmental stance . Social learning may also take on different forms due to diverse individual psychologies and values [39–41] . Such heterogeneities are known to affect the dynamics of a wide variety of systems [42] and can prevent population consensus by permitting development of echo chambers [43] . Our model also makes the simplifying assumption that individuals base their temperature projection on linear extrapolation of past temperatures . This could be generalised to a non-linear extrapolation to reflect an individual’s perception of ‘accelerating’ change . Extending socio-climate models to include these finer details should prove valuable in further investigations . Climate change is a manifestation of coupled human-environment dynamics and therefore we should start coupling climate models to social models [5 , 44] . Our simple coupled socio-climate model shows that the rate at which individuals learn socially strongly influences the peak global temperature anomaly , to the point that variation of this parameter within plausible ranges changes the peak temperature anomaly by more than 1°C . Therefore , it matters whether social processes cause slow or fast uptake of climate change mitigation measures . We found that social norms may not provide help when we most need it , although this finding could be nuanced by adding social heterogeneity . Finally , we illustrated how exploring the parameter space of socio-climate models suggests optimal paths for mitigating climate change . A more sophisticated policy impact assessment model based on a coupled socio-climate approach could therefore be useful to decision-makers facing a mandate to reduce GHG emissions with a fixed budget . In summary , it is essential for climate change research to account for dynamic social processes in order to generate accurate predictions of future climate trends , and the paradigm of coupled socio-climate modelling could help us address this challenge .
|
The importance of anthropogenic CO2 emissions on climate change trajectories is widely acknowledged . However , geophysical climate models rarely account for dynamic human behaviour , which determines the emissions trajectory , and is itself affected by the climate system . Here , using a coupled socio-climate model , we show how social processes can strongly alter climate trajectories and we suggest optimal intervention pathways based on the model projections . Steps to increase social learning surrounding climate change should initially be prioritised for maximum impact , making a subsequent reduction in mitigation costs more effective . Policymakers will benefit from a better understanding of how social and climate processes interact , which can be provided by socio-climate models .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"learning",
"anthropogenic",
"climate",
"change",
"ecology",
"and",
"environmental",
"sciences",
"chemical",
"compounds",
"atmospheric",
"science",
"social",
"sciences",
"neuroscience",
"learning",
"and",
"memory",
"physical",
"geography",
"simulation",
"and",
"modeling",
"cognitive",
"psychology",
"research",
"and",
"analysis",
"methods",
"climate",
"change",
"geography",
"behavior",
"chemistry",
"atmospheric",
"chemistry",
"physics",
"psychology",
"environmental",
"chemistry",
"climatology",
"earth",
"systems",
"earth",
"sciences",
"carbon",
"dioxide",
"biology",
"and",
"life",
"sciences",
"climate",
"modeling",
"greenhouse",
"gases",
"physical",
"sciences",
"geophysics",
"cognitive",
"science"
] |
2019
|
Charting pathways to climate change mitigation in a coupled socio-climate model
|
Lassa virus ( LASV ) , the causative agent of Lassa fever ( LF ) , is endemic in West Africa , accounting for substantial morbidity and mortality . In spite of ongoing research efforts , LF pathogenesis and mechanisms of LASV immune control remain poorly understood . While normal laboratory mice are resistant to LASV , we report that mice expressing humanized instead of murine MHC class I ( MHC-I ) failed to control LASV infection and develop severe LF . Infection of MHC-I knockout mice confirmed a key role for MHC-I-restricted T cell responses in controlling LASV . Intriguingly we found that T cell depletion in LASV-infected HHD mice prevented disease , irrespective of high-level viremia . Widespread activation of monocyte/macrophage lineage cells , manifest through inducible NO synthase expression , and elevated IL-12p40 serum levels indicated a systemic inflammatory condition . The absence of extensive monocyte/macrophage activation in T cell-depleted mice suggested that T cell responses contribute to deleterious innate inflammatory reactions and LF pathogenesis . Our observations in mice indicate a dual role for T cells , not only protecting from LASV , but also enhancing LF pathogenesis . The possibility of T cell-driven enhancement and immunopathogenesis should be given consideration in future LF vaccine development .
Lassa virus ( LASV ) is the causative agent of Lassa fever ( LF ) [1] . It accounts for an estimated number of 300′000 infections and several thousand deaths in endemic areas each year [2] , while imported cases have been reported from around the globe [3] . The virus is listed category A by the Center for Disease Control and Prevention [4] . So far , LASV vaccines have remained unavailable for clinical use , and Ribavirin , the only available therapy , has shown limited efficacy [5] . The development of effective vaccination strategies would therefore benefit from further insight into the mechanisms of successful LASV immune control , as well as into the processes underlying LF development and pathogenesis . It is generally agreed upon that the level of tissue damage observed at autopsy cannot by itself account for the severe nature of LF . Therefore , as with other viral hemorrhagic fevers [6] , [7] , a contribution of the host response to LF pathogenesis has long been a matter of debate . For instance , the manifestation of Dengue Hemorrhagic Fever ( DHF ) has long been accredited to pre-existing immunity [8] , [9] . Apart from serotype cross-reactive antibodies [8] , [9] , memory T cells were recently identified as important players in the disease process [10] , and susceptibility as well as resistance to DHF have been linked to particular MHC alleles [11] , [12] . In addition , infected monocytes and macrophages play an important role in DHF by secreting inflammatory cytokines [13] , [14] . Such contributions of the immune response to disease severity can represent a major hurdle in vaccine development [15] . For instance , formalin-inactivated vaccines to respiratory syncytial virus ( RSV ) and measles virus resulted in enhanced morbidity and mortality in response to natural infection [16] , [17] . Animal models for RSV have since provided evidence that T cell subsets play an important role in disease enhancement [16] , [18] . Interestingly , innate immune cells including eosinophils and polymorphonuclear granulocytes dominate the histological picture upon T cell-driven enhancement of RSV [18] . Similarly , inflammatory macrophage responses were found to be a common feature of viral hemorrhagic fevers [6] . In accordance with the “cytokine storm” hypothesis , macrophage-derived inflammatory cytokines [19] , [20] , [21] , [22] , [23] and nitric oxide ( NO ) [24] , [25] are candidate mediators of capillary leakage and shock [26] , and elevated levels of such mediators correlate with increased disease severity and worsened clinical outcome . Still , LASV lacks a clear pathognomonic signature , and clinical manifestations of LF are largely unspecific , making it difficult to diagnose the infection accurately via clinical criteria alone [27] . In contrast to other hemorrhagic fevers , coagulation abnormalities and bleeding are largely absent in LF [27] , [28] , leading some to argue on pathological grounds that Lassa fever ought not be considered a hemorrhagic fever at all [6] , [29] . More characteristic of severe LF cases are the vascular leakage with edema and effusions in the pleural and pericardial cavities [30] , [31] , [32] . At necropsy , liver and lung count among the organs most commonly affected during LF [29] , [30] , [31] , [32] , [33] . One of the few well-documented characteristics of primary LASV-directed immune response is that neutralizing antibody responses develop only weeks or months after the virus has been eliminated [28] . Also studies of vaccination-induced LASV immunity point toward a cell-mediated mechanism at the frontline of antiviral defense [34] , [35] . Still , this notion remains to be addressed and verified directly , and the responsible T cell subtypes to be characterized . Further , a potential disease-enhancing effect of T cell responses in LF has not yet been given sufficient consideration . Although normal laboratory mouse strains develop acute disease of the central nervous system when infected with LASV intracerebrally [1] , [28] , [36] , they remain resistant to the systemic disease so characteristic of human LF , irrespective of the route used to infect them . Research on LF has therefore been limited to the use of guinea pigs and non-human primates [28] , complicating mechanistic studies on immunity and pathogenesis . Here we report on a series of experiments triggered by accidental observations of serious disease in LASV-infected humanized mice ( HHD mice [37] , C57BL/6 background ) . HHD mice are genetically engineered to express a human/mouse-chimeric HLA-A2 . 1 molecule instead of the murine MHC class I gene products and are widely used to identify human HLA-A2 . 1-restricted peptide epitopes . Stimulated by these unexpected results , we were able to identify T cell-dependence of LASV control , but also of LF pathogenesis . These findings , combined with the propensity of LASV to target monocyte/macrophage lineage cells in vivo , followed by T cell-dependent activation of this cell population , provide a novel concept for virus-host relationship and pathogenesis of LASV . We anticipate that such understanding may aid rational refinement of both vaccine-mediated prevention and treatment of LASV infection .
We first compared viral replication in HHD mice and wild type C57BL/6 controls . C57BL/6 mice cleared LASV within about seven days after infection whereas HHD mice remained viremic for substantially longer periods of time ( Fig . 1A and data not shown ) . A detailed analysis of the initial phase of infection documented a virtually immediate uptake of the inoculum into tissues ( no virus in blood 2 . 5 hours after inoculation ) , followed by identical levels of viremia in wild type and HHD mice up to around day 4 ( Fig . 1A ) . This demonstrates that viremia reflected viral replication in tissues rather than residual inoculum , and that the early phase of virus replication was identical in HHD and C57BL/6 mice . Clear differences in virus control became evident no earlier than seven days after infection ( Fig . 1A ) . These differences in kinetics were compatible with differential adaptive immune control in HHD and C57BL/6 mice . Given that MHC class I ( MHC-I ) represents the only genetic difference between HHD and C57BL/6 mice , these findings suggested that H-2Db/H-2Kb-restricted T cell responses in C57BL/6 mice played an important role in virus control . Hence , we extended our study to further analyze the contribution of MHC-I- and MHC-II-restricted T cell responses to LASV control ( Fig . 1B ) . MHC-II-deficient mice ( lacking CD4+ T cells ) efficiently resolved the infection whereas MHC-I-deficient animals ( MHC-I-/-; targeted mutation of the β2-microglobulin gene; devoid of CD8+ T cells ) developed persistent high-level viremia . This corroborated the key role of MHC-I-restricted T cell responses in LASV control and indicated further that MHC-II-restricted responses were of lesser importance . A time course analysis of viral titers in kidney , lung , liver and spleen of LASV-infected HHD , C57BL/6 and MHC-I-/- mice confirmed that viral replication was comparable on day 2 and day 4 . By day 8 , however , LASV in the organs of C57BL/6 mice approached the detection limit whereas comparably high titers of virus persisted in tissues of HHD and MHC-I-/- mice . These data provided additional independent support for the above conclusions on productive replication of LASV in mice and the key role of MHC-I-restricted T cells in its control . Between 7 and 12 days after LASV infection , HHD mice developed ruffled fur and reduced spontaneous activity . Some of them rapidly and unexpectedly deteriorated and progressed to a state of agony and death: Five out of twenty three mice ( ∼22% ) infected in a total of five experiments succumbed to disease . In contrast but in accordance with the literature , all thirteen wild type C57BL/6 mice , serving as controls in three of these experiments , survived without clinical evidence of disease . Elevated serum aspartate aminotransferase activity ( AST ) represents the primary parameter for monitoring LF , and AST combined with viremia represent the best predictors for clinical outcome in humans [38] . In keeping with this manifestation of LF , serum AST activity remained mostly within normal ranges in wild type C57BL/6 controls but was significantly elevated in the serum of HHD mice , with a peak around day eight to twelve ( Fig . 2A ) . To further investigate how T cell responses restricted to H-2Kb/H-2Db and to HLA-A2 . 1 influenced LASV control and disease , we crossed HHD mice to C57BL/6 mice . C57BL/6 x HHD F1 mice express H-2Kb , H-2Db and HLA-A2 . 1 molecules owing to hemizygosity at all relevant genetic loci . C57BL/6 x HHD F1 mice controlled LASV infection as efficiently as did C57BL/6 wild type mice , and their AST levels remained within normal ranges ( Fig . 2B , C ) . This showed that H-2Kb/H-2Db-mediated virus control prevented disease even in the presence of the HLA-A2 . 1 molecule . At first it suggested also that persistent and high virus load was directly responsible for pathogenesis . Contrary to this notion , however , the experiments in MHC-I-deficient mice had not resulted in obvious disease despite persistent high-level viremia . This raised the possibility that primary T cell responses may contribute to LF in an immunopathological fashion , similar to the role of memory T cell responses in DHF [10] , [11] , [39] . To address this possibility , we infected HHD mice with LASV , and prior to infection depleted either CD8+ T cells or CD4+ T cells or both using monoclonal antibodies . MHC-I-/- mice are devoid of a CD8+ T cell compartment and were also included in the experiment . Unlike in untreated HHD mice and irrespective of comparably high levels of viremia in all groups ( Fig . 2E ) , serum AST levels of CD8/CD4-double-depleted HHD mice remained in normal ranges ( Fig . 2D ) . Also , depletion of only CD4+ or CD8+ T cells or genetic deficiency for MHC-I ( affecting the CD8+ but not the CD4+ T cell compartment ) afforded at least partial protection . In agreement with the results shown in Figs . 1B and 1C , these data suggested that T cells of HHD mice were unable to significantly influence viremia . Nevertheless CD8+ and CD4+ T cells played apparently an essential role in the pathogenesis of LF in HHD mice . To further characterize the role of T cells in the HHD mouse model for LF , we analyzed whether such animals could be immunized against LASV . For this , we used Mopeia virus ( MV ) , an apathogenic close relative of LASV . MV infection is known to elicit heterologous immunity against LASV in monkeys ( analogous to vaccinia virus protecting against smallpox ) , and MV and recombinants thereof have therefore been postulated as LASV vaccines [40] , [41] . MV infection of HHD mice did not result in detectable viremia ( Fig . 3A ) nor was AST elevation recorded at any time point ( Fig . 3B ) . This pattern of susceptibility of HHD mice to LASV but not MV reflected the one reported in non-human primates [40] . Next we tested whether MV immunization could induce HLA-A2 . 1-restricted immunity against LF . A recent study has characterized HLA-A2 . 1-restricted T cell epitopes in the glycoprotein ( GP ) of LASV [42] . Here we found that MV infection of HHD mice elicited high frequencies of CD8+ T cells specific for the GP42-50 epitope of LASV and a somewhat weaker but clearly detectable response against the GP60-68 epitope ( Fig . 3C ) . CD8+ T cells specific for a third known epitope in LASV-GP ( GP441-449 ) were not induced to a detectable extent , owing to only partial sequence homology of MV and LASV . When subsequently challenged with LASV , MV immunization prevented viremia and serum AST elevation in HHD mice ( Fig . 3D , E ) , and by day 14 after challenge the spleen , liver , lung and kidney of MV-immunized mice were free of detectable LASV ( data not shown ) . MV and LASV are serologically distinct i . e . neutralizing antibodies elicited against one virus do not crossreact nor crossprotect against the other [43] . This suggested that T cell immunity protected MV-immunized HHD mice against subsequent LASV challenge ( Fig . 3D , E ) , albeit primary T cell responses facilitated apparently the disease process in unvaccinated animals ( Fig . 2D ) . Next we set out to characterize tissue alterations in LASV-infected HHD mice and to study their dependence on T cells . In all LASV-infected HHD mice analyzed , the lung showed severe pneumonitis with interlobular septal thickening and collapse of the alveolar lumen ( Fig . 4A , B ) . In addition , macroscopic analysis at necropsy or in terminally diseased animals often revealed a substantial pleural effusion ( up to about 0 . 5 ml in each hemithorax , not shown ) . Both observations matched those in human LF [30] , [31] . In contrast to HHD mice , the lungs of C57BL/6 sacrificed at the same time point were only mildly affected or appeared normal ( Fig . 4A , B ) . CD8/CD4-depletion prevented these alterations in HHD mice , and also MHC-I-/- mice exhibited considerably milder signs of peumonitis . Interestingly , HHD lungs contained dense infiltrations of rounded Iba-I+ monocyte/macrophage lineage cells ( Fig . 4C ) , a finding that was less prominent or absent in C57BL/6 mice , T cell-depleted HHD mice or MHC-I-/- mice . Accumulation of T cells was also noted in HHD lungs ( Fig . 4D ) albeit to a lesser extent than for monocytes/macrophages ( compare Fig . 4C ) . Moreover , similar infiltrations were also found in resistant C57BL/6 wild type mice and thus did not correlate with disease . In the liver , nodules of mononuclear cells were found around the portal fields of all four groups of mice ( Fig . 4E ) . Striking differences were , however , noted in the distribution , shape and orientation of hepatic monocyte/macrophage populations ( including Kupffer cells , Fig . 4F ) . Like in uninfected mice , hepatic monocytes/macrophages of C57BL/6 and of T cell-depleted HHD mice formed predominantly a flat layer along liver sinusoids , oriented towards the central vein in a stellar pattern . In contrast , the architecture of this cell layer was disrupted in HHD mice ( unless depleted of T cells ) with the remaining cells enlarged , rounded up , disorganized and often accumulated in clusters , indicative of cellular activation and reminiscent of the vigorous hepatic macrophage response reported from human LF [29] , [30] , [31] , [32] , [33] . MHC-I-/- mice displayed an intermediate picture with only moderate monocyte/macrophage activation . Conversely , T cells were scattered at similarly moderate density throughout the liver parenchyma and in periportal inflammatory nodules of HHD and C57BL/6 mice ( Fig . 4G and data not shown ) . The number of hepatic T cells did therefore not correlate with disease , similar to the findings in the lung . Generalized immunosuppression is widely assumed to accentuate viral hemorrhagic fever [6] . A recent monkey study has tentatively attributed LASV immunosuppression to disorganization of the microarchitecture in secondary lymphoid organs [44] . Here we found that LASV infection of HHD mice resulted in disruption of the splenic white and red pulp compartments , whereas the spleens of C57BL/6 , CD8/CD4-depleted HHD mice and MHC-I-/- mice were less affected ( Fig . 4H ) . In correlation with these alterations , the marginal zone macrophage layer was lost in HHD mice but not in the other groups , and monocytes/macrophages were homogenously distributed throughout the splenic tissue of HHD mice ( Fig . 4I ) . T cells were largely absent in CD8/CD4-depleted mice , as expected , and were also somewhat scarce in HHD mice ( Fig . 4J ) , possibly indicating LASV-induced T cell depletion as reported from non-human primates [44] . The above morphological alterations had suggested T cell-dependent monocyte/macrophage activation in LASV-infected HHD mice . Classical activation of monocytes/macrophages e . g . by the T cell cytokine interferon gamma and cell-to-cell contact [45] , [46] , [47] , [48] triggers the secretion of NO and inflammatory cytokines , and expression of the former is mediated by inducible NO synthase ( iNOS ) [45] , [46] , [47] , [48] , [49] . iNOS expression can therefore serve as a histological marker for inflammatory differentiation of monocytes/macrophages [50] . On day 8 after LASV infection we detected numerous iNOS-expressing monocyte/macrophage clusters in the liver parenchyma of HHD mice ( Fig . 5A ) . Conversely , iNOS expression was not found in the liver of C57BL/6 mice , CD8/CD4-depleted HHD mice or MHC-I-/- mice infected with LASV . Furthermore , neither HHD nor C57BL/6 mice displayed hepatic iNOS expression or morphological evidence of monocyte/macrophage activation when assessed on day 2 and day 4 after infection ( Fig . S1 , analogous data in lung and spleen not shown ) , i . e . prior to the onset of the adaptive immune response . Synthesis of the inflammatory cytokine subunit IL-12p40 is restricted to macrophages , monocytes and dendritic cells , and its production is greatly enhanced by T cell stimulation [48] , [51] . Within eight days after LASV infection , susceptible HHD mice displayed vastly elevated serum IL12-p40 levels ( Fig . 5B ) . Resistant C57BL/6 mice showed comparably moderate IL-12p40 elevation . In agreement with the above results on iNOS , IL-12p40 secretion was strongly reduced by CD4/CD8-depletion in HHD mice , and MHC-I-/- mice exhibited an intermediate IL-12p40 response . Together , these findings suggested that T cells triggered an inflammatory differentiation of monocytes/macrophages with subsequent production of NO , IL-12p40 and likely other inflammatory mediators ( see Discussion section ) . This process occurred , however , solely under conditions of unchecked LASV replication i . e . in HHD mice but not in C57BL/6 mice where the infection and thus the antigen were rapidly cleared . We had noted high virus loads in lung , liver and spleen of HHD mice ( Fig . 1C ) where the above pathological changes were found . The cellular distribution of LASV in vivo remains unknown , albeit the virus has been shown to replicate productively in cultured primate macrophages and dendritic cells [21] , [52] , [53] . To better understand how viral replication might be associated with disease we assessed virus distribution in tissues by immunohistochemistry on day eight after infection ( Fig . 5C ) . LASV nucleoprotein ( NP ) was readily detected in a distinct population of cells in lung , liver and spleen . Morphological criteria suggested that these cells were predominantly monocytes/macrophages . We therefore performed immunofluorescence double-stains for LASV nucleoprotein ( LASV-NP ) and the monocyte/macrophage marker Iba-I ( Fig . 5D ) . By this method , 82±6 . 8% of LASV-NP+ cells in liver , 65±8 . 5% in the lung and 79±11% in the spleen ( mean±SD of six mice ) could be identified as monocytes/macrophages , suggesting that they served as a major target of LASV .
The data in mice presented here suggest a dual role for T cells during LASV infection: T cells appear essential for rapid clearance of the virus , but if failing to do so they may play a key role in the ensuing disease process , too . Involvement of T cells in the pathogenesis of viral hemorrhagic fever has precedence in Dengue virus infection [10] , [11] , [39] . Unlike dengue hemorrhagic fever ( DHF ) , where “original antigenic sin” of memory T cells appears to be involved , our findings suggest that T cell responses can have disease-enhancing effects during primary LASV infection . Importantly , we do not exclude roles for memory T cells in addition , but such aspects remain to be investigated ( see below ) . Our experiments modeled three prototypic scenarios of LASV-host balance , as defined by the parameters “T cells” ( relative efficacy of antiviral T cell responses ) and “Virus” ( persistent virus load ) , thus delineating the extremes of a spectrum and resulting in the following outcomes: Several mechanisms can be envisaged by which T cell responses enhance LF [10] , [11] , [39] but additional studies will be needed to address them directly . The histological picture in HHD mice supports the current view [7] that direct T cell-mediated cytolysis [54] unlikely is the main mechanism responsible for the tissue damage in LF , such as hepatocyte death with subsequent AST release . Based on the available evidence we postulate that during persisting viremia , T cells continuously encounter LASV epitopes on infected monocytes/macrophages in MHC-I and MHC-II context ( Fig . 5E ) . Additional interaction via co-stimulatory molecules , or stimulation via T cell cytokines may trigger infected monocytes/macrophages to differentiate and subsequently secrete inflammatory mediators of their own [45] , [48] , [49] , [51] . Overstimulation of macrophages can result in severe hepatic and pulmonary damage besides mediating a shock syndrome [26] , [55] , [56] , and such overstimulation provides a plausible mechanism for indirect T cell involvement in LF pathogenesis . LASV and related viruses are known to replicate in cultured macrophages without causing cellular activation or production of inflammatory cytokines [21] , [52] , [53] . Hence , classical T cell-driven monocyte/macrophage activation by IFNγ and direct cell-to-cell contact [45] , [46] , [47] , [48] may augment inflammatory differentiation and cytokine release from LASV-infected monocytes/macrophages in vivo [19] , [21] , [22] , similar to the ability of LPS to induce the activation of LASV-infected macrophages in culture [21] , [52] , [53] . T cell stimulation may thus facilitate a systemic inflammatory condition [57] as a potential pathogenetic correlate of the diverse and non-specific clinical manifestations of LF . This view of the role of T cells in LF correlates well with our observations in scenario I , where one would predict that efficient virus elimination by C57BL/6 and immune HHD mice results in only short and transient antigen presentation on a limited number of infected monocytes/macrophages , and therewith lack of disease , as was indeed found . Similarly , the concept explains our findings in scenario III , where the absence of T cells to properly activate infected monocytes/macrophages in CD8/CD4-depleted HHD mice results in mild or absent disease . In addition , at least two – seemingly paradoxical – observations in LASV-infected non-human primates would support our mechanistic postulate . The first observation is that high doses of LASV tend to be less lethal than low ones [28] . This phenomenon may be explained through the mechanisms of T cell “exhaustion” or “deletion” under high virus loads [58] . As such , a high initial virus inoculum may weaken the T cell response , thus attenuating disease through shifting conditions from scenario II towards scenario III . The second paradoxical observation stems from a vaccination study , in which a recombinant vaccinia virus expressing the NP protein of LASV was used in monkeys . The vaccine turned out to protect only a minority of the animals and , intriguingly , those LASV-challenged monkeys not protected by the vaccine displayed a more acute form of disease than control monkeys that had not been vaccinated at all [34] . A likely explanation may be that , although the vaccination may not have protected all animals , the still accelerated ( memory ) T cell response of non-protected animals also accelerated their disease process . Support for such a scenario comes from infection of mice with another arenavirus , lymphocytic choriomeningitis virus ( LCMV ) . LCMV-induced immunopathological disease of the central nervous system is T cell-dependent and , similar to the observation with LASV in monkeys , can be enhanced by prior vaccination with recombinant vaccinia viruses expressing LCMV antigens [59] . Last but not least , our study introduces a mouse model for LF , the lack of which has long hampered progress in this field of research . Only the general versatility of the mouse as a research model , including the availability of gene-targeted strains , makes mechanistic studies as presented here possible . Despite certain shortcomings as listed below , we think that the humanized mouse model could prove useful in further studies on LF pathogenesis , especially as the model lends itself well to the assessment of CD8+ T cell-based vaccines ( compare Fig . 3 ) . Although the ∼20% lethality we found in HHD mice using the Ba366 strain of LASV may contrast with the uniform lethality observed in LASV strain Josiah-infected strain 13 guinea pigs or monkeys [28] , our lethality rates match those reported for Josiah-infected outbred Hartley guinea pigs [60] , as well as inbred strain 13 and strain 2 guinea pigs inoculated with other LASV isolates [61] . Nevertheless , we cannot exclude the possibility that the HHD model fails to recreate certain aspects of human LF . For the mechanistic analyses presented here , we used relatively high intravenous LASV doses of 106 PFU . The need for such doses to cause severe disease likely reflects the imperfect adaptation of LASV to mice ( compare also the Methods section ) . However , preliminary experiments indicate that viremia and AST elevation can already be observed at lower doses , albeit with higher variability , and that LASV can replicate in HHD mice after subcutaneous administration ( data not shown ) . We would therefore argue that viremia and AST elevation in HHD mice may represent useful surrogates [38] to assess vaccine efficacy prior to an eventual confirmation in non-human primates . T cell densities in the spleen of LASV-infected HHD and C57BL/6 mice were comparable ( day 8: 3339±925 CD3+ cells/mm2 in HHD mice , 3826±2057 CD3+ cells/mm2 in C57BL/6 mice; mean±SD , n = 6; p = 0 . 61 ) , arguing against preferential T cell depletion [44] as a potential reason for defective virus control in HHD mice . To characterize disease enhancing ( HHD ) vs . protective ( C57BL/6 ) primary T cell responses against LASV , future studies will need to compare the magnitude , kinetics and effector-/cytokine-profile [62] of LASV epitope-specific CD4+ and CD8+ T cell responses . Similarly , serum IL-12p40 and iNOS expression have served as surrogates of the inflammatory monocyte/macrophage response in this study , but future work will have to determine its breadth in terms of cytokines , chemokines and inflammatory mediators such as NO , leukotrienes and prostaglandins , their production in tissues and systemic dissemination in blood . The experimental depletion and/or inhibition of monocyte/macrophage populations and of inflammatory mediators may provide additional insights into the cellular and molecular players in LF , indicating to which extent the “cytokine storm” hypothesis [57] can explain LF pathogenesis . With the availability of a mouse model for LASV such studies become possible , although the necessity for BSL-4 laboratory containment can represent a major practical hurdle . Taken together , our results in mice suggest a two faced role of T cells in LASV infection , both in virus control and also in enhancing LF pathogenesis . This extends our understanding of LASV-host interactions and raises the possibility that heterogeneity in MHC-I and in overall T cell immunocompetence represents one explanation for the wide spectrum of clinical outcomes in a human population exposed to LASV [2] . Perhaps even more important , we think that beside beneficial also detrimental aspects of T cell-responses and -immunity [34] , [59] should be given thorough consideration in future strategies for LF vaccine design .
C57BL/6 , β2-microglobulin-deficient mice ( MHC-I-/- ) [63] , MHC-II-/- [64] and HHD [37] mice were bred at the Institute for Laboratory Animal Sciences , University of Zurich , Switzerland . Experiments with Lassa virus were performed in the BSL-4 unit of the Bernhard Nocht Institute , Hamburg , Germany . Experiments with Mopeia virus were performed at the University of Geneva and at the University Hospital of Zurich , Switzerland . Permission for animal experiments was obtained from the authorities of the Freie und Hansestadt Hamburg , and from the Cantonal authorities of Geneva and Zurich , Switzerland , respectively . All experiments were performed in accordance with the Swiss and German law for animal protection , respectively . This model is based on the Ba366 [65] strain of LASV . Pilot experiments with a range of isolates representing the different endemic areas ( Josiah , Sierra Leone; Lib90 , Liberia; Ba366 , Guinea; AV , Ivory coast/Burkina Faso; CSF , Nigeria ) had indicated that Ba366 was the virus that most efficiently replicated in HHD mice . LASV and Mopeia virus ( AN21366 ) , were grown on BHK21 and Vero cells , respectively , and were administered to mice at a dose of 106 PFU i . v . unless stated differently . Virus stocks and viral infectivity in blood samples were determined in immunofocus assays as described [66] . CD8+ and/or CD4+ T cell populations were depleted by i . p . administration of monoclonal antibodies YTS169 ( anti-CD8 ) and YTS191 ( anti-CD4 ) on day -3 and day -1 of LASV infection as previously described [67] . The efficiency of depletion was verified by flow cytometry and was >99% . Serum AST and ALT activities were determined by using commercially available colorimetric assay kits ( Reflotron , Roche Diagnostics , Germany ) . Mouse tissues were fixed in 4% formalin and were embedded in paraffin . Sections were stained with hematoxilin/eosin ( H/E ) or processed for immunohistochemistry as follows: Upon inactivation of endogenous peroxidases ( PBS/3% hydrogen peroxide ) and blocking ( PBS/10% FCS ) sections were incubated with the primary antibodies rat anti-human CD3 ( crossreactive with murine CD3 on mouse T cells; Serotec ) , rabbit anti-Iba-1 ( monocytes/macrophages , Wako Pure Chemical Industries ) or rat anti-Lassa nucleoprotein ( see below ) . Bound primary antibodies were detected with biotinylated rat-specific ( DakoCytomation ) or rabbit-specific ( Amersham ) secondary antibody , followed by incubation with extraavidin peroxidase ( Sigma Aldrich ) , and bound peroxidase was visualized by 3 , 3′-diaminobenzidine as chromogen ( Sigma Aldrich ) . Haemalaun was used for counterstaining of nuclei . For fluorescence double labeling , primary antibodies were visualized using species specific Cy3- or Cy2-conjugated secondary antibodies ( all from Jackson ImmunoResearch Laboratories Inc . ) with DAPI ( Sigma-Aldrich ) nuclear staining . To determine the percentage of monocytes/macrophages ( Iba-1-positive cells ) amongst LASV-infected cells , a total of 41 ( liver ) and 27 ( lung ) randomly captured 40x visual fields were analyzed . Histological spleen sections stained with anti-CD3 antibody ( T cells ) and counterstained with Haemalaun ( nuclei , see above ) were scanned using the Dotslide System ( Olympus GmBH ) at a 200-fold magnification . For analysis , the images were automatically processed in a custom-programmed script of Cognition Network Language based on the Definiens Cognition Network Technology platform ( Definiens Developer XD software ) . The Cognition Network Language is an object-based procedural computer language , designed for automated analysis of complex , context-dependent image analysis . In brief , the programmed script first discriminates spleen tissue and tissue-free surroundings by spectral difference detection . The surface of the resulting region of interest ( spleen tissue , ROI ) is calculated . Subsequently “CD3 positive cells” within the ROI are detected based on brown anti-CD3 staining and are counted , to calculate the number of CD3+ cells per mm2 of tissue . Serum IL-12p40 was determined using a sandwich ELISA kit ( eBioscience ) according to the manufacturer's instructions . Recombinant NP of LASV strain BA366 was expressed in E . coli using the pET28 expression vector system ( Novagen ) . Supernatants of pET28 constructs were purified using the Talon Metal Affinity Resin ( Clontech ) in a batch procedure . Urea ( 8 M ) lysates were brought to nondenaturing conditions by increasingly substituting the buffer for sonication buffer during the resin-batch procedure . Proteins were eluted with 250 mM imidazole in sonication buffer on a gravity column ( Bio-Rad ) . Rat antisera were raised against purified recombinant NP by s . c . immunization with recombinant NP emulsified in complete Freund's adjuvant containing 1 mg of Mycobacterium tuberculosis ( H37RA; Difco Laboratories , Detroit , MI ) . Four weeks after the first immunization , animals were boosted with recombinant NP emulsified in incomplete Freund's adjuvant ( Difco ) . Terminal bleedings were performed 4 weeks after the boost . The specificity of the anti-LASV-NP antiserum was verified by immunofluorescence tests on LASV-infected cells as well as on LASV-infected or non-infected tissues . Pre-immune serum from the rats used for immunization was included as a control in both settings . Epitope-specific CD8+ T cells were enumerated by an intracellular cytokine assay for IFNγ as previously described [68] . In brief , 106 splenocytes were incubated in 200 µl of IMDM supplemented with 10% FCS and penicillin/streptomycin for 5 h at 37°C at a 10−6 M concentration of the LASV-GP-derived peptide epitope GP42-50 ( GLVGLVTFL ) , GP60-68 ( SLYKGVYEL ) , GP441-449 ( YLISIFLHL ) or with medium alone as a negative control . To enhance intracellular accumulation of IFN-γ , brefeldin A was added at a final concentration of 5 µg/ml for the last 3 . 5 hours of culture . Subsequently , the cells were washed with FACS buffer ( PBS supplemented with 2% FCS , 0 . 01% NaN3 and 20 mM EDTA ) and surface staining was performed with anti-CD8β-PE and anti-B220-PerCP antibody conjugates ( both from BD Biosciences ) for 30 min at 4°C . After washing twice with FACS buffer , the cells were fixed with 100 µl of 4% paraformaldehyde in PBS for 5 min at 4°C . Two milliliters of permeabilization buffer ( FACS buffer supplemented with 0 . 1% w/v saponin , Sigma ) were added and the cells were incubated for 10 min at 4°C . Subsequently , they were spun down and stained intracellularly with anti-mouse-IFNγ-APC ( BD Biosciences ) in permeabilisation buffer for 60 min at 4°C . After two washes with permeabilization buffer , the cells were resuspended in FACS buffer and were analyzed on a FacsCalibur ( Becton Dickinson ) . FACS plots were gated on B220− lymphocytes . Between group differences were analyzed by 1-way ANOVA and 2-way ANOVA for individual or multiple values of different groups , respectively , followed by LSD post tests . SPSS vs . 13 was used for analysis . P values <0 . 05 were considered statistically significant ( indicated as * in figures ) . P<0 . 01 was considered highly significant ( indicated as ** in figures ) . P>0 . 05 was considered as not significantly different ( “n . s . ” ) .
|
Lassa virus ( LASV ) is the causative agent of Lassa fever ( LF ) , accounting for substantial morbidity and mortality in West Africa . Yet the mechanisms leading to disease remain poorly understood . Here we propose a concept whereby the body's immune defense either defeats LASV rapidly or , if unsuccessful , becomes an essential facilitator of disease . This latter paradoxical postulate stems from observations in genetically engineered ( HHD ) mice , which we found to be susceptible to LF . HHD mice differ from resistant wild type mice in that they have a humanized repertoire of T cells , a main component of the mammalian immune system . Counterintuitively , we could protect HHD mice against LF by experimentally removing their T cells . We further found that LF correlated with widespread activation of macrophages , which again depended on T cells . Similar to T cells , macrophages are important players in our body's defense system , but their inflammatory products are also candidate mediators of LF . Taken together , these findings suggest that LF may represent an inappropriate host response to infection . Specifically , our study demonstrates a two-faced role of T cell responses against LASV . Such detrimental aspects of immune defense need to be given consideration in future LF vaccine development , to avoid enhancement of disease in vaccinated individuals .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"infectious",
"diseases/viral",
"infections",
"immunology/immunity",
"to",
"infections",
"immunology/immune",
"response",
"immunology/innate",
"immunity"
] |
2010
|
T Cell-Dependence of Lassa Fever Pathogenesis
|
TLR2 is a cell surface receptor which elicits an immediate response to a wide repertoire of bacteria and viruses . Its response is usually thought to be proinflammatory rather than an antiviral . In monocytic cells TLR2 cooperates with coreceptors , e . g . CD14 , CD36 and αMβ2-integrin . In an earlier work we showed that αvβ3-integrin acts in concert with TLR2 to elicit an innate response to HSV , and to lipopolysaccharide . This response is characterized by production of IFN-α and -β , a specific set of cytokines , and NF-κB activation . We investigated the basis of the cooperation between αvβ3-integrin and TLR2 . We report that β3-integrin participates by signaling through Y residues located in the C-tail , known to be involved in signaling activity . αvβ3-integrin boosts the MYD88-dependent TLR2 signaling and IRAK4 phosphorylation in 293T and in epithelial , keratinocytic and neuronal cell lines . The replication of ICP0minus HSV is greatly enhanced by DN versions of MYD88 , of Akt – a hub of this pathway , or by β3integrin-silencing . αvβ3-integrin enables the recruitment of TLR2 , MAL , MYD88 at lipid rafts , the platforms from where the signaling starts . The PAMP of the HSV-induced innate response is the gH/gL virion glycoprotein , which interacts with αvβ3-integrin and TLR2 independently one of the other , and cross-links the two receptors . Given the preferential distribution of αvβ3-integrin to epithelial cells , we propose that αvβ3-integrin serves as coreceptor of TLR2 in these cells . The results open the possibility that TLR2 makes use of coreceptors in a variety of cells to broaden its spectrum of activity and tissue specificity .
The toll like receptors ( TLRs ) constitute a major defensive system of the cell against invasion from bacteria and viruses , and endogenous DAMPs ( danger associated molecular patterns ) [1] . Some of them , including TLR2 and 4 , are present at the cell surface and mount the immediate branch of the innate response , before the invading microorganism or its components are internalized into the cell , and before the cytoplasmic sensors come into play . Initially described as an antibacterial sentinel [2] , TLR2 emerged also as an antiviral sentinel [3] , [4] , and , indeed , it is characterized by the wide spectrum of bacteria , viruses and DAMPs which it senses [5] . The general view is that TLR2 favours a proinflammatory response . In the past few years αvβ3-integrin and TLR2 were shown to act in concert to elicit a response to lipopeptide , to lipopolysaccharide ( LPS ) and to herpes simplex virus ( HSV ) , a large DNA virus [6]–[9] . Specifically , our laboratory reported that in cells positive for both αvβ3-integrin and TLR2 , IFN ( interferon ) α and β , and a specific set of cytokines - IL ( interleukin ) 2 and IL10 - were highly upregulated , and NF-κB was activated in response to HSV infection or exposure to a commercial source of LPS . By contrast , in cells negative for TLR2 , the IFN-α and -β production and the NF-κB response were very low . In loss of function experiments , the silencing of β3-integrin in TLR2-positive cells dramatically reduced the IFN-α and-β production and the NF-κB response . The β3-integrin-silencing in TLR2-negative cells practically abolished the IFN and NF-κB response [7]–[9] . Importantly , the activation of IFN-α and -β and of NF-κB was detected not only in the model 293T cells , but also in epithelial , keratinocytic and neuronal cell lines , i . e . in cells which are models of the cells targeted by HSV in vivo [7] . These findings highlight that the branch of the innate response dependent on the αvβ3-integrin and TLR2 axis makes a significant contribution to the overall IFN production in HSV-infected cells . They contrast with the prevalent view that the IFN antiviral effect is triggered mainly by the endosomal and the cytoplasmic sensors [10]–[13] . The herpes simplex virion component which elicits this immediate response is the envelope glycoprotein gH , which forms a heterodimer with gL [9] . gH/gL are part of the fusion machinery in HSV required for HSV entry into the cell [14] , [15] . They interact physically with αvβ3-integrin at low affinity [8] . They also interact with TLR2 by coimmunoprecipitation in αvβ3-integrin positive cells [9] . In turn , αvβ3-integrin and TLR2 interact with each other in a ligand-independent manner , i . e . in resting conditions [6] , [8] . HSV evades this immediate response as soon as it enters the cell , at the onset of viral protein synthesis , by aid of the immediate early protein ICP0 ( infected cell protein 0 ) [8] , [16] . Importantly , the same effects were elicited by HSV and by a commercial preparation of LPS , hence the repertoire of microbial components capable to elicit this branch of the innate response is broad [7] , [8] . The objective of this work was to shed light on the mechanisms by which αvβ3-integrin and TLR2 act in concert to elicit this branch of the innate response to HSV and LPS . At large , two scenarios were envisioned . In the first , the two receptors – αvβ3-integrin and TLR2 – increase the efficiency of virus entry into cell or of LPS binding-uptake into the cell . The alternative scenario envisions that , since each of the two receptors exhibits intrinsic signaling activity , the concerted response results from boosting of one or the other signaling activity . Against the first possibility argues the finding that the extent of HSV infection was essentially similar in wt- and in β3-integrin-silenced cells [8] . To preliminarily verify the second possibility we examined if αvβ3-integrin participates in this response through critical Y residues of its cytoplasmic tail , known to be involved in signaling . Subsequently , we examined the intermediates in the pathways downstream of TLR2 and of αvβ3-integrin and found that αvβ3-integrin boosts the recruitment of MYD88 ( myeloid differentiation primary response gene 88 ) to TLR2 and the phosphorylation of IRAK4 ( Interleukin-1 receptor-associated kinase ) . The reverse effect , boosting of αvβ3-integrin signaling by TLR2 , was not detected . Thus , the concerted αvβ3-integrin and TLR2 response rests on augmentation by αvβ3-integrin of the MYD88-dependent TLR2 signaling .
As a first line of evidence that the response to HSV , or to LPS , mediated in concert by αvβ3-integrin and TLR2 results from a signaling activity , we asked whether integrin participates through its cytoplasmic tail , namely through residues Y747 and Y759 known to undergo phosphorylation upon integrin activation . To avoid the interference from endogenous β3-integrin , we employed 293T cells in which β3-integrin was stably silenced ( named 293T sh-β3 cells ) [8] , and transfected them with plasmid encoding the wt- or mutant forms of β3-integrin carrying the single Y747F , or Y759F substitution , or both ( β3-integrinY747 , β3-integrinY759 , β3-integrinY747-Y759 ) [17] . 293T cells fail to express TLR2 , hence cells were transfected , or not , with TLR2 , as indicated . Comparison was with non-silenced 293T cells , expressing or not TLR2 , plus the indicated β3-integrin isoform . To measure the NF-κB response , cells were additionally transfected with a plasmid encoding firefly luciferase under NF-κB regulated promoter ( pNF-κB-luc ) . Renilla luciferase was included to account for variations in transfection efficiency . For virus-induced stimulation , we made use of the HSV mutant named R7910 , which carries the deletion of the gene encoding the immediate early protein ICP0 [18] . The immediate innate response to HSV is counteracted by this viral protein , and is therefore best detected in its absence [8] . Previously , we ascertained that HSV-1 virions entered β3-integrin-positive or -negative cells with similar efficiency . This result ruled out that the lower innate response seen in sh-β3 cells could be attributed to lower amount of virus infecting the cells [8] . The results in Fig . 1 A , right panel , shows that , in the presence of TLR2 , R7910 elicited NF-κB activation in sh-β3 cells expressing wt β3-integrin , but not any of the integrin mutants . For comparison , Fig . 1A , left panel , shows that R7910 elicited NF-κB activation in non silenced cells overexpressing wt β3-integrin , in agreement with previous data [8] . The activation was reduced by about 60% in cells overexpressing the integrin mutants ( note the different scale in Fig . 1A , right and left panels ) . As reported previously , the NF-κB activation is high in the presence but not in the absence of TLR2 , reflecting the αvβ3-integrin–TLR2 concerted action . In the next series of experiments , and in some of the subsequent experiments , we made use of gD−/− virions rather than of R7910 . gD is one of the essential glycoprotein for HSV entry into the cell [14] , [15] . gD−/− virions carry the deletion of the gD gene , lack gD in the virion envelope , are competent for attachment to but not for entry into cells . Because they elicit an innate response - albeit at levels somewhat lower than the gD-containing virions - they enabled us to rule out that any effect on the innate response seen with the β3-integrin mutant might be attributable to a reduction in virus entry . The response to a commercial LPS preparation able to elicit a TLR2 response [8] was assayed in parallel . Fig . 1 B shows that the substitution of wt-β3-integrin with the β3-integrinY747-Y759 grossly reduced the NF-κB activation elicited by gD−/− virions , or LPS; this occurred both in sh-β3 ( right panel ) and in wt cells ( left panel ) . Changes to endogenous NF-κB were measured through the degradation of IκB-α , an cytoplasmic inhibitory component of NF-κB pathway , which undergoes degradation when NF-κB is activated and then translocated to the nucleus . Fig . 1 C shows that in sh-β3 cells which express the β3-integrinY747-Y759 mutant and TLR2 , and do not enable NF-κB activation , IκB-α was not degraded . By contrast , in sh-β3 which express the wt β3-integrin and TLR2 , IκB-α was completely degraded 2 h after exposure of cells to R7910 . Previous studies showed that the production of IFN-α and -β , and of the specific cytokines followed the same pattern as NF-κB activation [7] , [8] . We ascertained whether the mutant form β3-integrinY747-Y759 hampered IFN-β response . sh-β3 or wt 293 cells , positive or negative for TLR2 , expressing or wt or mutant β3-integrin were infected with R7910 . The culture medium was harvested at 48 h after infection , and the secretion of IFN-β was measured by ELISA . Fig . 1 D shows that IFN-β was secreted only by cells expressing wt , but not the mutant β3-integrinY747-Y759 , in agreement with the NF-κB response . Gerold et al and our laboratory [6] , [8] showed that αvβ3-integrin and TLR2 interact in a ligand-independent manner , as seen by co-immunoprecipitation . We verified that the β3-integrinY747-Y759 mutant maintains the ability to interact with TLR2 . TLR2-Flag was immunoprecipitated from β3-integrin–silenced cells , transfected with wt-β3-integrin or β3-integrinY747-Y759 , plus TLR2-Flag . Fig . 1 E shows that β3-integrin was coimmunoprecipitated by TLR2-Flag , irrespective of mutations in the cytoplasmic tail ( compare lanes 3 and 4 ) . The results indicate that the innate response to HSV , or to LPS , dependent on the concerted αvβ3-integrin–TLR2 action is hampered when a mutant form of β3-integrin defective in phosphorylation replaces wt β3-integrin . Cumulatively , they demonstrate that the immediate innate response triggered by the integrin–TLR2 axis results from a signaling activity . To shed light on the mechanism by which αvβ3-integrin and TLR2 act in concert , we asked whether integrin boosts the TLR2 signaling response , or viceversa , whether TLR2 boosts the αvβ3-integrin signaling . To address the first question , we analyzed the typical intermediates downstream of TLR2 and asked whether their recruitment/activation was higher in integrin-positive ( wt ) than in β3-integrin-silenced cells . TLR2 signals through the recruitment of MYD88 , followed by phosphorylation of IRAK 1 and 4 [19] . TLR2-Flag and hemagglutin ( HA ) -tagged MYD88 ( MYD88-HA ) were expressed in 293T or sh-β3 cells . Cells were exposed to R7910 ( 60 PFU/cell ) or to LPS ( 300 ng/ml ) for 10 , 20 , 30 min . Fig . 2 A shows that the amount of MYD88-HA coimmunoprecipitated by TLR2-Flag was dramatically decreased in sh-β3 cells ( right panels ) , as compared to non-silenced cells ( left panels ) . The recruitment of MYD88 to TLR2 in response to LPS was also dramatically decreased in sh-β3 cells ( Fig . 2 A ) . The defective recruitment of MYD88 to TLR2 upon silencing of β3-integrin was seen also in cells other than 293T cells . We selected the keratinocytic HaCaT , the epithelial HeLa and the neuronal SK-N-SH cell lines , which are models of the cells targeted by HSV in vivo . Previous studies showed that silencing of β3-integrin in the above cell lines dramatically reduced both IFN-β production and NF-κB activation [7] . The indicated cells were transfected as indicated for panel A . Fig . 2 B–D show that also in HaCaT , HeLa and SK-N-SH cells , the silencing of β3-integrin resulted in strong impairment of MYD88 coimmunoprecipitated by TLR2 , upon exposure of cells to R7910 . Next , we investigated whether the expression of the β3-integrinY747-Y759 mutant , defective in signaling , and defective in the HSV- or LPS-induced activation of NF-κB , resulted in a reduced recruitment of MYD88 to TLR2 . sh-β3 cells , transfected with wt- or β3-integrinY747-Y759 were additionally transfected with TLR2-Flag and MYD88-HA . Fig . 2 E shows that the MYD88 recruitment to TLR2 was strongly reduced in cells expressing the β3-integrin mutant ( right panel ) . We then asked whether MYD88 can be recruited to a complex with integrin in the absence of TLR2 , and whether this complex formation is hampered when the wt β3-integrin is substituted with the β3-integrinY747-Y759 mutant . Sh-β3 cells were transfected with a form of αv-integrin carrying the double-strep epitope ( named αvstrep or αvst ) , wt-β3-integrin or β3-integrinY747-Y759 , plus or minus TLR2 . The transfected cells were exposed to R7910 for 10 , 20 , or 30 min . αvst-integrin was harvested by means of Strep-Tactin Sepharose . The co-precipitations in Fig . 2 F show that MYD88 was co-precipitated by αvstrepwt-β3-integrin ( left panel ) , but not by αvstrepβ3-integrinY747-Y759 mutant ( right panel ) , irrespective of the presence or absence of TLR2 . Whether this results from a direct MYD88–αvβ3-integrin interaction , or from an indirect interaction remains to be investigated . To provide evidence for a functional role of MYD88 in the αvβ3-integrin–TLR2 signaling pathway , we made use of a DN ( dominant negative ) version of MYD88 , named MYD88FW/AA [20] . MYD88FW/AA carries the indicated substitutions in α helix E ( αE ) . It can be recruited to TLR2 but is defective in downstream signaling [20] . 293T or sh-β3 cells , plus or minus TLR2 , were transfected with MYD88FW/AA or wt-MYD88 , plus NF-κB-luc and Renilla luciferase plasmids . The transfected cells were infected with R7910 for 6 h . Fig . 2 G shows that NF-κB activation was dramatically decreased in cells expressing the DN MYD88FW/AA . Downstream of MYD88 , the TLR2 signaling cascade involves the phosphorylation of IRAK1 , and of IRAK4 . We compared the extent of IRAK4 phosphorylation in 293T and in sh-β3 cells , transfected with MYD88 and TLR2 . Cells were infected with R7910 , or exposed to LPS . Endogenous IRAK4 was immunoprecipitated . The immunoblot for phospho-IRAK4 ( P-IRAK4 ) shows a dramatic decrease of the phosphorylated form in β3-integrin–silenced cells , while the total amount ( T-IRAK4 ) was not substantially modified ( Fig . 2 H ) . Cumulatively , the results show that the absence of αvβ3-integrin , or a C-tail mutant form of β3-integrin , greatly decreased the amount of MYD88 recruited to TLR2 in a number of cell lines which are models of cells targeted by HSV in vivo . The results further show that the amount of P-IRAK4 , and the HSV-induced NF-κB activation strongly depend on a signaling-competent form of MYD88 . We conclude that the αvβ3-integrin–TLR2 concerted action entails a strong boost of the MYD88-dependent TLR2 signaling by αvβ3-integrin . Having established that the αvβ3-integrin–TLR2 response entails an enhancement by αvβ3-integrin of the TLR2 signaling pathway , we asked the viceversa question , namely to what extent the αvβ3-integrin signaling branch contributes to the response . We focused on Src , because it belongs to a group of non-receptor tyrosine kinases typically activated by integrins . Furthermore , microarray analysis , validated by qRT-PCR , showed that some genes - Src , Syk , Card9 - were activated via αvβ3-integrin , in a TLR2-independent fashion , by HSV infection or exposure to LPS [8] . Of note , this group of genes was upregulated to a modest extent ( only 2–5 fold ) , whereas the IFN-α and -β genes were upregulated via the αvβ3-integrin–TLR2 axis about 100 fold . Here , we report that Src is phosphorylated very early - 10 minutes - following exposure of cells to HSV . The increase was much smaller in TLR2-positive cells than in TLR2-negative cells , hence it does not depend on TLR2 ( Fig . 3 A ) . Further , the HSV-induced Src phosphorylation did not ensue in β3-integrin–silenced cells ( Fig . 3 B ) , hence it depends on integrin alone , and not on the integrin–TLR2 cooperation . The functional role of Src was assessed through the effect of the PP1 ( 4-amino-5- ( methylphenyl ) -7- ( t-butyl ) pyrazolo- ( 3 , 4-d ) pyrimidine ) Src inhibitor on NF-κB activation induced by gD−/− virions . When TLR2 was present a very modest inhibition ( 20% ) was observed; in the absence of TLR2 , the inhibition 70% ( Fig . 3 C ) . Next , we investigated whether Src phosphorylation is dependent upon signaling by the C-tail of β3-integrin . sh-β3 cells expressing wt- β3 or β3-integrinY747-Y759 were exposed to R7910 for 10 or 20 minutes , and endogenous phospho-Src was analyzed . An increase in phosphorylation was clearly detected in cells expressing the wt-β3-integrin , but not in cells expressing the mutant β3-integrin ( Fig . 3 D ) , indicating that signaling by the β3-integrin C-tail targeted Src for phosphorylation . Together , the pattern of Src phosphorylation and the modest effect exerted by the Src inhibitor argue against an involvement of Src in the αvβ3-integrin–TLR2 concerted response , and , consequently , against an enhancement by TLR2 on the αvβ3-integrin signaling . The HSV-induced Src phosphorylation depends on αvβ3-integrin alone , in particular on signaling carried out by the C-tail . Akt constitutes a hub downstream of a number of signaling pathways . It participates in HSV-induced signaling cascades , both innate responses and downstream of FAK activation [21] , [22] . Inasmuch as Src is not a station in the αvβ3-integrin–TLR2 concerted signalling , we asked whether Akt participates in it . We first performed functional assays , and asked whether the MK2206 Akt inhibitor reduced the HSV-induced NF-κB activation and IFN-β production . Cells were exposed to the inhibitor for 1 h prior to infection and during virus absorption . In preliminary experiments we employed R7910 and observed that NF-κB activation was practically abolished by 5 µM MK2206 . Because , under these conditions , HSV infection was reduced to 50 and 40% in wt cells , and 5 and 30% in sh-β3 cells , and the reduction in NF-κB activation might reflect , in part , a reduction in virus entry , we made use of gD−/− virions . Fig . 4 A and B show that the NF-κB activation and the IFN-β production induced by gD−/− virions was almost abolished by MK2206 . We also tested the effect of a DN version of Akt , named PKB-CAAX , which carries a CAAX motif derived from Ki-Ras [23] . Wt- or sh-β3 293T cells were transfected with the wt version PKB , plus ten-fold excess of PKB-CAAX , or with wt-PKB and 10-fold excess of empty vector , plus or minus TLR2 . Cells were exposed to gD−/− virions . Fig . 4 A and B show a dramatic reduction in NF-κB activation and IFN-β production by PKB-CAAX both in the absence and in the presence of TLR2 . The reduction in activation of endogenous NF-κB by the above treatments was validated through measurements of IκB-α degradation , which paralleled the NF-κB activation seen in Fig . 4 , panels A and B ( Fig . 4 C , D ) . Next , we verified whether Akt undergoes phosphorylation , and whether the phosphorylation occurs in αvβ3-integrin/TLR2-dependent fashion . Fig . 5 A and B show the extent of phospho-Akt ( Ser473 ) ( P-Akt ) and total Akt ( T-Akt ) , as determined by western blotting ( WB ) , following exposure of wt-293T or sh-β3 cells to R7910 for 10 , 20 or 30 min . In these cells , Akt was overall modestly phosphorylated . Yet , the phosphorylation was consistently observed in numerous experiments . The HSV-induced increase occurred in the presence of TLR2 ( Fig . 5 A ) , was overall lower in sh-β3 cells ( Fig . 5 B ) . We next verified whether Akt phosphorylation was dependent on the signaling activity carried out by the C-tail of β3-integrin . The sh-β3 cells , transfected with wt-β3- integrin or C-tail integrin mutant were exposed R7910 for 10 , 20 , 30 min . Fig . 5 C shows that Akt phosphorylation was increased more than two-fold at 20 minute in cells expressing wt-β3-integrin; such increase was not seen in cells expressing the C-tail β3-integrin mutant . The pattern of Akt phosphorylation in 293T cells indicates that it is TLR2-dependent , and , in part , β3-integrin–dependent . This pattern of Akt phosphorylation - dependent on the αvβ3-integrin–TLR2 axis - was markedly different from that of Src phosphorylation . We checked the Akt involvement in the αvβ3-integrin/TLR2 signaling axis in cells other than 293T by analysis of Akt phosphorylation . The selected cells were those analysed in Fig . 2 for the αvβ3-integrin–enhanced MYD88 recruitment to TLR2 . HaCaT , HeLa , and SK-N-SH cells were silenced for β3-integrin , or non-silenced , and exposed to R7910 for 10 , 20 , 30 min . In all wt-cells , Akt underwent phosphorylation , or an increase in phosphorylation , upon exposure to R7910 ( Fig . 5 D–F ) . The increase was seen irrespectively of the differences in the basal level of Akt phosphorylation seen in unexposed cells . The virus-induced increase in Akt phosphorylation was not seen in the β3-integrin-silenced cells ( Fig . 5 D–F , right panels ) . Thus , in cells of different origin , the αvβ3-integrin/TLR2 signaling axis involves Akt as a downstream hub . HSV mutants deleted or mutated in ICP0 are strongly defective in replication , because they are defective in counteracting a number of the host defences . To shed light on the significance of the αvβ3-integrin/TLR2 response to the virus cycle , we measured R7910 replication in cells where the pathway was impaired through expression of DN mutants MYD88FW/AA or PKBCAAX , through silencing of β3-integrin , or both . Wt 293T or 293T sh-β3 cells were transfected with the wt or DN forms of the intermediates . Cells were infected with R7910 , and progeny virus titrated at 24 or 48 h after infection . Fig . 6 A shows that MYD88FW/AA or PKBCAAX rescued R7910 yield by about 2 Lg or more . The rescue was even higher in cells silenced for β3-integrin ( Fig . 6 B ) , in accordance with previous data [8] . Clearly , the αvβ3-integrin/TLR2 response is highly detrimental to the replication of the ICP0-minus HSV . A soluble form of gH/gL interacts physically with a soluble form of αvβ3-integrin at 10−6 M affinity [8] . We reported that in αvβ3-integrin–positive cells , TLR2 can co-immunoprecipitate gH/gL [9] . Here , we asked whether the observed gH/gL– TLR2 interaction was mediated by αvβ3-integrin , or was independent of it . We expressed gH/gL in αvβ3-integrin–positive or in sh-β3 cells , in the absence or presence of TLR2-Flag . Fig . 7 A shows that gH/gL was co-immunoprecipitated by TLR2 in β3-integrin-silenced cells ( lane 3 ) , as well as in wt 293T cells ( lane 4 ) . Hence , gH/gL interacts with TLR2 independently of αvβ3-integrin . We verified that gH/gL indeed cross-links αvβ3-integrin and TLR2-Flag . 293T cells were simultaneously transfected with TLR2-Flag , αv+β3-integrin , gH , gL . The controls were devoid of TLR2 . TLR2-Flag was immunoprecipitated . Fig . 7 B shows that TLR2-Flag coimmunoprecipitated both gH and β3-integrin , but not gD . Altogether , αvβ3-integrin and TLR2 interact one with the other under resting conditions , i . e . in the absence of ligands , [6] , [8] . In addition , gH/gL can interact with TLR2 and with αvβ3-integrin , independently one of the other , and indeed gH/gL recruit both of them to a complex , and cross-link them . To provide further evidence in support of gH/gL as the PAMP of the αvβ3-integrin/TLR2—mediated innate response , we checked whether gH−/− virions are defective in MYD88 recruitment to TLR2 , and in Src and Akt phosphorylation . gH−/− virions are deleted in gH [24] and were grown in non complementing cells . They can attach to cells but fail to infect them . The response elicited by these virions represents the immediate innate response to incoming virions , prior to their fusion with the target cells . Cells were exposed to gH−/− virions , and , for comparison , to R7910 . In cells exposed to gH−/− virions , TLR2 failed to recruit MYD88 ( Fig . 7 C ) ; Src phosphorylation was almost completely abolished ( Fig . 7 D ) ; the increase in Akt phosphorylation was moderately reduced in TLR2+293T as compared to that elicited by R7910 , in agreement with the modest Akt activation seen in these cells ( Fig . 7 E ) . Together , the previous finding that gH−/− virions fail to elicit NF-κB response [9] , the ability gH/gL and of αvβ3-integrin to interact with TLR2 , the defect of gH−/− virions to induce the MYD88 recruitment to TLR2 , and Akt phosphorylation argue for gH/gL as the HSV PAMP of the αvβ3-integrin/TLR2 system . CD14 is a costimulatory protein or co-receptor for a number of TLRs , including TLR2 [5] , [19] , [25] . It also enhances the HSV-induced TLR2 response [4] . Having established that the αvβ3-integrin–TLR2 concerted activity occurs through an enhancement by integrin of the TLR2 cascade , we asked whether CD14 plays a role in it . In particular , we asked whether CD14 augments the αvβ3-integrin–TLR2 concerted response , or whether CD14 can substitute for αvβ3-integrin in enhancing the TLR2 response . wt or β3-integrin-silenced 293T cells were transfected with CD14 , TLR2-Flag , or both , plus NF-κB-luc and Renilla luciferase , and induced with R7910 for 6 h , or LPS for 4 h . Fig . 8 A shows that CD14 increased the integrin–TLR2 mediated response by about 2 . 5 fold . CD14 had no effect on the silenced cells ( right panel ) , not even on those expressing TLR2 , indicating that it does not substitute for αvβ3-integrin in enhancing the TLR2 response . Further , we investigated whether the CD14-mediated enhancement of the NF-κB response involved the C-tail signaling portion of β3-integrin . In sh-β3 cells expressing the β3-integrinY747-Y759 mutant the CD14-mediated enhancement of NF-κB response was almost abolished ( Fig . 8 B ) . Altogether , the enhancement by CD14 of the αvβ3-integrin–TLR2 response strengthens the above conclusion that integrin–TLR2 cooperation rests on boosting of the TLR2 signaling by integrin , in particular by its C-tail . We asked which is the subcellular compartment where the αvβ3-integrin–enhanced recruitment of MYD88 to TLR2 takes place , and , in particular , whether the TLR2-MYD88 complex is assembled at or around lipid rafts . Wt or sh-β3 293T cells were transfected with TLR2-Flag and MYD88-HA . MAL , the connector between TLR2 and MYD88 , was included . Cells were exposed to R7910 for 30 min . Membranes were fractionated , and allowed to float in sucrose gradients . Previously , we showed that the top light fractions of the gradient contain molecules typical of lipid rafts , e . g . GPI-anchored receptors [26] . The middle fractions contain molecules that localize at or around lipid rafts . The bottom fractions contain the heavy membrane fractions . Fig . 9 shows the immunoblot analysis of gradient-partitioned membranes . Prior to cell exposure to virus , TLR2 , MYD88 and MAL partition with the heavy fractions of the gradient ( Fig . 9 A ) . Following exposure to R7910 , a portion of TLR2 , MYD88 and MAL partitions at light-middle fractions of the gradient ( Fig . 9 C ) , indicating that HSV induces a relocalization of these molecules at or around lipid rafts . The re-localization required integrin , since it did not occur in sh-β3 cells ( Fig . 9 , B and D ) . The results indicate that αvβ3-integrin participates in the initiation of the TLR2 signaling also by promoting the relocalization of TLR2 , MAL , MYD88 . Likely , the relocalization favours complex assembly . We next examined the contribution of CD14 to the lipid raft localization of the above molecules . The above experiment was repeated in the presence of CD14 . Fig . 9 E shows that , in the presence of CD14 , a portion of TLR2 , MAL , MYD88 partition to the middle fractions , i . e . are localized at or around lipid rafts , prior to exposure of cells to HSV . Exposure of cells to HSV further increases the lipid raft localization ( compare panel G to panel E ) . The HSV-induced compartimentalization to lipid rafts does not occur in sh-β3 cells ( Fig . 9 , F , H ) , indicating that it requires αvβ3-integrin .
We investigated how αvβ3-integrin and TLR2 act in concert to elicit the immediate branch of the innate response to HSV and to LPS . The key findings to emerge are that ( i ) αvβ3-integrin boosts the MYD88-dependent TLR2 signaling and defensive response; this was seen in all cell lines tested , i . e . the model 293T , and the keratinocytic , epithelial and neuronal cell lines , which are models of the HSV targets in vivo . In contrast , TLR2 exerts no effect on the αvβ3-integrin signaling pathway . ( ii ) The herpes simplex virion glycoproteins gH/gL serve as the PAMP and cross-link the two triggering receptors , ( iii ) αvβ3-integrin and TLR2 are relocated to lipid rafts in a ligand ( virion ) -dependent fashion; ( iv ) Akt serves as a hub of the signaling pathway ( see , Fig . 11 for a schematic overview ) . The initial investigations provided evidence that the basis of the concerted αvβ3-integrin–TLR2 action rests on signaling activity , and not on enhancement of HSV entry , or LPS binding/uptake , since the NF-κB activation , IFN-β secretion , and recruitment of MYD88 to TLR2 were strongly impaired when wt β3-integrin was replaced with a mutant carrying substitutions in the cytoplasmic tail , which prevent phosphorylation and signaling . Moreover , the gD−/− virions , competent for attachment but not for entry , triggered a response which required the signaling portion of β3-integrin C-tail , hence similar to that elicited by the replication competent HSV . This indicated that entry of HSV into the cell is not a requirement to initiate the immediate branch of the response mediated in concert by αvβ3-integrin and TLR2 . This response clearly differs from the one elicited by fusion of HS virions with the cell [28] . Numerous studies highlighted that the cell response to HSV can be differentiated in at least two temporal waves . The first , immediate response is that exerted by UV-inactivated ( able to enter cells but unable to express viral genes ) HSV , exemplified in current studies by gD−/− virions . They activate NF-κB and other lines of defence within a few minutes after infection [29] , [30] , along with activation of FAK [31] . More sustained NF-κB activation and cell defence occur at later times under the stimulus of additional viral gene products , including US3 , VP11/12 , UL31 [32]–[37] . While all these gene products play critical roles in eliciting the global cell defence , nonetheless , previous and current studies from our laboratory indicate that the response elicited by the αvβ3-integrin/TLR2 axis make a significant contribution also to the second wave of defence . Indeed , the absence of TLR2 , the silencing of β3-integrin , or both , dramatically decreased the NF-κB activation , and , importantly , the IFN-α and -β production [8] . We dissected the pathways downstream of αvβ3-integrin and downstream of TLR2 . The key findings were that αvβ3-integrin boosts the TLR2 signaling , which typically involves recruitment of TIRAP/MAL , MYD88 , followed by phosphorylation of IRAK1 and IRAK 4 [19] . In β3-integrin-silenced cells ( 293T , HaCaT , HeLa and SK-N-SK ) , or in 293T cells expressing the C-tail mutant of β3-integrin , MYD88 was not recruited to TLR2 , and consistently , IRAK 4 phosphorylation was almost abolished in β3-integrin-silenced 293T cells . Additional support for a functional role of the MYD88 was provided by a DN version of this molecule ( MYD88FW/AA ) , capable to be recruited to TLR2 but unable to signal downstream [20] . MYD88FW/AA drastically reduced NF-κB activation . HSV counteracts the αvβ3-integrin/TLR2 antiviral response , soon after it enters the cell , at the onset of viral protein synthesis by means of the immediate early protein ICP0 [8] . ICP0 alone can reduce the levels of MYD88 and MAL [38] , a property that lends indirect support to the αvβ3-integrin–mediated boosting of TLR2 signaling . The converse effect , i . e . the enhancement by TLR2 of the αvβ3-integrin signaling response was not detected . In particular , we ruled out an effect of TLR2 on the activation of the non receptor tyrosine kinase Src . This was the most appropriate candidate to be examined downstream of αvβ3-integrin , since Src is one of the most typical molecules activated in the integrin signaling pathway [39] , and , mainly , because Src , together with Syk and Card9 , were upregulated by HSV in an αvβ3-integrin-dependent TLR2-independent fashion in microarray analysis [8] . To further elucidate the signaling pathways downstream of the αvβ3-integrin–TLR2 axis , we considered Akt , a molecule on which several signaling pathways converge , and known to be involved in HSV infection [21] , [22] . Akt serves as a hub in the αvβ3-integrin–TLR2-mediated signaling . Thus , upon HSV infection , or LPS stimulation , Akt was phosphorylated in β3-integrin-integrin-silenced 293T , HaCaT , HeLa , SK-N-SK cells , in a TLR2-dependent fashion , in cells expressing the DN version of AKT ( PKB-CAAX ) [23] , or in cells exposed to the MK2206 specific inhibitor . gH/gL are essential glycoproteins in the process of HSV entry into the cells [15] . They are part of the conserved fusion apparatus across the Herpesviridae family , and transmit the activation signal which ultimately activates the fusion glycoprotein gB [14] . We identified the virion envelope glycoprotein gH/gL as the HSV PAMP of this system , based on three lines of evidence . First , interactions were documented here and elsewhere between gH/gL and TLR2 , or between gH/gL and αvβ3-integrin , independently one ( see Fig . 5 and references [6] , [8] , [9] . The latter was also documented between the virion gH/gL with αvβ3-integrin [40] . In addition , αvβ3-integrin and TLR2 interact one with the other under resting conditions , in the absence of ligands [6] , [8] . Secondly , we observed that gH/gL cross links the two receptors . Furthermore , gH−/− virions were defective in key steps of the αvβ3-integrin/TLR2 signaling , including the recruitment of MYD88 to TLR2 , and Akt phosphorylation . We propose that the gH/gL-mediated cross-linking of αvβ3-integrin and TLR2 represents the starting event in the activation of the signaling cascade . This mechanism differs from the αvβ3-integrin–TLR2 cooperation described by Gerold et al in response to lipopeptide [6]; in that case , the role of αvβ3-integrin was to bind the lipopetide and to present it to TLR2 , whereas in our system gH/gL cross-links the two receptors . It is important to note that , even though we have focused our studies mainly on the interaction of HSV with αvβ3-integrin and TLR2 , the main results were obtained also with LPS . Thus , HSV gH/gL should not be considered as the only PAMP in this system , and , in turn , this defensive system should not be considered as exclusively devoted to anti-viral activity . CD14 is a GPI-anchored accessory molecule , or coreceptor for a number of TLRs , including TLR2 . Its serves several functions , including to facilitate the presentation of LPS to TLR2 , TLR2 heterodimerization , the ligand-induced localization of TLR2 to lipid rafts , and the MYD88-independent signaling of TLR4 [5] , [19] , [25] . We found that CD14 can boosts the TLR2 signaling in an additive manner relative to αvβ3-integrin , but can not substitute for integrin , since it failed to exert any enhancing effect in cells in which β3-integrin was silenced , or mutated . Hence , a clear hierarchy between CD14 and αvβ3-integrin exists: αvβ3-integrin boosts TLR2 signaling irrespective of the absence or presence of CD14 . CD14 boosts the TLR2 response only in the presence of integrin . The boosting effect of CD14 on TLR2 response described here differs from that seen in macrophages , a system in which CD14 was a requirement [41] . From a mechanistic point of view , αvβ3-integrin relocates TLR2 , MAL and MYD88 at or around lipid rafts . CD14 further enhances both the ligand–independent and –dependent lipid raft localization of the receptors . The functional perturbation of lipid rafts abolished the HSV-induced NF-κB activation and IFN-β response . Of note , αvβ3-integrin relocates also relocates the HSV receptor nectin1 to lipid rafts [42] . Altogether , lipid rafts represent the platforms at , or around which the αvβ3-integrin-enhanced recruitment of MYD88 to TLR2 takes place . Under this respect , αvβ3-integrin behaves similarly to CD14 and CD36 coreceptors [43] , [44] . A key result is that this branch of the innate response is highly detrimental to the virus . R7910 replication was dramatically increased ( 2–4 Logs ) in cells in which MYD88 or Akt were replaced by DN mutants , and , especially in cells where β3-integrin was silenced and MYD88 or Akt were simultaneously replaced with DN mutants . We conclude that the cell deploys the αvβ3-integrin/TLR2 mediated response as a defensive antiviral system . The key finding of this work , that αvβ3-integrin boosts the MYD88-dependent TLR2 signaling ( see schematic overview in Fig . 11 ) , is best interpreted in the context of coreceptors that reinforce or broaden the activity of TLR2 , and their cell-type distribution . Among the cell surface TLRs , TLR2 emerges as the one capable to recognize a wide range of molecular patterns . The broad repertoire of exogenous and endogenous molecular patterns may well be achieved through interaction with co-receptors . In addition to CD14 mentioned above , which is preferentially expressed in monocytic cells , TLR2 coreceptors include CD36 , a member of the scavenger receptor family preferentially expressed in monocytic and endothelial cells [45]–[47] , certain leukocyte- monocyte-specific integrins , e . g . αMβ2-integrin and α3β1-integrin in monocytes-macrophages [5] , [19] , [48] , [49]; αMβ2-integrin positively regulates TLR4 in in dendritic cells [50] . Previously , we reported that the concerted αvβ3-integrin–TLR2 response represents a major innate response elicited by HSV in epithelial , keratinocytic and neuronal cell lines , i . e . in cells which are models of the cells targeted by HSV in vivo [7] . The silencing of β3-integrin in these same cells results in inhibition TLR2 signaling , seen as inhibition of MYD88 recruitment to TLR2 . Given the wide distribution of αvβ3-integrin in epithelial cells , we propose that αvβ3-integrin may well serve as the coreceptor employed by TLR2 in these cells , and what so far was described as the TLR2 response in epithelial cells is very likely the concerted αvβ3-integrin–TLR2 response . In essence , the role of the widely expressed αvβ3-integrin was so far simply unnoticed . Taken together , current and previous data argue that in several cell systems TLR2 actually requires one or another coreceptor , and that the coreceptors likely contribute to the cell type specificity and broad spectrum of the TLR2 response .
293T , HeLa , SK-H-SH and U20S cells were received from American Type Culture Collection and grown in Dulbecco's modified Eagle's medium containing 10% foetal bovine serum . HaCaT cells were received from Deutsches Krebsforshungzentrum , Heidelberg , and grown in high ( 4% ) glucose RPMI containing 10% foetal bovine serum . β3-integrin silenced 293T cells ( named sh-β3 ) and the β3-integrin silenced HaCaT , HeLa , SK-H-SH ( named HaCaT sh-β3 , HeLa sh-β3 , SK-H-SH sh-β3 ) were described [7] , [8] . R7910 is a HSV mutant deleted in the gene encoding ICP0 [18] . The wRR-1097 gD deletion HSV was described [51] . The gH deletion HSV was described [24] . The mammalian expression plasmids encoding HSV-1 gH , gL , gD , under the cytomegalovirus promoter , were described [52] . The αv , β3wt , β3Y747F , β3Y759F and β3Y747-759F expression plasmids were a generous gift from Dr . Blystone [17] . Plasmids encoding TLR2-Flag and NF-κB-luc were a generous gift from Dr . D . M . Knipe [4] . pCMV-HA-MYD88 , pCMV-HA- MYD88FW/AA , pEFBos-MAL-Flag and CD14 in pCDNA were form Addgene . pcDNA3 . 1 was from Invitrogen . Renilla luciferase plasmid was from Promega . Plasmids encoding HA-PBK and PKB-CAAX were generous gifts from Dr . B . Burgering [23] . The plasmid encoding strep-tagged αv-integrin ( named αvst or αvstrep ) was generated as follows . αv-integrin was excised from pcDM8 by insertion of two restriction sites , one at the 5′ ( NotI ) and one at 3′ end ( BamHI ) , just before the stop codon . The mutagenic oligonucleotides were 5′-gcttggcgtcccgcgGCcGcttcggcgatggcttttcc-3′ and 5′-ggaaaagccatcgccgaagCgGCcgcgggacgccaagc-3′ , and 5′- ggtgaaggaaactcagGGaTCCaactgcagtttttaagttatgc-3′ and 5′- gcataacttaaaaactgcagttGGAtCCctgagtttccttcacc-3′ , respectively . The excised αv open reading frame was cloned into a pcDNA plasmid containing the sequence encoding the One Strep tag ( GISGWSHPQFEKGGGSGSGGGSWSHPQFEK ) in frame with the C-ter of αv-integrin . The M2 monoclonal antibody ( MAb ) anti-Flag was from Sigma-Aldrich; MAb anti-HA was from Covance . Polyclonal antibody ( PAb ) to gH/gL and MAb H170 to gD were described [53] . PAbs to Akt , phospho-Akt ( Ser473 ) , Src ( 36D10 ) , phospho-Src ( Tyr416 ) , IRAK4 , phospho-IRAK4 ( Tyr345/Ser346 ) were from Cell Signaling . Strep-Tactin HRP ( horseradish peroxidase conjugate ) was from IBA GmbH ( Gottinghen ) . PAb 1932 to β3-integrin was from Chemicon . The PP1 ( 4-Amino-5- ( methylphenyl ) -7- ( t-butyl ) pyrazolo- ( 3 , 4-d ) pyrimidine ) inhibitor of Src and the MK2206 inhibitor of Akt phosphorylation were from Sigma and Merck , respectively . MAb to IκB-α was from Cell Signaling and anti-tubulin MAb was from Sigma-Aldrich . 293T or sh-β3 cells were transfected by means of Lipofectamine2000 ( Invitrogen ) with plasmid encoding firefly luciferase under a NF-κB regulated promoter , and Renilla luciferase in a 130∶1 NF-κB-luc∶Renilla ratio , plus TLR2-Flag or pcDNA 3 . 1 empty vector as indicated [8] . CD14 , HA-PKB , PKB-CAAX , pCMV-HA-MYD88 , pCMV-HA-MYD88FW/AA , β3wt , β3Y747F , β3Y759F or β3Y747-759F and αv-integrin , or a combination of plasmids , were included , as indicated in the text or figure legends . The transfected cells were maintained in pre-exhausted medium for two-three days prior to infection or LPS exposure [8] . Cells were exposed to 20 PFU/cell of the indicated virus for 6 h or LPS ( Sigma-Aldrich , # L2630 ) ( 100 ng/ml ) for 4 h . For treatment with inhibitors , cells were pre-treated for 1 h prior to exposure to virus ( 30 min for filipin III ) and during virus absorption with the compounds . Luciferase activity was quantified by means of Dual Glo-luciferase reporter assay system ( Promega ) . 293T , HaCaT , HeLa or SK-N-SH cells , silenced or not for β3-integrin ( cells were transfected by means of Lipofectamine2000 ( Invitrogen ) with TLR2-Flag-encoding plasmid ( 0 . 5 µg DNA for 10 cm2 dishes ) , plus plasmids encoding pCMV-HA-MYD88 ( 0 . 5 µg DNA for 10 cm2 ) . When indicated sh-β3 cells were transfected with TRL2-Flag and pCMV-HA-MYD88 plasmids plus αvwt plus β3wt or αvwt plus β3Y747-759F for 48 h . 2–3 days after transfection , cells were exposed to R7910 ( 60 PFU/cell ) , or gH−/− virions ( 60 PFU equivalent/cell ) , LPS ( 300 ng/ml ) for 10 , 20 , 30 min at 37°C and lysed in PBS plus 1% DOC ( deoxycholic acid ) , 1% Igepal containing the protease inhibitors Nα-p-tosyl-l-lysine chloromethyl ketone hydrochloride and Nα-p-tosyl-l-phenylalanine chloromethyl ketone ( final concentration , 0 . 3 mM each ) , as detailed [9] . TLR2-Flag was immunoprecipitated with anti-Flag M2 MAb [9] . The proteins retained by Protein G-Sepharose , were separated by polyacrylamide gel electrophoresis ( PAGE ) and WB with MAb to HA or to Flag . To detect the interaction of TLR2 with gH/gL , 293T or 293T sh-β3 were transfected with plasmids encoding full length gH and gL plus TLR2-Flag; immunoprecipitation was performed as described above by means of anti-Flag MAb; gH was detected by means of PAb to gH/gL . To detect the TLR2 and β3-integrin interaction , sh-β3 cells were transfected with αvwt plus β3wt or αvwt plus β3Y747-759F and TLR2-Flag . TLR2 was immonoprecipitated by means of anti-Flag MAb . β3-integrin was detected by WB with PAb 1932 . To determine IRAK4 phosphorylation , 293T or sh-β3 cells were transfected with TLR2-Flag-encoding plasmid , plus plasmids encoding HA-MYD88 . Cells were exposed to R7910 or LPS ( 300 ng/ml ) , as detailed above , and lysed in RIPA buffer ( 20 mM HEPES ( 4- ( 2-hydroxyethyl ) -1-piperazineethanesulfonic acid ) , 250 mM NaCl , 1 mM EDTA , 1 mM DTT , 0 . 5% Igepal ) containing the protease inhibitors and the phosphatase inhibitors cocktail ( Sigma-Aldrich ) . The endogenous IRAK4 was immunoprecipitated with PAb to IRAK4; phosphorylation was detected by means of anti phospho-IRAK4 PAb . The αv-integrin precipitation was carried out from sh-β3 cells , previously transfected with αvst plus β3wt or αvst plus β3Y747-759F , plus pCMV-HA-MYD88 plasmids , for 48 h; when indicated , TLR2-Flag was included . Following infection with R7910 , the cells were lysed in EA1 buffer plus ( 50 mM HEPES , 250 mM NaCl , 0 . 5% Igepal , pH 8 ) containing 0 . 3 mM protease inhibitors . αvst was harvested with Strep-Tactin Sepharose ( IBA , GmbH , Gottingen , Germany ) [53]; the retained proteins were separated by PAGE and blotted with MAb to HA to detect MYD88 , or to FLAG to detect TLR2 . 293T or 293T sh-β3 cells were transfected with plasmids encoding TLR2-Flag , plus pCMV-HA-MYD88 and pEFBos MAL Flag . When indicated , CD14 was included . 48 h after transfection , the cells were exposed for 30 min to R7910 ( 60 PFU/cell ) at 37°C or mock-infected . Cells were harvested , suspended in 1 ml of TNE buffer ( 10 mM Tris-HCl , pH 7 . 5 , 150 mM NaCl , 5 mM EDTA ) containing 1% Triton X-100 ( Sigma Aldrich , Milan , Italy ) and 0 . 3 mM protease inhibitors , and incubated on ice for 1 h . Membrane fractions were prepared essentially as described [26] . The samples obtained from fractionation of the sucrose gradient were subjected to PAGE and blotted with MAbs to HA ( for detection of MYD88 ) and to Flag ( for detection of TLR2 and MAL ) . 293T HaCaT , HeLa or SK-N-SH cells , silenced or not for β3-integrin were transfected or not with TLR2-Flag encoding plasmid . 293T sh-β3 cells were transfected with αv-integrin plasmid plus β3wt or β3Y747-759F and exposed to R7910 ( 60 PFU/cell ) or , gH−/− virions ( 60 PFU equivalent/cell for 10 , 20 and 30 min at 37°C . Cells were lysed with RIPA buffer . 300 ng of total proteins were subjected to PAGE . Src , phospho-Src , Akt and phospho-Akt were detected by WB with appropriate antibodies in two separate gels ( one for Src and one for Akt generated in parallel in the same experiment ) . 293T or 293T sh-β3 cells were transfected with plasmids encoding TLR2-Flag plus pCMV-HA-MYD88 , pCMV-HA-MYD88FW/AA , HA-PKB or PKB-CAAX . 24 h after transfection cells were infected with R7910 ( 1 PFU/cell ) for 90 min at 37°C . Extracellular virus was inactivated by means of an acidic wash ( 40 mM citric acid , 10 mM KCl , 135 mM NACl , pH 3 ) . Replicate cultures were frozen at 3 , 24 or 48 h after infection and viral progeny ( intracellular plus extracellular ) was titrated on U20S cells .
|
In an earlier work we showed that a relevant contribution to the overall IFN-based antiviral response of the cell to herpes simplex virus is exerted by αvβ3-integrin which acts in concert with TLR2 in eliciting this response . Major characteristics of this branch of the innate response are the secretion of IFN-α and -β , of a specific set of cytokines , and the activation of NF-κB . The response is elicited also by LPS , indicating that the αvβ3-integrin TLR2 sentinels sense both bacteria and viruses . The IFN response is usually thought to be elicited by the endosomal and cytoplasmic sensors . Here we have investigated the basis of the αvβ3-integrin–TLR2 response , and found that αvβ3-integrin acts through its signaling C-tail , and boosts the MYD88- IRAK4-dependent TLR2 response . This is seen also in epithelial and neuronal cells which exemplify targets of HSV infection . Altogether , the results argue that αvβ3-integrin may serve as a coreceptor of TLR2 in epithelial cells . A point of novelty is that the TLR2 coreceptors known to date - CD14 , CD36 and αMβ2-integrins - are typical of monocytic-derived cells ( macrophages , DCs ) . To our knowledge a TLR2 coreceptor for epithelial cells was not known to date .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"and",
"health",
"sciences"
] |
2014
|
The Epithelial αvβ3-Integrin Boosts the MYD88-Dependent TLR2 Signaling in Response to Viral and Bacterial Components
|
Mucosal mononuclear ( MMC ) CCR5+CD4+ T cells of the gastrointestinal ( GI ) tract are selectively infected and depleted during acute HIV-1 infection . Despite early initiation of combination antiretroviral therapy ( cART ) , gut-associated lymphoid tissue ( GALT ) CD4+ T cell depletion and activation persist in the majority of HIV-1 positive individuals studied . This may result from ongoing HIV-1 replication and T-cell activation despite effective cART . We hypothesized that ongoing viral replication in the GI tract during cART would result in measurable viral evolution , with divergent populations emerging over time . Subjects treated during early HIV-1 infection underwent phlebotomy and flexible sigmoidoscopy with biopsies prior to and 15–24 months post initiation of cART . At the 2nd biopsy , three GALT phenotypes were noted , characterized by high , intermediate and low levels of immune activation . A representative case from each phenotype was analyzed . Each subject had plasma HIV-1 RNA levels <50 copies/ml at 2nd GI biopsy and CD4+ T cell reconstitution in the peripheral blood . Single genome amplification of full-length HIV-1 envelope was performed for each subject pre- and post-initiation of cART in GALT and PBMC . A total of 280 confirmed single genome sequences ( SGS ) were analyzed for experimental cases . For each subject , maximum likelihood phylogenetic trees derived from molecular sequence data showed no evidence of evolved forms in the GALT over the study period . During treatment , HIV-1 envelope diversity in GALT-derived SGS did not increase and post-treatment GALT-derived SGS showed no substantial genetic divergence from pre-treatment sequences within transmitted groups . Similar results were obtained from PBMC-derived SGS . Our results reveal that initiation of cART during acute/early HIV-1 infection can result in the interruption of measurable viral evolution in the GALT , suggesting the absence of de-novo rounds of HIV-1 replication in this compartment during suppressive cART .
Acute infection with human immunodeficiency virus type 1 ( HIV-1 ) is a critical time during which host factors including innate and adaptive immunity converge with virologic characteristics to determine the course of clinical progression in infected individuals [1]–[3] . In the absence of combination antiretroviral therapy ( cART ) , HIV-1 infection is maintained in a chronic state- characterized by high levels of viral production with associated immune activation that is responsible in large part for progressive CD4+ T cell depletion [4] . Repeated rounds of infection of susceptible CD4+ cells of various types occur with continued generation of long-lived cells harboring replication competent virus , and that persist during therapy [5] . The HIV-1 treatment landscape was transformed when in 1996 , highly active antiretroviral therapy ( HAART ) became the standard of care for the treatment of HIV-1 infection . The advent of HAART is directly credited with the retardation of the overall progression of HIV infection to AIDS as well as the progression of AIDS to death . The end result has been a considerable decrease in morbidity , and increase in survival after AIDS diagnosis [6] , [7] . Combination antiretroviral drug therapy ( cART ) can dramatically suppress HIV replication and reduce the plasma HIV-1 viral load in compliant patients , resulting in immune reconstitution of memory CD4+ and CD8+ T cells and the restoration of T cell immunity [8]–[11] . Despite these advances , current regimens remain unable to eliminate the reservoir of latent virus in resting CD4+ T lymphocytes . As a result , cessation of therapy predictably results in the resurgence of virus replication [12]–[17] . The gut-associated lymphoid tissue ( GALT ) contains the vast majority , and most complex pool of immune cells [18] . In addition , intrinsic characteristics of the mucosal compartment , including the predominance of activated and well differentiated gastrointestinal ( GI ) mucosal CD4+ T cells with a memory phenotype [19] , [20] permit HIV-1 infection and accommodate its replication . Several studies have examined the critical role played by the GI tract in early simian immunodeficiency ( SIV ) pathogenesis . In the SIV/macaque model , SIV challenge has been shown to result in early , profound depletion of GI mucosal CD4+ lymphocytes [21]–[23] . Additional studies in the SIV/macaque model revealed that up to 60% of memory CD4+CCR5+ T cells may be selectively infected and lost during this early stage of infection [24] . We and others have shown that the human GI lymphoid system is similarly targeted during acute and early HIV-1 infection . A greater percentage of mucosal CD4+ lymphocytes express the CCR5 chemokine coreceptor [25]–[28] when compared to the peripheral blood . Preferential targeting of the CCR5+ memory CD4+ T cell subset therefore results in substantial depletion of CD4+ T cells in the gut-associated lymphoid tissue ( GALT ) . As a consequence , up to 60% of CD4+ T cells in the lamina propria of the lower gastrointestinal ( GI ) tract are lost as early as 2–4 wk after infection [29]–[31] . Cross-sectional analysis of a cohort of primary HIV-1 infection subjects demonstrated that despite immune reconstitution in the peripheral blood mononuclear cells ( PBMCs ) , persistent CD4+ T cell depletion and immune activation have been noted in the GALT in a majority of patients despite up to 5 years of suppressive cART [32] , [33] . The latent reservoir of HIV-1 is established early in infection and persists despite the initiation of cART during the primary phase of infection [12] , [13] , [34] . Employing a real-time reverse transcriptase-initiated PCR single copy assay ( SCA ) , Palmer et al . have demonstrated that the majority of infected individuals on “suppressive” ART ( defined by plasma HIV-1 RNA levels <50 copies/mL ) have persistent HIV-1 viremia below the current limit of detection of commercially available , and currently less sensitive assays . This real-time RT-initiated PCR assay quantifies HIV-1 RNA concentration down to 1 copy/mL of plasma with use of an internal control for viral pelleting and extraction followed by separate RT-PCR reactions for HIV-1 and virion control sequences [35] . Numerous studies have attempted to determine if measurable low-level residual viremia during “suppressive” cART is the result of ongoing cycles of HIV-1 replication . Studies have suggested that ongoing viral replication in patients on apparently suppressive cART may occur [36]–[46] . These include reports of decreases in low-level viremia with HAART intensification , persistent expression of unspliced ( US ) HIV-1 mRNA in PBMC , persistence of HIV-1 episomal cDNA in PBMC and modest evolution of viral envelope sequences over time [36] , [37] , [39] , [40] . Conversely and generally more recently , others have found no evidence of ongoing HIV-1 replication during suppressive cART [47]–[53] . These include studies that have generated phylogenetic data suggesting the absence of viral evolution in patients undergoing successful cART and support the notion that the reservoir is intrinsically stable [51]–[53] . Bailey et al . investigated plasma and cellular viral sequences obtained during suppressive cART and found that in some patients on antiretroviral therapy , the prolonged production of a small number of viral clones without evident evolution is a major mechanism for persistent viremia [47] . Most recently , in-vitro work by Sigal et al . proposes the possibility of cell to cell transmission as a mechanism of ongoing HIV-1 replication despite ART [54] . As a result , the important question of whether or not there is ongoing HIV-1 replication despite “suppressive” cART remains under debate . We hypothesized that one cause of persistent immune activation and CD4+ T cell depletion in the gastrointestinal lymphoid compartment despite immune reconstitution in the blood may be ongoing local HIV-1 replication in the GALT during apparently suppressive cART , which could potentially constitute a source for replenishment of the latently infected resting CD4+ T cell pool . We therefore endeavored to measure qualitatively and quantitatively the degree of HIV-1 evolution in the PBMC and GALT during cART to determine whether the lymphoid tissue of the GI tract is a reservoir of ongoing viral replication during suppressive cART . We hypothesized that ongoing viral replication in the GALT during suppressive cART would result in measurable viral evolution , with more divergent populations emerging in the GALT than in the PBMC over time . We also hypothesized that if continued HIV-1 replication in the GALT was responsible for the persistent immune activation in this compartment , the amount of HIV-1 env evolution would correlate positively with experimentally determined levels of immune activation , while correlating negatively with levels of immune reconstitution in the GALT . This research utilizes the single genome amplification ( SGA ) technique to generate HIV-1 variant sequences . This terminal dilution technique uses a single molecule of DNA or cDNA as the template for amplification and sequencing [55]–[58] . SGA has the ability to achieve a degree of detection of diversity greater than 20% within a repeatedly sampled viral population [59] and has been shown to decrease taq-induced recombination , template resampling , nucleotide misincorporation and cloning bias when compared to more conventional sequencing methods [55]–[57] , [60]–[62] . The use of this method therefore produces a more accurate representation of in-vivo HIV-1 quasispecies than compared to bulk sequencing methods [63] . The mucosal tissues harbor the primary targets of HIV-1 infection and serve as an important route for HIV-1 entry and replication . For these reasons , understanding the spectrum of HIV-1 quasispecies in the GI mucosa may be critical in revealing determinants of viral entry , persistence and developing treatment strategies to improve viral suppression and overcome an established obstacle to eradicate HIV-1 in the infected host .
In the experimental group , contemporaneous peripheral-blood and recto-sigmoid colonic mucosal tissue samples were obtained from three cART-naïve , HIV-1 seropositive males identified during acute/early HIV-1 infection ( Table 1 ) . Combination ART was initiated in this group within 72 hrs of flexible sigmoidoscopy . Experimental subjects reported >95% uninterrupted adherence to cART at the time of repeat phlebotomy and flexible sigmoidoscopy with biopsies 15 to 25 months post initiation of cART ( Table 1 ) . Levels of immune activation and immune reconstitution following cART for these individuals had been previously determined at the time of initial sample collection using flow cytometry and immunohistochemistry respectively . During early HIV-1 infection , all individuals demonstrated CD4+ T cell depletion at similar levels in the PBMCs , with CD4/CD8 ratios below 1 ( Figure 1A , 1B ) . The CD4/CD8 ratio inversion was more pronounced in the MMC compartment than in the PBMC compartment for all individuals at this early time point . Following 1–2 years of cART , all individuals had successful reconstitution of the peripheral CD4+ T cell compartment , with substantial increases in peripheral CD4+ T cell counts ( Table 1 ) and normalization of the CD4/CD8 ratio . Less successful CD4+ T cell reconstitution in the MMCs was seen , and none of the individuals reconstituted to a CD4/CD8 ratio >1 despite cART ( Figure 1B ) . Immunohistochemistry on GI biopsy specimens further demonstrated the inability of these individuals to fully reconstitute CD4+ T cells of the GI tract through specific examination of the lamina propria ( LP ) , an important immune-effector site where preferential CD4+ T cell depletion occurs during primary HIV infection [30] . Mean CD4+ T cells per unit area were severely depressed during early infection . Following 1–2 years of cART , each individual was able to reconstitute LP CD4+ T cells with varying degrees of success , although none quite to the 11 . 0+/−3 . 3 cells/unit area seen by Mehandru et al . in HIV-1 negative individuals [32] ( Figure 1D , 1E ) . Similarly , during primary HIV-1 infection prior to initiation of cART , levels of immune activation ( as measured by %CD8+CD45RO+HLA-DR+ T cells ) were abnormally elevated in the CD8+ PBMCs and MMCs for all subjects ( Figure 1C ) . After >1 year on cART , levels of immune activation in the PBMC compartment dropped dramatically to levels that approximate those seen in HIV-1 negative individuals ( 4 . 8%+/−3 . 9% ) [32] . However , in the MMC compartment , while decreased from primary infection , levels of memory CD8+ T cell immune activation remained elevated above that seen in historic HIV-1 negative control individuals ( 19 . 8%+/−9 . 8% ) [32] , for our High Immune Activator ( HIA ) at 35 . 5% . GALT immune activation levels for the Intermediate Immune Activator ( IIA ) and Low Immune Activator ( LIA ) were 25 . 2% and 12 . 1% respectively ( Figure 1C ) . Given the importance of the GI LP as a target for CD4+ T cell depletion during primary HIV-1 infection , the link between persistent immune activation in the CD8+ T cell compartment to adverse HIV-1 disease progression [64] and the negative correlation in these three individuals between the ability to reconstitute CD4+ T cells in the GALT and the degree of residual CD8+ T cell activation in the MMCs , subjects were chosen based on the exhibition of one of three phenotypes in the GI lymphoid tissue at the time of 2nd biopsy: ( 1 ) High Immune Activation ( HIA ) with 35 . 5% CD8+ CD45RO+HLA-DR+ T cells in the mucosal mononuclear cell ( MMC ) compartment and an absolute CD4+ T cell count of 4 . 3 cells/unit area in the GI lamina propria ( 2 ) Intermediate Immune Activation ( IIA ) with 25 . 2% CD8+CD45RO+HLA-DR+ T cells and an absolute CD4+ T cell count of 6 . 7 cells/unit area and ( 3 ) Low Immune Activation ( LIA ) with 12 . 1% CD8+CD45RO+HLA-DR+ T cells and an absolute CD4+ T cell count of 7 . 2 cells/unit area ( Figure 1C , 1E ) . All three had plasma HIV-1 viral loads of <50 copies/ml at the time of the 2nd GI biopsy and CD4+ T cell counts of 774 , 903 and 556 cells/ml respectively ( Table 1 ) . Repeat testing of stored plasma corresponding to the 2nd PBMC sampling time point revealed all individuals to have plasma HIV-1 viral loads of below 20 copies/ml ( Roche Taqman , v . 2 . 0 ) . For individual IIA , GALT TP2 occurred approximately 1 year after PBMC TP2 . Although plasma HIV-1 RNA was determined to be <50 copies/ml at GALT TP2 , no PBMCs or plasma were stored at GALT TP2 for retrospective analysis . As a result , although the plasma HIV-1 RNA for IIA one year prior to GALT TP2 was clearly <20 copies/ml , we were unable to determine if the plasma viral load for this individual at GALT TP2 was also <20 copies/ml . For the control group , peripheral-blood mononuclear cells ( PBMCs ) were obtained during their presentation with early HIV-1 infection ( Table 1 ) . Both participants elected to remain naïve to cART , and were continuously viremic to the second PBMC sampling time point , 1 . 8 to 1 . 9 years later . Recto-sigmoid colonic mucosal tissue sampling was not performed for the positive controls . Single genome amplification ( SGA ) of full-length HIV-1 env ( >2 . 5 kb ) was performed on proviral genomic DNA from cryopreserved PBMCs and GALT tissue where indicated for each individual using a modification of the method of Salazar-Gonzalez et al . [63] , [65] . A total of 378 confirmed single genome sequences ( SGS ) from the three experimental and two positive control patients were obtained as described in Methods . On average , 23 SGS were obtained per patient , per time-point , per compartment . Phylogenetic analysis demonstrates that sequences from each experimental and control patient form tight and distinct clusters ( Figure 2 ) . This is consistent with the absence of contamination between patient samples during PCR [66] . All experimental and control participants were found to harbor HIV-1 Subtype B virus using the REGA HIV-1 subtyping tool [67] , [68] . Along with CD4 , HIV-1 typically uses the CCR5 chemokine coreceptor for entry early in infection [69] . Genotypic changes allowing the virus to use CXCR4 have been associated with the more rapid progression of HIV-1 disease [70] . Analysis of translated HIV-1 V3 loop sequences for all viral quasispecies from the positive controls and participants HIA and IIA were predicted by the Web Position Specific Scoring Matrix ( PSSM ) tool ( SINSI matrix ) to utilize the CCR5 HIV-1 co-receptor for viral entry [71] , [72] . Interestingly , translated V3 loop sequences from individual LIA were predicted to use the CXCR4 co-receptor for viral entry . The PSSM bioinformatic method has a reported sensitivity of 84% sensitivity and 96% specificity for the prediction of CXCR4 usage [72] . Given the relative infrequency of transmitted CXCR4 virus [73]–[75] , a cyropreserved pre-treatment plasma specimen from LIA was sent for cell-based confirmation of co-receptor usage for entry into CD4+ cells . The LIA plasma sample submitted for the determination of HIV-1 tropism using the monogram assay was collected on the same day as the pre-treatment samples used for SGA of HIV-1 env from PBMC and GALT in this individual . Using pseudotyped virus engineered to express the LIA HIV-1 env , Trofile assays ( Monogram Biosciences ) [76] reported a dual/mixed virus population with the ability to use CXCR4 and/or CCR5 co-receptors to enter the CD4+ cell . The reported relative light units ( RLUs ) for CCR5 and CXCR4 usage in the population were 132 , 786 and 1 , 439 respectively ( near the limit of detection for CXCR4 ) . Given the uniform predictions for all LIA quasispecies in PSSM scoring , we concluded the LIA viruses were likely to represent a dual mixed virus population with the ability to utilize both co-receptors for cellular entry . To demonstrate our ability to detect measurable viral evolution when expected , we generated HIV-1 env SGS from the PBMCs of two individuals who chose not to initiate cART during primary HIV-1 infection . For these two cases , a total of 98 confirmed SGS were obtained ( 43 for POS1 and 55 for POS2 ) . Highlighter plots ( http://www . hiv . lanl . gov ) were used to visualize individual nucleotide polymorphisms within sequences under consideration [77] . Both phylogenetic trees and Highlighter plots for each intra-patient viral variant for POS1 and POS2 are shown ( Figure 3 ) . For both individuals , SGA-derived full-length HIV-1 env sequences unequivocally revealed that over time , in the absence of cART , there is predictable diversification away from the virus population initiating the infection . The Tamura-Nei substitution model in MEGA 4 . 0 . 2 [78] was used to calculate average pair-wise distances ( APD ) between intra-patient HIV-1 env quasispecies . These calculations allowed for the quantification of genetic diversity within and between PBMC populations at each intra-patient time point . Several sequences were found to have a high statistical likelihood ( p<0 . 05 ) of G→A hypermutations using the Hypermut program available via the Los Alamos website [79] . In contrast to true sequence evolution , hypermutations are believed to be the product of a single replication cycle and as such , they cannot be regarded as evidence of gradual sequence evolution [80] , [81] . As a result , these sequences were excluded from quantitative determinations of genetic diversity . For POS1 the average within-patient env nucleotide diversity at PBMC time point #1 was 0 . 13% and increased to 1 . 23% by the time of the second PBMC sampling 1 . 9 years later . For POS2 , the within-patient env nucleotide diversity at PBMC time point #1 was 1 . 11% . It has been estimated that the maximum within-patient diversity plausibly developing from infection with a single viral variant within 100 days is approximately 0 . 6% [63] . The within-patient env nucleotide diversity of 1 . 11% within 54 days of infection in POS2 therefore suggests the transmission of more than one viral variant in this case . Examination of the maximum likelihood ( ML ) phylogenetic tree and Highlighter analysis ( Figure 3 ) for this individual was consistent with this hypothesis , as several distinct sequence clusters with ML bootstrap values >85% supported the transmission of more than one viral variant . The high multiplicity of HIV-1 infection in this individual is compatible with previous results by Li et al , in which SGA of HIV-1 env revealed 36% of their men who have sex with men ( MSM ) cohort to have evidence of productive infection with more than one HIV-1 Subtype B virus [82] . At the time of the second PBMC time point 1 . 8 years later , the within-patient env nucleotide diversity in this individual was found to be 1 . 08% . To quantify the amount of evolution that occurred during the time interval studied , we determined the APD between SGS obtained from ARV naïve time points #1 and #2 . For POS1 , the experimentally observed HIV-1 env nucleotide diversity between the two time points was 1 . 68% . For POS2 , given the evidence of infection with more than one viral variant , we conservatively compared the APD between a monophyletic subset of time point #1 SGS phylogenetically closest to the time point #2 sequences ( labeled Variant 1 in Figure 3 ) and the time point #2 sequences . Using this approach , the HIV-1 env nucleotide diversity between the two time points was 1 . 36% . Given a published rate of evolution of the C2-V5 region of the HIV-1 env gene of approximately 1 . 0% per year [83] , the experimentally observed env nucleotide diversity between PBMC time points #1 and #2 for each individual are within the expected range for evolving viral quasispecies over a period of 1 . 8 ( POS2 ) to 1 . 9 ( POS1 ) years . Altogether , these results demonstrate our ability to detect evolution using SGA when multiple replication cycles have occurred in untreated patients . As detailed above , the three study subjects achieved a significant degree of immune reconstitution in the peripheral blood based on CD4+ T cell count . We hypothesized that if persistent immune activation and CD4+ T cell depletion in the GALT was the result of local ongoing virus replication , evolution of GALT-derived HIV-1 env would be most likely to occur in the individual with the highest levels of residual immune activation and lowest level of immune reconstitution despite cART . Therefore , we anticipated that if we were to observe measurable evolution of viral quasispecies during cART , that two scenarios would be apparent: ( 1 ) The amount of evolutionary divergence between GALT-derived HIV-1 env populations sampled 1–2 years apart would be greater than between longitudinal , contemporaneous PBMC-derived HIV-1 env populations and ( 2 ) The amount of genetic diversity developing in an individual's GALT-derived HIV-1 env quasispecies over time would correlate positively with the level of immune activation and CD4+ T cell depletion measured experimentally . As a result , we would expect the greatest amount of within-patient evolution of HIV-1 viral quasispecies to occur in HIA , followed by IIA , with the lowest levels of evolution in LIA . Phylogenetic analyses were first used to estimate sequence divergence as described for the positive control cases . Figure 4 shows the ML trees with bootstrap values for HIV-1 env sequences derived from pre- and post-treatment PBMC and GALT for participants HIA , IIA and LIA . Corresponding Highlighter plots allow for visualization of individual nucleotide polymorphisms within sequences under consideration [77] . Visual inspection of maximum likelihood phylogenetic trees and Highlighter plots of full length HIV-1 env clearly reveal highly homogeneous viral populations during early HIV-1 infection in both the PBMC and GALT immediately prior to the initiation of cART for individuals HIA and LIA . This is consistent with productive infection in these individuals with a single virus . In each individual , there is a high degree of similarity between SGA-derived HIV-1 env sequences from each compartment within the viral variants establishing primary infection . In contrast to what is observed in the absence of cART , the population structure of these ML trees demonstrates no visual evidence of evolutionary diversification away from the virus populations identified during primary infection after 1–2 yrs of suppressive cART in either the PBMC or GALT . Using the same sequence alignments , phylogenetic trees were also generated using the neighbor-joining method [84] implemented in the MEGA program Version 4 . 0 [78] ( data not shown ) . Bayesian estimation of phylogenies for experimental and control individuals were also generated using MrBayes 3 . 1 . 2 [85] yielding similar population structure results ( Figure S1 ) . SGA followed by phylogenetic analysis indicates that individual IIA was productively infected with a minimum of two , and possibly three distinct HIV-1 viral variants . In this individual , a small number of in-vivo inter-lineage recombinant viruses ( Figure 4B and Table S1 ) were detected through visual interrogation of the Highlighter plot . The Hudson-Kaplan test [86] implemented in the DnaSP software [87] found a minimum of 9 recombination breakpoints among the recombinant sequences . Recombination breakpoints were confirmed and corresponding levels of significance determined using the Recco program [88] ( Table S1 ) . The noted recombinants were observed in both the PBMC and GALT and at both time points analyzed . Recombination of HIV-1 env quasispecies results in non-clock genetic evolution that can incorrectly bias quantitative estimates of population divergence [89] . Additionally , owing to their relative rarity in the population of HIV-1 envs sampled from participant IIA , the recombinants sampled 1–2 years into cART are just as likely to have been present at the initial sampling time point and therefore do not represent evidence of viral evolution during suppressive cART . Evidence for the absence of HIV-1 env evolution in the PBMC and GALT of the three experimental individuals , regardless of level of residual GALT immune activation or CD4+ T cell depletion is readily found in the phylograms and Highlighter plots ( Figure 4 ) . In an effort to quantify evolutionary relationships represented in the phylogenetic trees , we determined estimates of population diversity and divergence over sequence pairs within and between groups by measurement of average pairwise distances ( APDs ) . As in the positive control group , sequences with a significant probability of hypermutation were excluded from this analysis . To detect statistically significant shifts in population structure that might not be evident from APD calculations alone , we used a nonparameteric test for panmixia [90] . This test was derived from a geographic subdivision detection test proposed by Hudson et al . [91] . In an investigation of viral evolutionary divergence in the SIV/pigtail macaque model [92] , p<104 was used as a significance threshold for panmixia . This stringent p- value threshold was chosen to correct for multiple nucleotide comparisons and make panmixia detection robust to potential sampling error by SGS [92] . In this analysis , we have instead chosen to fix the p value threshold for panmixia at p<0 . 05 , apply a Bonferroni correction to p-values obtained from testing of multiple compartments across time points , and require a second , confirmatory test to infer true shifts in population structure in a sample . The second test used for analysis of population structure was the Slatkin-Maddison ( SM ) test [93] . Implemented in HyPhy [94] . Given an input tree and predefined groups of sequences in that tree , the SM test uses parsimony to determine the number of migration events , i . e . the number of times a sequence from one group migrated to another group; the observed number of migrations - the degree of intercompartment segregation - is then compared to a null distribution obtained by repeated shuffling ( 1000 iterations ) of the group assignments . This approach applies a parsimony criterion to the evolution of each character on the maximum likelihood gene phylogeny in question , and assesses the degree of variation from the normal distribution of simulated sequences over the tree to assess the degree of intercompartment segregation . The number of base substitutions per site averaged over all sequence pairs both within and between groups is shown as well as results for tests of panmixia ( Tables 2 and 3 ) . As the diversity between viral lineages at transmission represents mutations accumulated in the donor , the determination of viral evolution as a function of changes in genetic diversity is only valid when comparing diversification away from known parental variants . As a result , for participant IIA , only one distinct group of sequences corresponding to one of the multiple infecting viral variants demonstrated in Figure 4B was considered following the removal of both hypermutated and recombinant viral variants: population ( Variant 1 ) with 48 SGS ( Figure 4B and Table 3 ) . Given the need to further distinguish between SGA-derived HIV-1 env quasispecies obtained from PBMC pre- and post-cART and GALT pre- and post-cART , remaining populations contained too few sequences in each compartment for meaningful quantitative and statistical analyses . As a result , only the largest of the IIA infecting populations ( Variant 1 ) was analyzed ( Table 3 ) . Similarly , given evidence of multiple infecting viral variants for POS2 ( each with a small number of variants ) , statistical analysis for POS2 could not be performed ( Table 2 ) . In the event of ongoing viral replication resulting in viral evolution , we would expect to document the all of the following as seen for POS1: ( 1 ) Increasing nucleotide diversity within groups over time ( increases in within-group APD between time point #1 ( TP1 ) and time point #2 ( TP2 ) ; ( 2 ) Between group ( BG ) APDs consistent with nucleotide divergence between groups over time and; ( 3 ) Significant p- values for tests of panmixia reflecting significant population divergence over time . ( 4 ) Significant p-values for confirmatory Slatkin-Maddison test reflecting true differences in population structure . Average pairwise distances within PBMC groups during early HIV-1 infection for HIA , IIA ( infecting variant #1 ) and LIA were expectedly low ( 0 . 0011 , 0 . 0006 and 0 . 0012 respectively ) ( Table 3 ) . At the second time point after 1 . 1 to 1 . 3 years of suppressive cART , within group APDs remained low for each individual but were slightly increased for HIA ( 0 . 0015 ) and decreased for both IIA ( 0 . 0005 ) and LIA ( 0 . 0008 ) . In each individual , tests of panmixia between PBMC TP1 and PBMC TP2 yielded non-significant p-values , indicating a clear mixing of populations and providing assurance that the measured values for between group APDs ( BG APDs ) were not large enough to represent statistically significant nucleotide divergence between the populations . As a result , we conclude that the small increase in within group APD from PBMC TP2 SGS for HIA is not large enough to represent evidence of viral evolution over time , and the small decrease in within group APD from PBMC SGS for IIA and LIA does not represent evidence of loss of viral diversity over time . Average pairwise distance determinations within and between GALT groups for HIA reveal a similar story . During early HIV-1 infection , GALT populations in HIA , IIA ( Variant 1 ) and LIA were also homogenous , with low APDs within populations of infecting quasispecies ( 0 . 0009 for HIA , 0 . 0005 for IIA Variant 1 and 0 . 0014 for LIA ) . In the case of HIA , the average within-GALT nucleotide distance prior to the initiation of cART was 0 . 0009 . Following 1 . 3 years of cART , the average within-group nucleotide distance was 0 . 0007 . Sequence compartmentalization between GALT TP1 and TP2 was supported by the test of panmixia ( p = 0 . 027 ) . However , confirmatory testing using the Slatkin-Maddison test revealed no significant evidence of population structure in the GALT compartment ( p = 0 . 780 ) or over the entire phylogenetic tree as a whole ( p = 0 . 589 ) . This finding , combined with the low APD between GALT TP1 and TP2 ( 0 . 0008 ) and absence of an increase in GALT viral diversity allowed us to conclude that the numerical decrease in the within group APD was not reflective of a significant loss of viral diversity in the GALT of participant HIA at the second time point , and that there was no evidence of HIV-1 env evolution in the GALT from the individual we hypothesized most likely to exhibit residual HIV-1 replication in this compartment given their immunologic profile 1 . 3 years after the initiation of cART ( highest degree of GALT immune activation and lowest degree of CD4+ T cell reconstitution ) . For IIA , we saw a similar decrease in within group APD between GALT TP1 ( 0 . 0005 ) and GALT TP2 , ( 0 . 0002 ) arguing against any evidence of viral evolution in this compartment due to viral replication . However , given the small number of SGS in GALT TP 2 for this variant ( <10 SGS ) , the high potential for sampling error prevented us from performing a statistical test of panmixia between these two groups . As a result , we were unable to determine if the measured decrease in within-group GALT TP2 viral diversity was indicative of a significant loss of diversity over time . Slatkin-Maddison testing for the IIA Variant 1 phylogram ( with GALT TP2 sequences removed ) was non-significant ( p = 0 . 078 ) . Finally , for LIA , the within-group APDs for GALT TP1 ( 0 . 0014 ) and GALT TP2 ( 0 . 0007 ) SGSs again fails to reveal an increase in HIV-1 env diversity over time . However , the panmixia probability for this compartment is significant ( p = 0 . 003 ) and confirmed by Slatkin-Maddison testing ( p<0 . 001 ) of the compartment , and of the complete phylogram ( p<0 . 001 ) . Statistical tests of population structure across all compartments and across all time points for this individual reveals significant compartmentalization using both the panmixia and Slatkin-Maddison tests only for those compartments that are compared against GALT TP1 . Evaluation of the phylogram and Highlighter plots demonstrates a number of identical ( monotypic ) sequences . Given this observation , and the knowledge that statistical estimates of compartmentalization can be biased by identical sequences [95]–[98] , we then collapsed identical sequences into a single sequence within each compartment , at each time point . Repeat Slatkin-Maddison testing on the reconstructed maximum-likelihood phylogram revealed the absence of any significant population structure in the tree ( p = 0 . 086 ) . As mentioned earlier , evaluation of pre-treatment plasma and proviral DNA sequences revealed that patient LIA was infected with a dual-tropic HIV-1 variant capable of using both CCR5 and CXCR4 for cellular entry . The extent , if any , to which this may have influenced the expansion of pre-treatment variants in the GALT , where the target cells are overwhelmingly CD4+CCR5+ is unclear . When compared to their corresponding PBMC population at TP1 , low between group APDs and non-significant p-values for tests of panmixia reveals the absence of compartmentalization between HIV-1 env variants in PBMC and GALT during early infection with HIV-1 for individuals HIA and IIA ( Variant 1 ) . Following removal of duplicate sequences in the phylogeny , this is true for LIA as well . Taken together , in concert with the corresponding maximum-likelihood phylogenies and Highlighter plots , these results provide no evidence of significant evolution of HIV-1 env in the PBMC or GALT of individuals initiating cART during early HIV-1 infection . Additionally , this data provides no evidence that levels of immune activation in the GALT at time points during which viral suppression has been achieved in the periphery is associated with or predictive of any degree of viral evolution .
In this study we have shown that in a group of HIV-1 infected individuals initiating therapy during early infection , no evidence of substantial viral evolution could be found in HIV-1 env variants derived from the peripheral blood mononuclear cells or gut-associated lymphoid tissue after 1–2 years of suppressive cART . We initially hypothesized that if continued HIV-1 replication in the GALT was responsible for persistent immune activation in this compartment , the amount of HIV-1 env evolution would correlate positively with experimentally determined levels of GALT immune activation , while correlating negatively with levels of immune reconstitution . In the absence of quantifiable evolution in any of the three carefully selected individuals characterized as having High , Intermediate , or Low levels of Immune activation , we find no evidence to support our initial hypothesis . In an effort to make the search for evolution as rigorous as possible , this research utilized the single genome amplification ( SGA ) technique to generate HIV-1 variant sequences . SGA has been shown to decrease Taq-induced recombination , template resampling , nucleotide misincorporation and cloning bias when compared to more conventional sequencing methods [55]–[57] , [60]–[62] . The use of this method therefore produces a more accurate representation of in-vivo HIV-1 quasispecies and this work represents the first application of the SGA method to an evolutionary study of GALT-derived HIV-1 full-length HIV-1 env populations . Controversy in the field regarding the ability of suppressive cART to prevent ongoing cycles of HIV-1 replication remains . In the recent literature , the majority of studies using phylogenetic methods have suggested the absence of HIV-1 evolution in patients undergoing suppressive cART [51]–[53] although dissenting findings exist [45] . Recent literature using alternative experimental methods such as measuring the accumulation of 2-LTR circles in the periphery during intensification of cART-suppressed individuals suggests ongoing replication [41] . These findings however , are also controversial , as other recent intensification studies in patients on suppressive cART have shown no effect on endpoints such as plasma HIV-1 RNA levels or T cell activation in the PBMCs or sigmoid colon in HIV-infected patients with a suboptimal CD4+ T cell response [99] as well as no effect on HIV-1 RNA levels or 2-LTR circles in the plasma [100] . Examining the effect of intensification of suppressive cART with the integrase-inhibitor raltegravir , Yukl et al . concluded that in addition to effecting no significant decrease in HIV-1 RNA in plasma , no HIV-1 RNA decreases were noted in the duodenum , colon or rectum [101] . In contrast , the authors did report a decrease in unspliced HIV-1 RNA per 10 CD4+ T cells in the ileum in the majority of patients studied . These data suggest that if the GI tract does have the ability to act as a reservoir for HIV-1 replication , that ability may be compartmentalized to areas outside of the recto-sigmoid colon investigated in our work . One may also consider that although robust phylogenetic studies may not reveal evidence of viral evolution during cART , the ability to identify ongoing HIV-1 replication during suppressive cART may be possible using other experimental methods . With regards to the more specific question of whether or not the GALT serves as a reservoir for continued viral replication in the presence of cART , our findings are consistent with those of two recent phylogenetic studies examining this unique compartment . Imamichi et al . [102] examined compartmental differences between two sites in the gut ( colon and terminal ileum ) and peripheral blood in chronically infected HIV+ individuals . Contemporaneous intra-patient clonal sequences spanning the C2-V3 region of the HIV-1 env gene were examined from cell-associated DNA and RNA and virion RNA . Phylogenetic analysis revealed no evidence of compartmentalization of HIV-1 between the gut and peripheral blood and in two individuals , neighbor-joining trees revealed no indication of viral evolution during 12 months of suppressive cART . Additionally , Lerner et al . conclude through the bulk sequencing of GALT-derived viral variants rebounding from individuals interrupting cART after initiation of cART during primary HIV-1 , that GALT was unlikely to be a major contributor to post-interruption plasma viremia [103] . In the presence of cART , the plasma viral burden is decreased from 4–6 log in our study participants . As a result , the scale of expected evolutionary change is perhaps less than would be expected in the absence of ART . Several studies have looked at HIV-1 env replication at this lower limit of detection and their findings are informative . Mens et al . looked at evolutionary rates of HIV-1 pro-rt and env in longitudinal samples from HIV-1 controllers with median plasma HIV-1 RNA levels of 0 . 3 to 0 . 8 copies/ml and found evidence of ongoing replication [89] based on increasing viral divergence over time . In addition , while levels of viremia in the HIV-1 non-controllers was significantly higher than that of the HIV-1 controllers studied , the measured rates of evolution of HIV-1 env were not very different in the groups . Also , in a recent study by Anderson et al . [104] SGS-derived HIV-1 envs from the plasma of two individuals on suppressive cART with low level viremia was found to be identical to HIV-1 env sequences recovered from outgrowth assays from pools of resting CD4+ T cells . Finally , in the SIV model , Kearney et al . found that in well-suppressed pathogenic SIV infection , those animals with SIV RNA viral load <20copies/ml showed no evidence of evolution in the SIV pol in plasma over a 20 week period [92] . The extent to which the dual tropic nature of the infecting virus in the low immune activator ( LIA ) patient may have influenced the overall burden of HIV-1 infection in the GALT and subsequent levels of immune activation in that compartment ( given that the preferred targets for HIV-1 infection are believed to be mucosal CCR5+ memory CD4+ T cells [20] , [27] , [28] ) is unclear . While removal of duplicate sequences in the LIA phylogram abolished any statistical evidence of compartmentalization between tissues ( PBMC vs GALT ) and time points ( pre- and during- cART ) , we also must consider the extent to which the elimination of these sequences may have altered the power to detect real differences in the dataset . Known differences in HIV-1 co-receptor expression on CD4+ T cells of the peripheral blood and GI tract , coupled with the dual tropic nature of the infecting viral population in this individual could make compartmentalization at the time of infection between variants in the PBMC and GALT a possibility . Persistence would in our minds more likely reflect a founder effect as opposed to evidence of the generation of viral diversity over time given the absence of evolution on the phylogram , as well as a failure of the sequences to increase in diversity . It is also interesting that for subject IIA , there is an apparent difference in distribution of sequence types among the two or three potential clusters between GALT TP1 and GALT TP2 . Based on the amount of dilution necessary to generate single genomes , HIV DNA levels in the GALT of this individual were approximately ∼80× lower in TP2 than in TP1 . We hypothesize that this apparent shift in GALT populations may be a reflection of selective clearance of target cells ( perhaps a particular subtype of target cell ) in the GALT initially harboring a population of viral variants . We do also have to consider , particularly in the face of dramatic declines in HIV-1 DNA in this compartment , that the apparent phylogenetic shift may reflect under-sampling of the compartment . Limitations of the present study include the generation of sequence data using genomic DNA as the template . This was experimentally necessary; given the need to generate multiple HIV-1 env variants from GALT specimens obtained in the setting of plasma HIV-1 RNA levels <50 copies/ml . As a result , one must consider that the sampling of approximately 20 to 25 DNA sequences per tissue per time point from a provirus population may have hindered our ability to adequately detect and sample from the small subset of infected cells containing replication competent virus . Use of viral RNA would more accurately limit the population sampled to those env sequences derived from replication competent virus at the time of sampling . However , inclusion of archived variants in the studied population should not have been expected to preclude the discovery of evolved variants if they existed . It is also important to note the small number of subjects studied to date ( 3 experimental and 2 positive control individuals ) , as well as the relatively short duration of cART in these studies . Characterization of the study individuals using immune activation phenotypes at only one time point on suppressive cART , as well as the absence of more than one time point following viral suppression for the generation of SGS are further experimental limitations of note . Additional work on the generation of a model of expected levels of evolution in the presence of HAART would be informative . Furthermore , the experiments we report here involve the SGA of full length HIV-1 env . This method was chosen because it interrogates a significant fraction ( >2 . 5 kb ) of the viral genome that has been shown to evolve at a measurable rate in a number of publications utilizing the SGA technique . We cannot , however , rule out the possibility that we might have been able to document evidence of evolution in other regions of HIV-1 such as gag , pol or nef had they been interrogated . Also , as mucosal biopsy sampling does lead to some degree of bleeding into the tissues , we cannot fully exclude the possibility of some contamination of GALT tissue samples with contemporaneous PBMC . While this may lead to appearance of PBMC derived SGS appearing in the GALT compartment , it does not explain the absence of evolving forms from GALT founder sequences . Finally , an important consideration in these studies is the unique nature of the study population . In each of the cases , participants were identified during acute/early HIV-1 infection , and in the case of the experimental subjects , initiated cART during this period . Typically , HIV-1 seropositive patients are identified during the chronic phase of infection . It is possible that the findings in our group of individuals cannot be completely generalized to a greater population initiating cART at later time points , with a higher degree of pre-treatment diversity and potentially larger pool of latent CD4+ T cells . However , for this work , the study of individuals initiating therapy during primary infection is preferred , as interpretation of longitudinal nucleotide changes is greatly facilitated by knowledge of the viral quasispecies present close to the time of viral acquisition . In summary , our data is consistent with the observations made by others suggesting that the success of cART is due to the complete or nearly complete suppression of viral replication achieved with cART [105] , [106] . Using the robust technique of SGS of full length HIV-1 env , we were unable to identify evolved forms in the PBMC or GALT of individuals with known immunologic evidence of residual GALT immune activation despite clinical evidence of plasma HIV-1 RNA suppression . The absence of evidence of evolved HIV-1 variants during cART described in the present study supports the conclusion that at a minimum , currently available regimens of suppressive cART have the ability to abrogate de-novo rounds of HIV-1 replication in the gastrointestinal lymphoid tissue in individuals initiating such therapy during primary infection .
This study was approved by the Institutional Review Board of the Rockefeller University ( New York , New York , United States ) . All participants provided written informed consent prior to sample acquisition and all clinical investigation was conducted according to the principles expressed in the Declaration of Helsinki . Study subjects were chosen from the Aaron Diamond AIDS Research Center Acute Infection Program . All study individuals self-identified as men who have sex with men ( MSM ) and had a clinical history consistent with the acquisition of HIV-1 during sexual contact . Duration of infection at presentation was estimated as 2 weeks prior to the onset of acute retroviral illness . Serum specimens were tested for viral specific antibodies using the HIV-1/HIV-2 PLUS O enzyme immunoassay ( EIA ) and Vironostika less sensitive ( LS ) EIA ( the serologic testing algorithm for recent HIV seroconversion [STARHS] ) . HIV-1 Western Blots were performed by the Public Health Laboratory of the City of New York Department of Health and Mental Hygiene . Plasma samples were tested for HIV-1 RNA and by routine commercial testing battery including the Roche Cobas Amplicor HIV-1 Monitor Test version 1 . 5 and the Roche COBAS AmpliPrep/ COBAS TaqMan 48 System with AmpliLink Software Version 3 . 2 as per manufacturer's instructions . These results were used to stage subjects according to the Fiebig classification system for acute and early HIV-1 infection [107] . Contemporaneous peripheral blood and recto-sigmoid colonic mucosal tissue samples were obtained from three cART-naïve , HIV-1 seropositive males identified during acute/early HIV-1 infection . Positive control peripheral blood samples were obtained from two additional cART-naïve , HIV-1 seropositive males identified during acute/early HIV-1 infection . All participants were enrolled with a nonreactive detuned ELISA result or a documented negative HIV-1 test within the 6 months of biopsy . Endoscopic biopsies were obtained from the colon from macroscopically normal mucosa in all cases via flexible sigmoidoscopy and processed as previously described [30] . Briefly , the biopsies were taken using large-cup endoscopic-biopsy forceps ( Microvasive Radial Jaw , Boston Scientific , Boston , Massachusetts , United States ) ( outside diameter 3 . 3 mm ) and ( 1 ) placed immediately in tissue-culture medium ( RPMI 1640 , Mediatech , Herndon , Virginia , United States ) ; ( 2 ) placed into 2-ml pre-labeled cryovials ( Nalgene , Rochester , New York , United States ) and immediately frozen in liquid nitrogen; or ( 3 ) placed in formalin to preserve tissue architecture . Formalin-fixed tissues were washed with phosphate-buffered saline ( PBS ) , transferred to 100% alcohol and processed for immunohistochemistry . Endoscopic biopsies were not obtained for the two positive control participants . Phlebotomy was undertaken immediately prior to endoscopy where applicable . Immediately after acquisition , mucosal mononuclear cells ( MMCs ) were enzymatically isolated from mucosal biopsies using a 30-min incubation in collagenase type II ( Clostridio-peptidase A , Sigma-Aldrich , St . Louis , Missouri , United States ) followed by mechanical separation through a blunt ended 16-gauge needle . The digested cell suspension was strained through a 70-µm disposable plastic strainer . Immediately after isolation , cells were washed with PBS and resuspended in PBS containing antibodies for flow cytometry . Peripheral blood mononuclear cells ( PBMCs ) were prepared by centrifugation on a Ficoll-Hypaque density gradient ( Mediatech ) . At protocol defined participant visits , PBMCs were also stored at −70 C in 1 ml of freezing media ( 10% DMSO 90% FCS ) for genomic DNA isolation at a later time . Cell surface expression of lymphocyte antigens was performed as previously described [30] . Briefly , freshly isolated MMCs were subject to monoclonal antibody staining , followed by flow cytometry using a FACSCalibur ( Becton-Dickinson , Palo Alto , California , United States ) with analysis using CellQuest software ( Becton-Dickinson ) . Monoclonal antibodies used in this study included: anti-human CD3-fluorescein isothiocyanate ( FITC ) ( clone UCHT1; Becton-Dickinson ) , anti-human CD3-phycoerythrin ( PE ) ( clone SK-7; Becton-Dickinson ) , anti-human CD3-peridinin chlorophyll-α protein ( PerCP ) ( clone SK-7; Becton-Dickinson ) , anti-human CD4-allophycocyanin ( clone RPA T4; PharMingen , San Diego , California , United States ) , anti-human CD8 PE ( clone RPA T8; PharMingen ) , anti-human HLA-DR PerCP ( clone L243; BD Biosciences Pharmingen ) , anti-human CD45RO PE-Cy7 ( clone UCHL1 , BD Biosciences Pharmingen ) and the appropriate isotype controls . During flow cytometry , lymphocytes , initially identified by their forward- and side-scatter characteristics , were subject to phenotypic analysis . Dead cells were excluded from analysis using 7-aminoactinomycin D ( Calbiochem , San Diego , California , United States ) . To determine the percentages of CD4+ and CD8+ cells in the T cell population , gated lymphocytes were initially examined for the expression of CD3 . The CD3+ lymphocytes were then analyzed for expression of CD4 and CD8 receptors . To examine for activated memory cells , gated CD8+ lymphocytes were examined for the expression of CD45RO and HLA-DR . Endoscopic biopsy tissue sample were fixed in 4% neutral-buffered formalin and embedded in paraffin . Sections ( of 5 µm thickness ) were cut and stained with hematoxylin-and-eosin and Giemsa stains for light-microscopic evaluation . Immunohistochemistry was performed on paraffin-embedded sections after high-temperature antigen retrieval as previously described [30] . The sections were incubated with 1∶25 dilution of antibody to CD4 ( NCL-CD4-IF6 , Novocastra Laboratories , Newcastle-upon-Tyne , United Kingdom ) or to 1∶100 dilution of antibody to CD8 ( C8/144B , DakoCytomaton , Glostrup , Denmark ) for 60 min , followed by incubation with a 1∶20 dilution of rabbit anti-mouse secondary antibody ( DakoCytomaton code 259 ) for 20 min . The tertiary antibody ( APAAP-Complex Monoclonal Mouse , DakoCytomaton ) was applied in 1∶50 dilution . The incubations were carried out at room temperature and were followed by rinsing in Tris-buffered saline ( pH 7 . 4 ) for 5 min each . The alkaline phosphatase was revealed by New Fuchsin as the chromogen . CD4+ or CD8+ cells in the LP ( effector site ) and the organized lymphoid tissue ( OLT ) ( inductive site ) were quantified separately . Using a 40× objective , a standard area was set ( unit area ) , and a photomicrograph was taken with a Zeiss AxioImager M1 microscope equipped with AxioCam MRc5 digital camera ( Zeiss , Jena , Germany ) . Fifteen nonoverlapping unit areas were selected for the LP . Using AxioVision ( Release 4 . 5 ) software ( Zeiss ) , positive cells showing lymphocyte morphology were counted . Total genomic DNA from cryopreserved PBMCs from all participants and GALT from experimental participants was extracted by routine methods using the QIAamp DNA Mini Kit ( QIAGEN , USA ) . To minimize the risk of within-patient cross contamination of samples , only one participant sample from one compartment ( PBMC or GALT ) and time point was processed on any given day . Following limiting-dilution , single proviral molecules of full-length HIV-1 subtype B env gene ( >2 . 5 kb ) were amplified by nested PCR using a modification of the SGA method of Salazar-Gonzalez et al . [63] , [108] . Briefly , genomic DNA was serially diluted and distributed in replicates of 10 PCR reactions in MicroAmp 96-well plates ( Applied Biosystems , Foster City , CA ) . Poisson distribution dictates that the DNA dilution that yields PCR products in no more than 30% of wells contains one amplifiable DNA template per positive PCR more than 80% of the time [63] . Therefore , genomic DNA was endpoint diluted in 96-well plates such that fewer than 30% of the PCRs yielded an amplification product . Additional PCR amplifications were performed using this dilution in 96-well reaction plates . PCR amplification was carried out in presence of 1× High Fidelity Platinum Taq PCR buffer , 2 mM MgSO4 , 0 . 2 mM each deoxynucleoside triphosphate , 0 . 2 uM each primer , and 0 . 025 units/ul of Platinum Taq High Fidelity polymerase in a 20-ul reaction ( Invitrogen , Carlsbad , CA ) . The nested primers for generating full length env were as follows: 1st round sense primer env5out 5′-TAGAGCCCTGGAAGCATCCAGGAAG-3′ , 1st round antisense primer env3out 5′- TTGCTACTTGTGATTGCTCCATGT-3′ , 2nd round sense primer env5in 5′-TTAGGCATCTCCTATGGCAGGAAGAAG-3′ and 2nd round antisense primer env3in 5′-GTCTCGAGATACTGCTCCCACCC-3′ . PCR parameters were as follows: 94°C for 2 min , followed by 40 cycles of 94°C for 15 s , 58 . 5°C for 30 s , and 68°C for 4 min followed by a final extension of 68°C for 5 min . The product of the first-round PCR was used as a template in the second-round PCR under same conditions with the following PCR parameters: 94°C for 2 min , followed by 45 cycles of 94°C for 15 s , 61°C for 30 s , and 68°C for 4 min followed by a final extension of 68°C for 15 min . The resulting amplicons were then inspected on a 1% agarose gel ( Sigma-Aldrich , St . Louis , MO ) . All PCR procedures were carried out under clean PCR conditions with appropriate negative controls . A target of 20–30 single-genome sequences were generated for each compartment at each time point . This was followed by direct sequencing of the uncloned amplicons . HIV-1 env gene products were directly sequenced using an automated ABI Prism 3730xl DNA analyzer . ( Applied Biosystems , Inc . ) . Both strands of DNA were sequenced with partially overlapping fragments . All sequencing chromatograms were carefully inspected for sites of ambiguous sequence ( double peaks ) . Sequences for which any chromatogram revealed double peaks were excluded from further analysis , as this was indicative of amplification from more than one template or early Taq polymerase error . Individual sequence fragments for each amplicon were assembled using the CAP3 DNA sequence assembly program [109] . Multiple alignments of nucleotide sequences were produced using Clustal W [110] with the following parameters: pairwise alignment gap opening penalty 10; gap extension penalty 0 . 1; multiple alignment gap opening penalty 3; gap extension penalty 1 . 8 . All resulting alignments were inspected and corrected manually using the Alignment Explorer in the MEGA 4 . 0 software [78] when warranted . Columns with gaps were then removed from the multiple alignments using GapStrip ( www . hiv . lanl . gov ) [79] . In order to characterize co-receptor usage in our patients , V3 loop nucleotide sequences were extracted from multiply aligned full-length HIV-1 env for each participant using coordinates 7110–7216 on the HXB2 reference genome via the Gene cutter program on the HIV Los Alamos website www . hiv . lanl . gov . Translated V3 loop sequences were scored using the SINSI position-specific scoring matrix [PSSM] as per Jensen et al . [72] . The “distance matrix” calculation in MEGA 4 . 0 . 2 [78] was used to determine average pairwise genetic distances within or between compartments . Pairwise distances among HIV-1 env genes were determined using Tamura-Nei substitution model in Mega 4 . 0 . 2 . Sequences were analyzed for average within compartment-and between compartment diversity . The Findmodel tool on the Los Alamos HIV database site http://www . hiv . lanl . gov/was used to determine the most appropriate nucleotide substitution model for data description . Overall , the phylogenetic model found to best describe the data while allowing for distance matrix calculations to be performed in MEGA 4 . 0 . 2 was the Tamura-Nei model [111] . Phylogenetic trees were constructed by the maximum likelihood ( ML ) method using the Tamura-Nei evolutionary model in the PhyML program [112] . Bootstrap test of phylogeny were performed on 1000 replicates to evaluate the reliability and robustness of each internal branch in the resulting phylogenies . Each set of sequences was then visually inspected using the Highlighter tool available through the Los Alamos HIV website ( www . hiv . lanl . gov ) . Bayesian Markov Chain Monte Carlo ( MCMC ) inference of phylogeny was also performed at the nucleotide level using MrBayes version 3 . 1 . 2 [85] with the Tamura-Nei nucleotide substitution model of evolution . Neighbor-Joining phylogenies were used as a starting point . Two simultaneous independent runs were performed for each dataset , each with 4 chains of chain length 1×10∧7 sampling every 100 generations . The analyses were run until convergence , as determined by an average standard deviation of split frequencies <0 . 01 . Discarding the first 25% as burn-in , a majority-rule consensus of trees sampled from the posterior distribution was used to derive node support . Enrichment for mutations with APOBEC3G/F signatures was assessed using Hypermut 2 . 0 ( www . hiv . lanl . gov ) [79] . For each intra-patient set , the most common form in the first sampled time point was used as the reference sequence . Sequences that yielded a Fischer's exact p-value of 0 . 05 or lower were considered significantly hypermutated and excluded from analyses of sequence diversity . Recombinant sequence identification for IIA was performed using Recco [88] and by visual inspection of Highlighter analysis plots . One thousand permutations were run . The method of Hudson and Kaplan [86] as employed in the DnaSP 5 . 10 software package [87] was used to estimate the minimum number of recombination events required to explain sequence datasets . Sequences with evidence of recombination in Recco were excluded from analyses of sequence diversity . A nonparameteric test for panmixia was used to calculate shifts in population structure [90] . The online version of this test was applied from the site at http://wwwabi . snv . jussieu . fr/~achaz/hudsontest . html . The online portal allows for the input of intra-patient SGA derived HIV-1 env multiple alignments from the compartments ( PBMC and GALT ) and time points under consideration . This test was derived from a geographic subdivision detection test proposed by Hudson et al . [91] and compares an estimate of the degree of genetic differentiation in subpopulations of SGS chosen for comparison . In the absence of genetic differentiation between subpopulations , random reassignment of SGSs to different groups would be expected to recapitulate a new , imaginary population with population structures with the same distribution as the experimentally observed subpopulation . Ten thousand ( 10 , 000 ) re-labelings/permutations were used to generate a p-value for the probability that the randomized SGS-derived population structures between compartments are statistically equivalent . The Slatkin-Maddison test [93] as implemented in the HyPhy software package [94] , was also used to detect population structure amongst HIV-1 env sequences within individual ML phylograms as indicated . The significance of group separation was determined using the permutation test ( 1000 permutations ) . All non-hypermutated HIV-1 env sequences discussed in this manuscript have been deposited in GenBank ( accession numbers JQ250832-JQ251198 ) .
|
This study was undertaken to determine if the gastrointestinal tract is a site of ongoing viral replication during suppressive combination antiretroviral therapy ( cART ) ( defined by plasma HIV-1 RNA levels below 50 copies/ml ) . We found no evidence of substantial viral evolution in HIV-1 envelope sequences derived from peripheral blood mononuclear cells or cells of the gastrointestinal tract lymphoid tissue in participants initiating cART during early HIV-1 infection . To our knowledge , this is the first application of the single genome amplification technique to the comparative analysis of HIV-1 quasi-species derived from the gastrointestinal tract , demonstrating that in these individuals , cART has the ability to halt measurable evolution of HIV-1 envelope in this compartment . These findings suggest the absence of de-novo rounds of HIV-1 replication during suppressive cART and by extension , that experimentally observed , persistently elevated levels of immune activation in the gastrointestinal lymphoid tissue seen after the early initiation and uninterrupted use of cART ( despite relative immune reconstitution in the blood ) is likely due to factors other than ongoing viral replication . This implies that in this virally suppressed population , cART intensification is unlikely to significantly impact persistent CD4+ T cell depletion or increased levels of immune activation in the gastrointestinal tract .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"infectious",
"diseases",
"gastroenterology",
"and",
"hepatology",
"hiv",
"biology",
"genomics",
"viral",
"diseases",
"genetics",
"and",
"genomics"
] |
2012
|
Absence of HIV-1 Evolution in the Gut-Associated Lymphoid Tissue from Patients on Combination Antiviral Therapy Initiated during Primary Infection
|
The single glycoprotein ( G ) of rabies virus ( RABV ) dictates all viral entry steps from receptor engagement to membrane fusion . To study the uptake of RABV into primary neuronal cells in culture , we generated a recombinant vesicular stomatitis virus in which the G protein was replaced with that of the neurotropic RABV CVS-11 strain ( rVSV CVS G ) . Using microfluidic compartmentalized culture , we examined the uptake of single virions into the termini of primary neurons of the dorsal root ganglion and ventral spinal cord . By pharmacologically disrupting endocytosis at the distal neurites , we demonstrate that rVSV CVS G uptake and infection are dependent on dynamin . Imaging of single virion uptake with fluorescent endocytic markers further identifies endocytosis via clathrin-coated pits as the predominant internalization mechanism . Transmission electron micrographs also reveal the presence of viral particles in vesicular structures consistent with incompletely coated clathrin pits . This work extends our previous findings of clathrin-mediated uptake of RABV into epithelial cells to two neuronal subtypes involved in rabies infection in vivo . Chemical perturbation of endosomal acidification in the neurite or somal compartment further shows that establishment of infection requires pH-dependent fusion of virions at the cell body . These findings correlate infectivity to existing single particle evidence of long-range endosomal transport of RABV and clathrin dependent uptake at the plasma membrane .
Rabies virus ( RABV ) , a member of the Rhabdoviridae family , is a neurotropic pathogen that causes fatal encephalitis in animals and humans . The neurotropism of RABV is conferred by its single attachment and fusion glycoprotein ( G ) [1] . Virulence of specific RABV strains correlates with the neuroinvasiveness of their G proteins [2] , such that exchange of G of an attenuated strain with that of a pathogenic strain and vice versa confers the corresponding level of pathogenicity [1 , 3–5] . Although differential glycosylation [6 , 7] , dysregulation of G expression levels [8 , 9] , and increased induction of apoptosis [8] all contribute to G-dependent attenuation of RABV strains , it is apparent that a predominant mechanism by which G modulates rabies virulence is by dictating affinity for and spread between neurons . Following the bite of a rabid animal , peripheral neurons serve as conduits of the virus to the CNS . Although both sensory and motor neurons can be infected [10–14] , retrograde transmission of RABV dictates that motor neurons serve as the primary gateway for CNS invasion [15] . The predominant route of rabies virus entry into cells appears to be clathrin-mediated endocytosis ( CME ) [16–19] . Electron microscopic examination of chick embryo fibroblasts [18] and hippocampal neurons [20] show the presence of virions in coated pits . The relationship of those internalization events to infection , however , is not well established and existing studies that correlate the route of entry to eventual infection are restricted to non-neuronal cells [19] . Such studies also utilize vaccine RABV strains which may behave differently than their neurotropic counterparts . Available evidence suggests that RABV exploits existing cellular mechanisms that relay molecular signals from distal synapses to the somatodendritic compartment [21–23] . Long-range microtubule ( MT ) networks connect neuronal termini to the perinuclear region and mediate bidirectional axonal transport of proteins [24] , mRNAs , organelles and endosomes [24–26] . Other neurotropic viruses exploit these routes to invade the CNS , but differ in directionality of transport and mode of MT engagement [27 , 28] . For example , polio- [29] and adeno- [30] viruses are transported within endosomes tethered to MTs via host proteins , whereas alpha herpesviruses [31] interact with cellular motors directly via capsid and tegument proteins . For RABV , single viral particles incorporating fluorescently-tagged transmembrane and RNP proteins appear to translocate intact within axons [21] . Consistent with this , receptors recruited to virions at the plasma membrane appear to remain associated during long-range axonal transport [22 , 23] . Collectively these studies provide evidence that rabies viruses are transported intact within endosomes , but the significance of this transport for productive infection has not been examined [21–23] . In the present study , we combine infectivity and single particle imaging approaches to study rabies internalization and fusion from the termini of neurons of the dorsal root ganglion ( DRG ) and motor neuron-rich ventral spinal cords ( V SC ) mediated by the neurotropic Challenge Virus Strain 11 ( CVS ) G protein . To model natural RABV infection at neuronal termini , we adapt a polydimethylsiloxane ( PDMS ) microfluidic culturing platform [32 , 33] which physically separates neuronal cell bodies from their neurites . We demonstrate that CVS G-mediated infection is reliant on dynamin-dependent uptake processes and that the predominant endocytic route is clathrin-mediated . We provide evidence that productive infection requires viral fusion at the somatodendritic compartment following long-range transport from the neuronal termini . This work extends previous findings from epithelial cells providing evidence that infection of neuronal cells by rabies virus occurs through a clathrin-dependent entry pathway with subsequent membrane fusion occurring at the cell body .
To study uptake of RABV in neurons , we generated a recombinant VSV ( rVSV ) in which we replaced the glycoprotein ( G ) gene with that of the neurotropic rabies strain , CVS ( rVSV CVS G; Fig 1A ) . This virus expresses eGFP as a marker of infection . Transmission electron microscopy ( TEM ) of purified rVSV CVS G showed characteristic bullet-shaped particles with readily discernible glycoprotein spikes ( Fig 1B ) consistent with efficient incorporation of CVS G . SDS-PAGE analysis of the protein composition of purified virions demonstrates comparable incorporation of CVS G into rVSV compared to RABV vaccine strain SAD B19 or VSV G ( Fig 1C ) . CVS G migrates as a doublet consistent with the reported existence of two glycosylation variants of G [34] ( Fig 1C ) . Neuroinvasiveness is a defining feature of pathogenic RABV G , and incorporation of CVS G results in a corresponding shift in rVSV tropism . BSR T7/5 monolayers were less susceptible to rVSV CVS G than their non-neuroinvasive counterparts , rVSV and rVSV SAD B19 G , as evident from a reduction in the number and size of the foci ( Fig 1D ) . In addition , rVSV CVS G displayed greater capacity for spread resulting in larger plaque sizes than its non-neurotropic counterparts in N2a cells . A viral dose equivalent to MOI = 3 in N2a cells results in a calculated effective MOI < 0 . 05 in BSR T7/5 assuming the standard Poisson model of infection ( Fig 1E ) . We exclude the possibility that differential lipid composition of virus grown in N2a cells is responsible for the difference in tropism by comparing the infectivity of rVSV grown in BSR T7/5 ( rVSVBSRT7 ) or N2a cells ( rVSVN2a; Fig 1E ) . Cumulatively , these observations are consistent with a restricted tropism conferred by the CVS G . We next investigated RABV uptake in neurons of the dorsal root ganglia ( DRG ) and ventral spinal cord ( V SC ) that project into muscle tissue ( Fig 2A ) . Dissociated cultures , obtained from embryonic rats , yield neuronal populations with extensive projections as determined by immunofluorescence against the phosphorylated neurofilament H ( NF ) and the neuronal marker , Neuronal Nuclei ( NeuN; Fig 2B ) . We selectively manipulated neuronal termini by culturing neurons in the S compartments of microfluidic devices , which allow spatial isolation of neurites from cell bodies ( Fig 2C ) . By 10 days in vitro we detect significant outgrowth of projections into the distal , neurite ( N ) compartment ( Fig 2D ) , and calcein staining demonstrates that neurites are contiguous across the channel ( S1 Fig ) . Furthermore , we demonstrate by staining against phosphorylated NF that the majority of projections into the N compartment are axonal ( Fig 2E ) . We restrict diffusion of molecules between the N and S compartments by controlling liquid levels and , therefore , hydrostatic pressure across the microchannels . Consistent with effective fluidic isolation , transferrin ( Tfn ) tagged with AlexaFluor 488 added to the N compartment is retained with no detectable diffusion across the channels ( Fig 2F ) . We next evaluated the effect of endocytosis inhibitors on viral infection from the neuronal termini of cultured DRG and V SC neurons ( Fig 3 ) . For this purpose , we use dynasore and 5- ( N-ethyl-N-isopropyl ) amiloride ( EIPA ) to determine the requirement for dynamin- and macropinosome-mediated uptake respectively , and measure eGFP expression as a marker of viral infection . Consistent with dynamin-dependent endocytosis of rVSV CVS G , dynasore results in a near total block of infection in either neuronal population ( Fig 3A and 3B ) . Treatment with EIPA also results in a reduction of infection; however , inhibition was not as pronounced as with dynasore and not significant in DRG neurons . Because neurite projections cannot be assigned to specific neurons in the somal chamber , the magnitude of the inhibitory effects of dynasore and EIPA cannot be precisely quantified . This is particularly apparent for V SC cultures which exhibit limited N compartment projection relative to the number of cultured neurons . For DRG cultures which exhibit a greater efficiency of neurite outgrowth ( Fig 3C ) , dynasore reduces the percentage infected neurons from 43% to 2% . The inhibitory effect of EIPA in DRG infection was approximately two-fold with significant variability across experiments . This differed qualitatively from the more pronounced and reproducible inhibition observed in V SC cultures ( Fig 3B ) . Reductions in infection by either inhibitor were not due to off-target effects on viral replication since addition of inhibitor at 2 hpi had limited effect on infection . We also exclude chemical degradation of projecting neurites as a contributing factor: intact neurite structures were retained at the experimental endpoint in both treated and untreated cultures ( Fig 3D ) . We previously demonstrated that epithelial uptake of rVSV incorporating G from a vaccine strain of rabies ( rVSV SAD B19 G ) is clathrin-dependent [19] . To identify differences in uptake between epithelial and neuronal cells , we tested the effect of dynasore and EIPA on rVSV SAD B19 G uptake into DRG neurons ( Fig 4 ) . Treatment with dynasore resulted in an almost complete block of infection whereas EIPA treatment resulted in a more modest reduction ( Fig 4B ) . These results suggest a shared mechanism for initial uptake at the plasma membrane during SAD B19 and CSV G-mediated entry . To determine whether chemical inhibition of infection was due to a reduction in the number of virions delivered to the soma , we next examined the cellular distribution of rVSV CVS G virions labeled with AF647 ( rVSV CVS G AF647 ) by confocal microscopy ( Fig 5 ) . By 26 hpi neurons ( Fig 5A and 5B ) demonstrate efficient uptake and transport of virus: cell-associated particles are readily detected in all compartments including the cell bodies of some uninfected cells . In contrast , dynasore treatment restricts viral localization to the N compartment ( Fig 5A ) whereas EIPA has a limited effect on virus transport to the soma with viral particles present in both eGFP-positive and -negative neurons ( Fig 5B ) . Irrespective of the inhibitor used , viruses efficiently bound N compartment neurites indicating that association with the plasma membrane is unaffected . At 26 hpi , remaining intact viral particles represent a population that did not contribute to infection . We , therefore , examined rVSV CVS G AF647 uptake disruption at an earlier timepoint , 6 hpi ( Fig 5C and 5D ) . Here , we did not detect chemical disruption of viral association with the neuronal membrane ( Fig 5C ) but find differential viral accumulation within microchannel neurites . Since extracellular diffusion into the microchannels is restricted , viruses must access this compartment only via axoplasmic transport and are therefore intracellular . Both dynasore and EIPA administered at the time of inoculation abrogate viral accumulation in the microchannels ( Fig 5D ) . Dynasore also impacts viral accumulation when added at 2 hpi , although the effect was less pronounced . A strong positive correlation between the number of cell-associated viruses at the opening to and the number within individual microchannels is observed in untreated controls . Dynasore decouples N-compartment and microchannel accumulation of virus consistent with a block to particle uptake . By contrast , following EIPA treatment that correlation remains positive suggesting that EIPA delays , rather than blocks , endocytosis and transport of virus . Fluorescent transferrin ( Tfn ) and dextran ( Dex ) are commonly used markers of clathrin-mediated and fluid phase endocytosis respectively . To corroborate our inhibitor studies , we next determined whether rVSV CVS G AF647 are colocalized with Tfn or Dex during entry ( Fig 6A and 6B ) . In both neuronal populations , we observe a limited colocalization of virions with Dex in the N compartment and channel neurites ( Fig 6A–6C ) . By contrast , a third of incoming rVSV CVS G associate with Tfn in the N compartment ( Fig 6B ) , and this colocalization is enriched in the microchannels . This result suggests that internalization via clathrin shunts virus into long-range axoplasmic transport ( Fig 6B and 6C ) . The transportation bias for particles originating from clathrin-coated pits is particularly pronounced within DRG neurons where 87% of virions within the channels colocalize with Tfn versus 33% in the N compartment ( Fig 6C ) . Although fewer viral particles associate with Tfn in V SC channel neurites ( Fig 6B and 6C ) we note a similar enrichment from 31% in the N compartment to 58% in the channels . This observation is consistent with results showing greater sensitivity of infection to EIPA in the V SC neurons ( Fig 3B ) . However , transmission electron micrographs of rVSV CVS G uptake in V SC displays viruses exclusively associated with endocytic structures consistent with clathrin mediated endocytosis ( Fig 6D ) . This may suggest an off-target effect of EIPA in neurons which , by dysregulating Na+/H+ exchangers , may indirectly impact neuronal endocytic dynamics via disruption of the excitation properties of the plasma membrane [35] . Alternatively , 2hpi may be too early for the unambiguous identifications of macropinocytic structures in these cells . Furthermore , a caveat of the Tfn uptake within V SC neurons is the reported absence of transferrin receptors ( TfnR ) on the axons of mature motor neurons [36] . Axonal TfnR expression can be maintained by culturing of neurons in the presence of Tfn [37] and during axonal regeneration such as during outgrowth following axotomization inherent to the dissociation process [38] . Although the observation of Tfn uptake suggests that these primary cultures retain the receptor at some level , it is possible that a low expression of TfnR may result in an under-reporting of clathrin-mediated uptake in these neurons . To ascertain that viral particles are being transported , we performed live imaging studies in DRG neurons ( S1 Video and S2 Video; Fig 7 ) . Within the channels , we recorded single viral particles in the process of long-range axoplasmic transport . Many of these were transported concomitantly with fluorescent Tfn ( S1 Video ) . Within highly polarized neurons , fusion can occur locally at the site of uptake or following endosomal transport at intermediate locations along the axon or at the cell body . Live imaging data of cotransport with Tfn corroborates previously published findings that RABV virions are sorted into a long-range vesicular transport route with delayed acidification and release at the cell body [21–23] . To investigate the role of delayed fusion on productive infection we administered Bafilomycin A1 ( BAF A1 ) , an inhibitor of vesicular H+ ion pumps , to the N or S compartment to respectively block localized or delayed fusion . We monitored the impact of this selective treatment on infection ( Fig 8A–8C ) . Cells whose neurites were treated with BAF A1 displayed widespread infection in the S compartment despite presence of the inhibitor ( Fig 8A and 8B ) . Under these conditions , we also observed enhanced viral spread relative to untreated counterparts in both DRG and V SC neurons . In contrast , somal treatment with BAF A1 resulted in robust inhibition of infection ( Fig 8A–8C ) . This effect depended on administration of the drug early in infection: we observed lesser inhibition following treatment at 9–12 hpi . At this point incoming viruses from the N compartment have likely already fused at the cell body . We excluded potential cytotoxicity or off-target effects on the viral life cycle by assessing the effect of BAF A1 in non-compartmentalized culture . Post-entry administration at 2 hpi did not interfere with infection ( Fig 8D ) . We corroborated our inhibitor experiments by demonstrating that incoming fluorescently-labeled viral particles colocalize with LysoTracker in the cell bodies of DRG neurons ( Fig 9 ) . LysoTracker accumulates in acidified endosomes where the luminal pH is low enough to trigger fusion . No colocalization was observed between LysoTracker and virions within microchannel neurites or neurites proximal to the somatodendritic compartment ( Fig 9; S3 Video ) . We did not investigate colocalization of virions with LysoTracker at the growth cones or in distal neurites . Although we were unable to conclusively determine whether viruses enter acidified organelles prior to arrival in the cell body , the presence of acidified endosomes within the neurites is consistent with the possibility of fusion at earlier timepoints ( Fig 9 ) . Together , these results demonstrate that the majority of incoming virions are transported within endosomes to the cell body where acidification occurs . Fusion at the cell body is then a prerequisite for efficient infection .
In this study , we provide corroborating evidence of long-range axoplasmic transport of rhabdoviruses incorporating RABV G . Co-transport of virions with Tfn is consistent with trafficking within endosomes . Furthermore , exclusive sensitivity of infection to disruption of endosomal acidification in the S compartment demonstrates that viral fusion events leading to productive infection occur in the vicinity of the soma . Previous studies have reported long-range endosomal transport of intact vaccine strain RABV virions within differentiated neuroblastoma cells and DRG neurons [21 , 23] . Additionally , lentiviral vectors pseudotyped with G from the attenuated CVS-B2C strain have been tracked during retrograde endosomal transport in compartmentalized motor neuron culture [22] . We extend those earlier observations to the pathogenic CVS-11 G , and demonstrate that subsequent efficient establishment of infection is facilitated by pH-dependent fusion at the perikaryon . Previous work with the SAD vaccine strain of RABV provided evidence for transport within acidified endosomes [23] . This is in contrast to our observation that rVSV CVS G is exposed to acidic endosomal pH primarily at the cell body of neurons ( Fig 9; S3 Video ) , with a minority of particles in acidified endosomes in the microchannel neurites . Vaccine strains of RABV are characterized by their reduced neurotropism and neuroinvasiveness , which are dictated by their respective glycoproteins and may account for the difference between our observations . The site of imaging might also be a contributory factor in the apparent difference because studies of acidified endosome distribution in neurons suggest a high density of these structures in areas proximal to the growth cone , but a significant reduction at intermediate locations on the axon [39] . The increase in infection and spread following inhibition of endosomal acidification at the neurites in both DRG and V SC neurons suggests that some virus-containing endosomes engaging long-range axoplasmic transport may undergo acidification before delivery to the soma ( Fig 8 ) . The presence of RABV in acidified organelles at the neuronal termini or within distal neurites has been previously suggested based on colocalization between RABV antigens and LysoTracker at NMJs [20] , and cotransport of vaccine strain RABV and LysoTracker within DRG neurons [23] . Studies of tetanus neurotoxin trafficking suggest that sorting of certain cargoes into axonal carriers involves passage through intermediate acidified endosomes [40] . Although we did not independently verify the presence of infectious particles within acidified organelles in the distal neurites and growth cones , we cannot exclude that a similar sorting step occurs during transportation of RABV virions . If rhabdoviral particles traffic through analogous structures , exposure to the acidified environment may result in early fusion of a subset of incoming virions . Our results suggest that such early triggering of fusion leads to a delay or possibly abrogation of infection . Accordingly , if this early acidification event is circumvented–for example by BAF A1 inhibition of the vacuolar ATPase–viruses that would otherwise have fused prematurely now proceed along the transportation route and are subject to subsequent acidification events leading to infection . Although this observation is provocative it is important to emphasize that here we are studying infection initiated by delivery of the VSV RNP core and we cannot ignore the possibility of a RABV specific RNP transit to the cell body . The reported interaction between RABV P , and dynein LC8 [41 , 42] could facilitate RNP engagement of microtubules for retrograde transport . Mutation of the motif in P responsible for this interaction does not , however , abolish neuroinvasiveness of RABV in vivo [43 , 44] suggesting that direct transport of RNPs is not the primary mechanism of RABV axoplasmic transport . Previous studies investigated RABV uptake into non-neuronal [18 , 19] , neuroblastoma [17] or hippocampal neurons [16] . Our work extends these studies into DRG and V SC neurons , the first populations of neurons invaded by rabies in vivo . We identify clathrin-mediated endocytosis ( CME ) as the primary mechanism of productive RABV uptake . We base our conclusion on three observations: ( i ) susceptibility of infection and single particle uptake to disruption of dynamin; ( ii ) co-packaging and -transport of incoming virions with transferrin; and ( iii ) detection of rhabdoviral particles within coated pits by electron microscopy . Because co-transport with transferrin was assessed during long-range axoplasmic transport , we cannot exclude that rhabdovirus-containing endosomes fuse or coalesce with Tfn-positive endosomes following internalization at the plasma membrane . However , transmission electron micrographs of uptake into V SC neurons provide direct evidence for clathrin-dependent uptake as viruses associated with coated endocytic structures that resemble clathrin-coated pits . Ultrastructural data from transmission electron micrographs also allows us to infer additional structural and mechanistic similarities between rhabdovirus-containing endosomes in neuronal and epithelial cells . The clathrin-coated pits share the elongated profile characteristic of their incompletely coated counterparts observed in BS-C-1 cells for both VSV and RABV ( Fig 6D ) [19 , 45 , 46] . It is likely , therefore , that actin polymerization is a requirement for completion of envelopment and scission also from the neuronal plasma membrane . Sensory DRG neurons and motor neurons are both susceptible to RABV infection in the host . Due to the morphological and functional differences between these neuronal populations , we explored the possibility of non-identical uptake mechanisms for RABV based on the neuronal subtype . Our infectivity experiments using pharmacological perturbation of endocytic processes reveal that productive infection in either neuronal population is dynamin-dependent . Single particle experiments further implicate clathrin-mediated uptake as a major route of RABV endocytosis in both cell types . Accordingly , 90% of particles in DRG microchannel neurites , and 55% in V SC neurites are co-packaged with transferrin . We also identified some differences in uptake between the two neuronal populations . In V SC culture , a significant fraction of incoming particles do not colocalize with transferrin , or fluorescent dextran . This suggests that a fraction of particles enter in a manner independent of clathrin or macropinocytic uptake mechanisms . Such differences in uptake appear to be cell-type dependent and may be dictated by differential engagement of particular receptors . Lentiviral vectors expressing CVS-B2C G undergoing axoplasmic transport in motor neurons were found to colocalize with all three known RABV receptors: p75 neurotrophin receptor , neural cell adhesion molecule , and nicotinic acetylcholine receptor [22] . Studies of uptake of putative rabies receptors following engagement with endogenous ligands , crosslinking antibodies or toxins indicate that p75NTR , NCAM and nAChR internalize by different cellular mechanisms . p75NTR bound to neurotropic factors internalizes via clathrin-mediated endocytosis [47–49] . Antibody crosslinking of NCAM induces its internalization primarily by clathrin-mediated endocytosis with caveolae playing a secondary role [48 , 49] . Finally , nAChR uptake into filamentous invaginations from the plasma membrane is clathrin-independent [50 , 51] . These observations raise the possibility that RABV interaction with specific receptors may contribute to the endocytic sorting of the virus . It will therefore be of interest to identify which receptors are internalized at the plasma membrane in complex with RABV and to relate these interactions to establishment of infection in these two neuronal populations .
Device masters were manufactured by two-layer soft photolithography onto 3 inch mechanical grade silicon wafers for spin coating ( UW3MEC , University Wafers ) utilizing established methods [52] . Two negative photoresists were used: SU-8 2002 . 5 ( MicroChem ) for the 3 μm microchannel layer followed by SU-8 2050 ( MicroChem ) for the 100 μm culturing compartment layer . Photoresist was patterned by UV-crosslinking through custom 20 , 000 dpi transparency masks ( CAD/Art Services ) , processed and cured according to supplier’s instructions . Following a final hard cure at 150°C for 15 min , masters were treated with ( tridecafluoro-1 , 1 , 2 , 2-tetrahydrooctyl ) trichlorosilane for 45 min to facilitate removal of cured polydimethylsiloxane ( PDMS ) following moulding . Devices were cast by applying a 10:1 prepolymer:curing agent mixture of Sylgard 184 ( Dow Corning ) to the master and curing at 65°C for a minimum of 1 h . After curing and release from the master , PDMS devices were cut , and wells punched out with round biopsy punches . We irreversibly bonded devices to acid cleaned glass coverslips by oxygen plasma bonding in a 500-II Plasma Etcher ( Technics ) . Bonded devices were sterilized under UV in a biosafety cabinet for 10 minutes prior to consecutive overnight coatings with 300 μg ml-1 poly-D-lysine ( P7886 , Sigma-Aldrich ) dissolved in 2X borate buffer solution ( 28341 , Thermo Scientific ) and 10 μg ml-1 laminin ( L2020 , Sigma-Aldrich ) in sterile water . Pregnant Spraque-Dawley rats ( Taconic ) were a kind gift from C . Cepko . All animal work included in this study was approved by the Harvard Medical Area Standing Committee on Animals under protocol 428-R98 of the Institutional Animal Care and Use Committee ( IACUC ) of Harvard Medical School . Animals were housed and handled in accordance with the Guide for the Care and Use of Laboratory Animals . Euthanasia of pregnant Sprague-Dawley rats was performed by controlled exposure to carbon dioxide from compressed gas cylinders such that suffering and distress was minimized . Whenever possible , animals were kept in their housing cages during the procedure . Non-responsive animals were further subjected to cervical dislocation to exclude any possibility of accidental revival . Only once this sequence of procedures was completed did we remove the uterus and embryos from the carcass . E15 embryos were euthanized by removal from the amniotic sac and subsequent decapitation . Neuronal tissues were dissected from E14 . 5-E15 . 5 embryonic Sprague-Dawley rats ( Taconic ) . Dorsal root ganglia ( DRG ) were dissected , dissociated by trypsinization and cultured in Neurobasal media ( Gibco ) supplemented with B27 ( 1:50; 17504–044 , Gibco ) , β-nerve growth factor ( 100 ng ml-1; 450–01 , Peprotech ) , 5% fetal bovine serum ( FBS , Tissue Culture Biologicals ) , 2 mM glutamine ( 101806 , MP Biochemicals ) , 25 mM HEPES ( 0511 , AMRESCO ) pH 7 . 4 and 25 μg ml-1 β-D-arabinofuranoside ( AraC; C1768 , Sigma-Aldrich ) [53] . Ventral spinal cord neurons were dissected by adapting the strategy outlined for the harvest and culture of dorsal spinal cord commissural neurons [54] . Instead of harvesting the dorsal portion of the spinal cord , we retained the ventral portion for our cultures . Dissociation of the neurons was also carried out as detailed with the exclusion of the Opti-Prep purification step . V SC neurons were cultured in Neurobasal media ( Gibco ) supplemented as outlined by Leach et al . [55] . AraC treatment for selective kill-off of dividing non-neuronal cells was included and maintained in DRG media from first plating . For V SC culture AraC was applied after 48 h in culture . For preparation of compartmentalized cultures , we seeded dissociated neurons into the S compartment: 1 . 5 × 105 DRG or 1 × 105 V SC neurons were dispensed per device . A half-media swap was performed for both DRG and V SC culture following 2–3 days in culture; at this point , 25 μg ml-1 AraC ( Sigma-Aldrich ) was administered to the V SC cultures and maintained . Subsequently , media was periodically supplemented to counteract evaporation in culture . Experimental infections were carried out once adequate neurite outgrowth was observed in the N compartment: typically , from day 7 in culture for DRG and day 10 for V SC neurons . Devices were discarded following 12 days in culture . Neurons , mouse neuroblastoma Neuro-2a cells ( N2a; ATCC CCL-131; American Type Culture Collection , Manassas , VA ) , baby hamster kidney BSR T7/5 cells ( gift of U . J . Buchholz ) [56] , and African green monkey kidney BS-C-1 cells ( ATCC CCL-26; American Type Culture Collection , Manassas , VA ) were maintained at 37°C and 5% CO2 . Non-neuronal cells were cultured in Dulbecco’s modified Eagle medium ( DMEM; Corning ) supplemented 10% fetal bovine serum ( Tissue Culture Biologicals ) . N2a media was further supplemented with 2mM glutamine ( Sigma ) and 25mM HEPES pH 7 . 4 . rVSV eGFP and rVSV eGFP SAD B19 G were amplified , purified and maintained as previously described [45 , 46] . rVSV eGFP CVS G ( rVSV CVS G ) was generated by insertion of the CVS-11 glycoprotein coding region into MluI and NotI restriction sites in a modified rVSV eGFP ΔG backbone [57–59] . A pUC57 plasmid containing the cDNA of CVS-11 G ( Genbank: GQ918139 . 1 ) with flanking MluI and NotI sites was commercially synthesized by GenScript . A P0 stock of the virus was recovered by standard methods in BSR T7/5 monolayers [45] . Individual viral clones from the P0 stock were isolated by fluorescent focus assay in N2a culture , and further amplified in N2a monolayers . rVSV CVS G stocks were passaged and expanded in N2a monolayers . Infections were carried out according to standard technique [45] . At 24hpi supernatant and infected N2a cells were collected and subjected to 2 min sonication in a Branson 1510 ultrasonic cleaner ( Branson , Richmond , VA ) followed by 30 s vortex to release cell bound viruses . Cell debris was pelleted by centrifugation and the resultant virus supernatant purified by ultracentrifugation . Virus pellets were resuspended overnight in phosphate buffered saline ( PBS ) + 25 mM HEPES pH 7 . 4 + 50 mM EDTA . We sonicated the virus resuspension for an additional 2 min , followed by 30s vortex , immediately prior to a final gradient purification over a 15%-45% ( wt/vol ) sucrose gradient in PBS as previously described [45] . Viral titers were determined by fluorescent focus assay in N2a monolayers . We used established methods to label gradient-purified viral particles with 40 μg ml−1 Alexa Fluor ( AF ) -conjugated succinyl esters ( Molecular Probes , Invitrogen ) [45] . Titration of mock- or AF647-labeled virus preparations showed that dye conjugation had a negligible effect on infectivity: mock-labeled rVSV CVS G had a titer of 2 . 1 x 1010 ffu/mL compared to 1 . 5 1010 ffu/mL titer of the AF647-labeled virus . Viral proteins of gradient purified virions were separated by SDS-PAGE in a 10% polyacrylamide ( wt/vol ) and 0 . 13% ( wt/vol ) bis-acrylamide gel . Protein bands were visualized with SimplyBlue SafeStain according to manufacturer’s instructions . Viral protein amounts relative to N protein were determined using ImageJ ( U . S . National Institutes of Health , Bethesda , Maryland; http://rsb . info . nih . gov/ij/ ) . Neuronal cytoplasms in the N and S compartments were stained with calcein ( diluted 1:1000; C3099 , Molecular Probes ) or CellTracker ( diluted 1:500; C34552 Molecular Probes ) , and nuclei with NucBlue Live Cell Stain ( diluted 1:50; R37605 , Molecular Probes ) in Neurobasal for 30 min prior to inoculation with virus . Stains were washed once with Neurobasal following removal , prior to infection or imaging by direct fluorescent microscopy . The following chemicals were administered at the listed concentrations: 0 . 1 μM bafilomycin A1 ( BAF A1; 196000 , Calbiochem , EMD Chemicals ) ; 150 μM dynasore ( Sigma-Aldrich ) ; and 25 μM 5- ( N-ethyl-N-isopropyl ) amiloride ( EIPA; A3085 , Sigma-Aldrich ) . Infections were carried out exclusively in the N compartment . For N-compartment inhibitor treatment , culturing media in the S compartment was supplemented to replace evaporated liquid volume , and maintained throughout the experiment . For BAF A1 experiments where inhibitor treatment was also carried out in the S compartment , S compartment media was replaced with Neurobasal alone or BAF A1 diluted in Neurobasal . Neuronal culturing media in the N compartment was replaced with 30 μL of inoculation media . Inoculation media consisted of Neurobasal media alone or with inhibitor as indicated . 106 fluorescent foci forming units of virus were administered to the N compartment in 25 μL of inoculation media . Liquid volume differential between the S compartment and N compartment was optimized to eliminate diffusion of molecules across the channels . Inoculum was maintained for 2 h and then replaced with 55 μL treated or untreated Neurobasal media . For BAF A1 experiments where inhibitor was administered at 9–12 hpi , an additional media swap was carried out . Expression of eGFP was assessed at 26 hpi either by direct live fluorescence microscopy for DRG neurons or following fixation and immunofluorescence for V SC neurons . This end point was determined by monitoring appearance of eGFP and selecting a timepoint at which robust eGFP expression first occurs . In addition , we verified infection of neurons solely in the vicinity of the microchannel openings , consistent with primary infection without spread to secondary infection sites . For all co-uptake experiments neurons were first prestained with calcein or CellTracker as indicated . Infections were carried out , as described , in the N compartment with inocula containing Tfn conjugated to AF594 ( 50 μg ml-1; Molecular Probes ) , AF488-conjugated Dextran ( MW 10 000 , 1 μg ml-1; Molecular Probes ) or LysoTracker DND-26 ( 75 nM; Molecular Probes ) . Tfn and Dextran results were analyzed in both fixed and live samples . All LysoTracker images were collected by live confocal microscopy . For fixed experiments , N compartments were washed twice at 2 or 5 hpi with Neurobasal , and fixed with 2% ( wt/vol ) paraformaldehyde in PBS + 5% ( wt/vol ) sucrose . For live imaging , uptake experiments were carried out in devices bonded to FluoroDish glass bottomed culture dishes ( FD35-100 , World Precision Instruments , Inc . ) and imaged by high resolution spinning disk confocal microscopy . Neurons were fixed with 2% ( wt/vol ) paraformaldehyde in PBS + 5% ( wt/vol ) sucrose . Cell membranes were permeabilized with 0 . 2% Triton-X in PBS . Cells were consecutively stained with mouse monoclonal antibody against phosphorylated neurofilament H , SMI-31 ( 1:1000; NE1022 , Calbiochem ) , and AF-conjugated anti-mouse secondary antibody ( Molecular Probes , Invitrogen ) . When indicated , neuronal cells were detected by staining against Neuronal Nuclei ( NeuN ) with rabbit polyclonal antibody ( 1:500; ab104225 , Abcam ) and AF-conjugated anti-rabbit secondary antibody ( Molecular Probes ) . Nuclei were stained with DAPI ( 1:10 , 000; Molecular Probes ) . Devices processed for immunofluorescence underwent a final wash with PBS and were imaged in solution . Non-compartmentalized neurons , cultured on coverslips , were mounted onto slides with ProLong Diamond ( Molecular Probes ) . Devices were illuminated with a Mercury-100W mercury lamp ( Chu Technical Corporation ) and imaged using a Nikon Eclipse TE300 inverted microscope , outfitted with 4× Plan Fluor , 10× and 20× Plan Fluor objective lenses ( Nikon ) . Images were collected using a SPOT RT Monochrome camera ( Spot Imaging Solutions , Diagnostic Instruments Inc . ) and recorded with the manufacturer’s Spot 3 . 5 Advanced software . Devices were imaged using a Marianas system ( Intelligent Imaging Innovations ) based on a Zeiss observer microscope ( Carl Zeiss MicroImaging ) outfitted with a CSU-22 spinning-disk confocal unit ( Yokogawa Electric Corporation ) and a 63× ( Plan-Apochromat , NA 1 . 4; Carl Zeiss Microimaging ) objective lens . Excitation wavelengths were 491 nm for AF488 , 561 nm for AF594 , and 660 nm for AF647 . For three-dimensional acquisitions , the vertical position was manipulated in 0 . 3 μm increments using a PZ-2000 automated stage ( Applied Scientific Instrumentation ) . Live imaging experiments were carried out on a temperature controlled sample holder ( 20/20 Technology Inc . ; Wilmington , NC ) maintained at 37°C and 5% CO2 . Images were collected using a Photometrics Cascade II electron multiplication camera ( Photometrics ) . SlideBook 5 . 0 ( Intelligent Imaging Innovations ) was used to command the hardware devices , and visualize and export the acquired data . Subsequent image manipulation was conducted using ImageJ ( U . S . National Institutes of Health , http://rsb . info . nih . gov/ij/ ) . To visualize viral morphology , we deposited gradient purified rVSV CVS G virions onto carbon-coated copper grids and stained them with 2% phosphotungstic acid ( wt/vol ) in H2O ( pH 7 . 5 ) . To visualize viral uptake , we inoculated V SC neurons cultured on Aclar with rVSV CVS G at a multiplicity of infection ( MOI ) exceeding 1 , 000 for 2 h at 37°C . rVSV uptake samples were prepared by inoculating BS-C-1 cells with rVSV at an MOI of 1 , 000 for 15 min at 37°C . Samples were fixed and processed for ultrathin sectioning as previously described [45 , 60] . Virus particles and ultrathin sections of cells were viewed using a Tecnai G2 Spirit BioTWIN transmission electron microscope ( FEI ) .
|
Rabies virus is the causative agent of a generally fatal and incurable disease of the central nervous system ( CNS ) . Rabies lethality requires that the virus invade the brain , a feat accomplished by neuronal transmission from the site of infection to the CNS . Using cultures of peripheral neurons and chemicals that perturb specific cellular entry pathways we characterize the mechanism of rabies uptake . Using high resolution confocal microscopy , we visualize individual viral particles in the process of internalization and the establishment of infection by expression of a genetically encoded marker for infection . We show that clathrin-coated pits mediate internalization of the virus into endocytic vesicles that transport the virus to the cell body . We further demonstrate that release of the viral genomic core at the cell body is required to efficiently establish infection , and provide evidence that a subset of incoming virus particles fuse at non-productive sites prior to arrival at this site . This study extends the prior knowledge by identifying the entry mechanism and the site of fusion required for effective establishment of infection in neurons .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"cell",
"physiology",
"medicine",
"and",
"health",
"sciences",
"vesicles",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"cell",
"processes",
"microbiology",
"viral",
"structure",
"neurites",
"neuroscience",
"motor",
"neurons",
"viruses",
"rna",
"viruses",
"cellular",
"structures",
"and",
"organelles",
"neuronal",
"dendrites",
"endosomes",
"rabies",
"virus",
"animal",
"cells",
"medical",
"microbiology",
"microbial",
"pathogens",
"virions",
"endocytosis",
"cellular",
"neuroscience",
"lyssavirus",
"cell",
"biology",
"secretory",
"pathway",
"virology",
"viral",
"pathogens",
"neurons",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"organisms",
"cell",
"fusion"
] |
2016
|
Rabies Internalizes into Primary Peripheral Neurons via Clathrin Coated Pits and Requires Fusion at the Cell Body
|
HIV-1 can disseminate between susceptible cells by two mechanisms: cell-free infection following fluid-phase diffusion of virions and by highly-efficient direct cell-to-cell transmission at immune cell contacts . The contribution of this hybrid spreading mechanism , which is also a characteristic of some important computer worm outbreaks , to HIV-1 progression in vivo remains unknown . Here we present a new mathematical model that explicitly incorporates the ability of HIV-1 to use hybrid spreading mechanisms and evaluate the consequences for HIV-1 pathogenenesis . The model captures the major phases of the HIV-1 infection course of a cohort of treatment naive patients and also accurately predicts the results of the Short Pulse Anti-Retroviral Therapy at Seroconversion ( SPARTAC ) trial . Using this model we find that hybrid spreading is critical to seed and establish infection , and that cell-to-cell spread and increased CD4+ T cell activation are important for HIV-1 progression . Notably , the model predicts that cell-to-cell spread becomes increasingly effective as infection progresses and thus may present a considerable treatment barrier . Deriving predictions of various treatments’ influence on HIV-1 progression highlights the importance of earlier intervention and suggests that treatments effectively targeting cell-to-cell HIV-1 spread can delay progression to AIDS . This study suggests that hybrid spreading is a fundamental feature of HIV infection , and provides the mathematical framework incorporating this feature with which to evaluate future therapeutic strategies .
The course of HIV-1 infection is typified by three phases; acute infection characterized by a rapid viraemia peak ( 3–6 weeks post-infection ) followed by a rapid fall in virus levels , a stable chronic phase of variable length characterized by low level viraemia and slowly declining CD4+ T cell numbers , and a final stage ( Acquired Immune Deficiency Syndrome , AIDS ) characterized by multiple opportunistic infections and a rapid fall in CD4+ T cell count . The cellular and viral changes which drive each phase of this complex infection have been the subject of intense debate , in which mathematical models have played an important role in delineating HIV-1 pathogenesis and informing antiretroviral therapy [1–3] . The rich literature appertaining to mathematical modeling of intra host HIV dynamics has been reviewed several times recently [1 , 4–6] . Recent studies incorporate sophisticated models of immune selection [7 , 8] , as well as the formation of a latent reservoir of quiescent infected cells [9 , 10] . However the interplay of cell-to-cell spread and increased CD4+ T cell activation , that are likely to have profound influences on the progression of the disease have been hitherto little studied . Here we have addressed this and developed a unified model that can explain the complex progress of the infection in all its phases and its variable timescale . Such a unified model is important not only to understand the HIV-1 infection dynamics , but also to estimate the long term effects of therapeutic strategies on HIV-1 progression . In this paper , unless otherwise stated , “T cells” refers to CD4+ T cells . HIV-1 predominantly replicates in CD4+ T cells in vivo , and is now known to spread between T cells by two parallel routes . According to the classical model of HIV-1 spread , virus particles bud from an infected T cell , enter the blood/extracellular fluid and then infect another T cell following a chance encounter ( termed cell-free spread ) . Because diffusion of virus particles is much faster than cell migration , and there is extensive flow of blood and fluid , this mode of spreading can be characterized by a well mixed epidemic spreading model . In this scenario , the probability of infection for a particular cell will be proportional to the concentration of extracellular infectious virus . However , HIV-1 can also disseminate by direct transmission from one cell to another by a process of cell-to-cell spread . Two pathways of cell-to-cell transmission have been reported . Firstly , an infected T cell can transmit virus directly to a target T cell via a virological synapse [11–13] . Secondly , an antigen presenting cell ( APC ) can also transmit HIV-1 to T cells by a process that either involves productive infection ( in the case of macrophages ) or capture and transfer of virions in trans ( in the case of dendritic cells ) [11] . Whichever pathway is used , infection by cell-to-cell transfer is reported to be much more efficient than cell-free virus spread [14–16] . A number of factors contribute to this increased efficiency , including polarised virus budding towards the site of cell-to-cell contact , close apposition of cells which minimizes fluid-phase diffusion of virions , and clustering of HIV-1 entry receptors on the target cell to the contact zone [11 , 12] . Cell-to-cell spread is thought to be particularly important in lymphoid tissues where CD4+ T lymphocytes are densely packed and likely to frequently interact . Indeed , intravital imaging studies have supported the concept of the HIV-1 virological synapse in vivo [17 , 18] . Hybrid spreading is in fact a feature of other viral infections [19] , but is also shared in other “epidemic” scenarios such as spread of computer worms [20 , 21] , or of mobile phone viruses [22] . The mathematical analysis of hybrid spreading has received significant previous attention [22–25] . However , the importance of hybrid spread to HIV-1 dissemination and disease progression , has not been explored from a mathematical point of view . In this paper we develop a new mathematical model which incorporates the basic principles of previous host-centric models including a virus-dependent immune response [8] , viral latency and a progressive increase in cell activation [26 , 27] . Notably , the model additionally includes explicit terms for the two modes of virus spread , parametrised from experimental observation . The model faithfully replicates the overall three phase course of HIV-1 infection . The model predictions are consistent with both a set of longitudinal data ( viral load and CD4+ T cell count ) from a cohort of treatment naive HIV-1 infected patients and the results of the Short Pulse Anti-Retroviral Therapy at Seroconversion ( SPARTAC ) trial that aims to evaluate how the short-course antiretroviral therapy ( ART ) delays HIV progression [28] . The results of our study reveal the importance of two modes of HIV-1 spread , highlight the close link between cell-to-cell spread and cell activation in driving the progression of HIV-1 infection to AIDS and support early therapeutic intervention ( i . e . “test-and-treat” initiatives ) to delay disease progression in infected individuals . Since cell-to-cell spread is likely to present a considerable barrier to HIV-1 eradication , our data suggest that efforts to target this mode of viral spread whilst simultaneously manipulating CD4+ T cell activation may be a fruitful strategy to help control virus infection and halt progression to AIDS .
We here introduce a model of HIV-1 infection as depicted in Fig . 1A . We consider four distinct CD4+ T cell states: activated , uninfected susceptible ( S ) cells; activated and productively infected ( I ) cells; quiescent , uninfected ( Q ) cells; and quiescent latently ( L ) infected cells . The total CD4+ cell density ( N ) changes with time and is given by the sum of these four terms , i . e . N ( t ) = Q ( t ) + S ( t ) + I ( t ) + L ( t ) . The model can be described by an Ordinary Differential Equation ( ODE ) system ( Equation 1 ) and is illustrated in Fig . 1A . d Q d t = - γ Q + r S S - a N M N Q + b d S d t = - γ S S - r S S + a N M N Q - c I S N θ β 1 - S V β 2 + p N M - N N M S d I d t = - γ I I - r I I + a N M N L + c I S N θ β 1 + S V β 2 - κ I I + 0 . 1 N N M I d L d t = - γ L + r I I - a N M N L d V d t = - γ V V + g I ( 1 ) The density variables ( Q , S , I , L , V ) and parameters are defined in Table 1 . The densities are measured as numbers of cells or virions in a μl of blood/extracellular fluid . We set a density variable to zero when it drops to below 10−12/μl , accounting for the fact that when density of cells or virions drops to such low level , there is a high probability that it would die out ( density becomes zero ) . The default value of parameters , shown in Table 1 , are taken from the literature or estimated from clinical and experimentally observed data . The killing coefficient κ equals to its value in Table 1 when t ≥ D; otherwise κ = 0 when t < D . The production rate of new quiescent T cells from sources , such as thymus , within the human body is represented by b . Quiescent T cells are activated and become susceptible at a variable activation rate a ( NM/N ) , where a is the activation coefficient and NM is the density of T cells at which proliferation stops . The activation rate a ( NM/N ) increases as the total T cell density ( N ) falls ( caused by the HIV-1 progression ) . The detailed mechanism of how the activation rate increases with the progression of HIV-1 is still unclear , and may include increased rates of co-infections , danger signals from dying T cells or homeostatic regulatory loops . This term , a ( NM/N ) , here is an approximation , which encompasses the combined effects of all these different mechanisms . Quiescent T cells die at a rate of γ . Susceptible T cells turns into quiescent T cells at a rate r . They proliferate at a variable rate p ( 1 − N/NM ) , where p is the proliferation coefficient , N is the total T cell density , and NM is the T cell density at which proliferation stops . This variable proliferation rate is a reasonable approximation [29] to the real T cell proliferation process , based on evidence [2] that T cell proliferation rate is density-dependent and would slow as the T cell density becomes high . Susceptible T cells die at a rate γS . Susceptible T cells become infected through both cell-to-cell and cell-free infection . For cell-free infection , the number of newly infected T cells per unit period of time is β2 SV , where β2 is the infection rate of susceptible T cells by free virus . For cell-to-cell infection , we consider that the T cells are randomly moving , i . e . a T cell has an equal chance of contacting any other T cells . Let c represent the number of effective contacts each T cell makes in a unit period of time . Then the number of contacts made by all infected T cells per unit period of time is cI . Among those contacts , S/N are contacts with a susceptible T cell that could potentially end up with a new infection . Let θ represent the Synapse rate: the average probability that two T cells form a virological synapse once they have made an effective contact . Then the number of newly infected T cells through cell-to-cell infection per unit time period can be represented as cI ( S/N ) θβ1 , where β1 is the cell-to-cell infection rate when an infected T cell and a susceptible T cell form a synapse . In reality , cell-to-cell transfer occurs locally involving only the infected cell and its immediate neighbours . The model abstracts this process by averaging infection over all cells . In practice , local effects will only distort this average when target cells in the vicinity of an infected cell become limiting . This limit seems unlikely to be reached except very late in infection , given that infected cells continue to migrate , albeit at a slower rate ( personal observations and [17] ) , and uninfected target cells continue to migrate into the vicinity of an infected cell . More complex spatial models will be required , however , to understand the detailed anatomical distribution of HIV-1 infected cells over time . The amount of cell-to-cell transfer in the model depends on the number of infected T cells I and the proportion of susceptible cells S/N at any given time . Since N , the total density of CD4+ cells , is not held constant , but in fact declines over time , cell-to-cell spread becomes increasingly effective as HIV-1 progresses . The model does not explicitly distinguish antigen-presenting cell to T cell ( APC/T ) from T cell to T cell ( T/T ) transmission . APC/T transmission may potentially be most important very early in establishing infection [38] , a process which is not examined in detail in this model . These two types of transmission are in fact both likely to occur most frequently and efficiently in the microenvironment of an APC/T cluster , where APC/T interactions lead to T cell activation , and hence favour also T/T interaction . Furthermore , there are very few quantitative estimates of the parameters of APC/T interaction in vivo . The incorporation of an additional cell type is therefore unlikely to have a major effect on the model behaviour , but would add significantly to model complexity and uncertainty . The cause of T cell death in HIV-1 infection continues to be controversial , and probably includes several effects including lysis of infected cells by effector cells such as CD8 T cells and NK cells , apoptosis/pyroptosis and bystander death [27] . Non-immunological death of infected T cells is represented by a death rate of γI . And we use the term κ I I + 0 . 1 N N M to model the death of infected T cells by the cellular immune response . κ is initially 0 and changes to a higher value in Table 1 when the cellular immune response kicks in D ( default value: 30 ) days after the initial infection . The term I I + 0 . 1 captures the relationship between the strength of the immune response and the density of infected CD4+ T cells [8] . The term N/NM captures immune exhaustion caused by HIV-1 infection . It falls from around 1 before infection towards 0 as HIV-1 progresses ( because NM is a constant and N , the total density of CD4+ T cells , gradually declines with the HIV-1 progression ) . Infected cells return to a quiescent state , and become latent , at rate rI . Latent cells die and are activated ( i . e . becomes infected cells ) at the same rates as quiescent cells . Infected T cells release free viruses at rate g . Free viruses die at a rate of γV . The abortive infection of quiescent cells is not considered in this simplified model , similarly to most previous modelling studies [1 , 3 , 39 , 40] . HIV-1 immune escape mutants [41] are not directly modelled in this paper but their effects on degrading cellular immune response are reflected in the immune exhaustion in our model . The numerical solutions of each of the variables are shown in Fig . 1C . Fig . 1B shows the combined CD4+ T cell counts ( N ) and virus load ( V ) , which are measured routinely in the clinic to monitor HIV-1 infection . Notably , the qualitative behaviour of the model accurately reflects the three main phases of disease that are observed clinically . The model reproduces an acute infection phase , where the virus replicates rapidly ( reflecting the absence of any pre-existing adaptive immunity ) , peaks and then returns to a low level by approximately 5 weeks . This metastable level of virus represents the clinical “set-point” . Virus then remains stable for a prolonged period ( note interruption and change of scale in x axis ) , during which time T cells decline very slowly . Finally , T cell numbers start to drop faster , and viral loads rise . The model calculation is stopped when CD4+ level reaches 200 cells/μl . It is important to note that the term cI ( S/N ) θβ1 we introduced for cell-to-cell spread is mathematically different from the classic term β2 SV which captures cell-free spread . N , the total CD4+ number , is not a constant but decreases as a function of time . Indeed , if N in the cell-to-cell spread term is fixed , disease progression is not observed in the model ( S1 Fig ) . Furthermore , N cannot be expressed as a simple function of V . In Fig . 1B the same value of V ( free virion density ) on each side of its peak around t = 30 days corresponds to two different N ( CD4+ T cell density ) values . In addition , I , the number of infected cells , is not always proportional to V , especially early in infection ( S2 Fig ) . Thus the term for cell-to-cell spread cI ( S/N ) θβ1 shows different dynamics from the classic cell-free term β2 SV . Most of the parameters of our model , and especially those determining HIV infectivity via cell-to-cell or cell-free spreading , were obtained from experimental observations ( Table 1 ) . It was therefore important to test that the model with this parametrization accurately fits real clinical data sets . We therefore evaluated our model against a set of longitudinal T cell count and virus load measurements obtained from a cohort of HIV-1-infected individuals who were recruited following clinical presentation with symptomatic acute HIV-1 infection and followed over time with serial measurement of plasma viral RNA levels and circulating T cell counts . The subjects selected for inclusion in this study all chose not to receive antiretroviral treatment in acute or early infection , and remained untreated until progression towards AIDS , evidenced by a substantial decline in their circulating CD4+ T cell count ( See Materials and Methods ) . We use our model to theoretically reproduce the HIV-1 infection courses in the data . The values for HIV infectivity were fixed , as derived from the literature or our own observations ( Table 1 ) . Values for five parameters ( Q0 , S0 , NM , κ and D ) describing the characteristics of the immune response were chosen for each patient to minimise the error of the predicted quasi-stable level of T cell counts ( Ns ) and viral load ( Vs ) , and the time of progression to AIDS ( tA ) . All other values remained fixed at the default values in Table 1 . The predicted progression results are compared against the actual measurements in Fig . 2 and S2 Table . The predicted Vs and tA for each patient were negatively correlated ( correlation coefficient = −0 . 46 ) , in agreement with the well-established relationship between these two clinical values . Remarkably , the model can fit all patients by modifying the five immune-relevant parameters over a narrow range . Furthermore , the parameter values which gave the best results for the patients ( see S1 Table ) are all very close to those in Table 1 , which were derived independently from experimental measurements . Having confirmed that the model gives realistic estimations and predictions of real clinical data , we investigated the behaviour of the model in more detail . The role of the two spreading routes was further examined by systematic variation of the cell-to-cell infection rate , β1 , and the cell-free infection rate , β2 . The predicted outcome of infection are shown in Fig . 3 . When either route is abolished , infection is blocked completely; T cell level returns to normal and virus is cleared after the cellular immune response kicks in . If cell-to-cell spread is removed from the model ( β1 = 0 ) even a doubling in cell-free infection rate does not result in infection progression . In contrast , a doubling of cell-to-cell infection rate increases the set-point of viral load , and greatly speeds up the progression of infection even in the absence of cell-free infectivity . Thus the model suggests cell-to-cell spread may be an important force in allowing virus to establish infection in lymphoid tissue [38] . In the context of the model , the transition from phase 1 ( acute ) to phase 2 ( stable chronic ) is driven by a balance between several processes , including viral spreading through two parallel modes , and the cellular immune response , i . e . killing of infected cells as the cytotoxic CD8+ T cell response becomes active . Paradoxically , in the stable chronic phase , the activation of T cells , which is the hallmark of adaptive immunity and is aimed at protecting the host , in fact contributes to the persistence of HIV-1 . The role of CD4+ T cell activation is explored in Fig . 4A . In this model , the rate of T cell activation a ( NM/N ) increases as the number ( N ) of T cells falls , which can be considered to represent a type of homeostatic regulation reinforcing immunological activity relevant to the progressive damage of the immune system and its consequences . In the absence of this feedback ( i . e . when activation rate is fixed ) , HIV infection would not progress to AIDS after the onset of the cellular immune response . In contrast , if the activation rate is doubled , then infection progresses significantly faster to AIDS . These results confirm and extend the findings of DeBoer and Perelson [29] , which suggested an increasing rate of cellular activation was important in establishment of chronic infection and progression to AIDS . The results are also consistent with evidence that non-pathogenic SIV infection in the natural host species results in viral replication in the absence of chronic immune activation and no AIDS [42] . Fig . 4B depicts the number of CD4+ T cells newly infected via either cell-to-cell spread or cell-free spread as the infection progresses . The model predicts that cell-to-cell transfer becomes increasingly dominant as the total number of CD4+ T cells falls , the proportion of susceptible cells rises ( Fig . 4B inset left y axis ) and the strength of immune response falls ( because of immune exhaustion , see Fig . 4B inset right y axis ) . Modelling can help evaluate the long term effects of different treatments on HIV-1 progression . Once the start time , duration , and effectiveness against two modes of HIV-1 spread are known for a treatment , its effects on HIV-1 progression can be evaluated by the model . To validate the model’s ability to evaluate HIV-1 treatments , we use it to theoretically reproduce the results of the Short Pulse Anti-Retroviral Therapy at Seroconversion ( SPARTAC ) trial [28] . The SPARTAC trial aims to evaluate how the short-course antiretroviral therapy ( ART ) delays HIV progression . The patients ( 366 in total ) who participated in the trial were randomly assigned to three groups: standard care , 12-week ART treatment , and 48-week ART treatment . For these three groups of patients , the primary end point tp ( defined as when CD4+ count ≤ 350 cells/μl or the start of long-term ART ) on average reached 157 weeks ( standard care ) , 184 weeks ( 12-week ART ) , and 222 weeks ( 48-week ART ) after randomisation . Randomisation is the time when the trial starts . We first estimate the trial time points ( randomisation , start of the treatments , and primary end points tp ) in terms of days after the initial infection ( See Materials and Methods ) . We then use the average CD4+ count and virus load of all patients at randomisation , and the primary end point of the patients in standard care group to fit the five immune-relevant model parameters ( Q0 , S0 , NM , κ and D ) . These fitted parameters represent an average patient in the trial . We then evaluated the effects of 12 and 48 week therapy using the model . We assumed that the therapy was 100% effective against cell-free transmission , but we evaluated the model for both 100% and 50% efficiency against cell-to-cell transmission , since cell-to-cell transmission has been reported as being more resistant to some forms of therapy [43 , 44] . Both modalities reproduced the observed effects of therapy well , and the model results were robust to changes in the efficiency ( Table 2 ) . The model therefore not only fits known data sets ( standard care group ) but also accurately predicts the effects of new treatment regimes on two independent patient groups ( 12-week and 48-week ) . We therefore further explored the sensitivity of the model to perturbation as a function of treatment starting time ( Fig . 5 ) . The “treatment” lasts for 30 days , during which both cell-free and cell-to-cell infection are completely blocked . Once “treatment” is finished , two modes of HIV-1 infection resume . Early treatment in this model ( 3 days after infection , i . e . post-exposure prophylaxis ) leads to no decline in CD4+ T cell density , and no chronic infection phase . The same treatment applied when T cell density reaches the levels ( 500 CD4+ T cells/μl and 350 CD4+ T cells/μl ) at which the World Health Organization recommends ART initiation [45] is followed by a rapid virus rebound after the treatment stops , and the disease progresses according to its normal course . Thus , as HIV-1 progresses it becomes increasingly difficult to control infection in this model . Finally , we looked at the interactions between treatment starting time , activation rate and efficiency of therapy against cell-to-cell spread ( Fig . 6 ) . In general , increased efficiency of therapy and earlier treatment both prolonged time to progression to AIDS . However , the effects of altered activation depend in a complex way on the context of the intervention . Blocking activation early is beneficial , since it will reduce the number of susceptible cells HIV-1 can infect; while blocking activation late , when the latent HIV-1 reservoir has been established , will prevent latent HIV-1 from being activated and eradicated . Increasing cellular activation , which has been proposed as a means of flushing out the latent reservoir [46] , can be effective in increasing time to AIDS when given in the context of efficient anti-viral therapy , but can shorten the time if concomitant anti-viral therapy is incomplete . This is because increasing cellular activation increases both the number of susceptible cells ( activated from quiescent cells ) and the number of infected cells ( activated from latent cells ) at the same time . Thus if it is used together with an effective anti-viral therapy , the latent HIV-1 reservoir will be flushed out and killed by the antiviral drugs . But if the concomitant anti-viral therapy is not efficient enough to clear the increased number of infected cells , the spread of HIV-1 will speed up .
The aim of this study was to develop a model that reduces the enormous biological complexity of the course of HIV-1 infection to a few well-defined equations but nevertheless retains the main dynamic features of the disease . In particular , the model was required to predict the evolution of a long lived metastable state of low level viral infection , which ultimately breaks down to uncontrolled viral growth and a precipitous fall in CD4+ T cells , two hallmarks of AIDS . The model incorporates both an immune response , which is believed to be a major factor limiting viral expansion by killing of infected cells , and the formation of long lived latently infected cells , which are believed to play an important role in limiting the long-term effects of antiviral therapy . The key distinguishing features of the model are that it incorporates explicitly cell-to-cell spread of virus as well as classical spread via cell-free virus . The motivation of building such a model was to investigate the role of these two modes of spread in determining the outcome of infection , and thus complement the limited in vivo experimental and clinical data available addressing this question . A fundamental prediction of the model is that , given the known experimentally derived parameters of cell-to-cell and cell-free spreading , the two modes of spread complement each other and both make important contributions to disease progression . In addition , it is clear that there is a close relationship between the proportion of activated T cells , cell-to-cell spread and disease progression . Specifically , cell-to-cell spreading is strong when the percentage of activated , and therefore susceptible cells is high in the population , since an infected cell is then more likely to encounter an effective target ( a susceptible cell ) to infect . When the percentage of susceptible cells is low , infected cells will mostly encounter quiescent/resting cells that will provide ineffective targets . These conditions may occur both early and late in HIV-1 infection . The site of infection itself ( for example the vaginal mucosa ) may contain a large proportion of activated T cells some of which may be interacting with APC , particularly if there is a concomitant sexually transmitted infection . The well-documented association between HIV-1 infection and other mucosal infections may therefore reflect the key importance of a high concentration of APC and activated T cells in early transmission of virus [47] . There is also convincing evidence that gut associated lymphoid tissue is a major site of viral replication early on in disease [48 , 49] . This tissue is characterized by an unusually large proportion of T cells capable of supporting HIV-1 replication , presumed to result from chronic exposure to the gut microbiome . Since cell-to-cell spread is much more efficient , and under these conditions the number of activated target cells are not limiting , our model predicts that gut lymphoid tissue would provide an ideal microenvironment for rapid propagation of HIV-1 , at least until the majority of target cells are infected or die . Further development of the basic model , to allow heterogeneity associated with different anatomical compartments would allow this prediction to be tested directly . The model also predicts a dominant role for cell-to-cell spread in the late phase of HIV-1 . As in the scenario early in infection in the gut lymphoid tissue , our model predicts that late stage disease will be associated with a large number of infected cells , combined with a large proportion of activated target cells . A high number of infected cells is a simple corollary of the very high levels of free virus late in infection in the absence of treatment . There is also substantial experimental evidence for increased immune activation in the late phases of HIV-1 [26] , and indeed this has been proposed as an important contribution to pathogenesis . Thus cell-to-cell spread is likely to become the dominant mode of transmission in the late stages of HIV-1 . This may be important in light of recent data showing that the different components of current HAART display variable efficacy against cell-to-cell spread [43 , 44] . Interestingly , the hybrid spreading mechanism employed by HIV-1 is reminiscent of those of some computer worms such as Red Code II and Cornficker [20 , 21] , which allocate their resources between probing for susceptible target computers in local area networks and globally across the internet . Similarly to HIV-1 , local interactions have a high chance of success but access only a limited number of targets while global spread targets a much larger number of targets with a much lower probability of success . Modelling studies have shown that this hybrid spreading is required to explain the large outbreak of such worms on the Internet [25] . It is tempting to speculate that hybrid spreading may contribute to the pathogenicity and dynamics of infection of other viruses that employ parrallel spreading mechanisms , for example Hepatitis C virus [19] . The current model incorporates some simplifying assumptions . Nevertheless , the model does provide some insights into the effectiveness of therapy at different stages of disease . Specifically , the model strongly supports the hypothesis that interfering with viral infection early in HIV-1 progression is likely to have a major impact on the subsequent progress of the disease . Interfering with viral transmission is predicted by our model to be much more effective early in HIV-1 infection . The clinical decision about when to start therapy remains a matter of debate . Current WHO guidelines suggest commencing treatment at CD4+ T cell density of > 350cells/μl and < 500cells/μl [45] . However , studies exploring much earlier commencements of treatment , have claimed increased efficacy [28 , 50] . For example , the recent “Short Pulse Anti-Retroviral Therapy at Seroconversion” ( SPARTAC ) trial , demonstrated a long term clinical benefit of a limited period of ART soon after seroconversion [28] . Our model accurately predicted the results of the SPARTAC trial providing further support for the model’s generality and robustness . Anti-viral therapies that specifically target cell-to-cell spread are not currently available , but are clearly important therapeutic goals [44] . A number of previous studies have proposed combining antiviral therapies with therapies that either limit CD4+ T cell activation ( thus reducing the number of target susceptible cells ) or increased T cell activation , thus flushing out residual latent cells . These approaches have not given clear cut clinical benefits [51 , 52] . Indeed our model suggests that the outcome of such manipulation of cellular activation will be critically dependent on the time at which it is administered , and the efficiency of concomitant antiviral therapy . Targeted suppression of CD4+ T cell activation , in combination with antiviral therapies may nevertheless offer a useful approach , if used early on in infection . Mathematical models provide an important tool for understanding and predicting the course of natural HIV-1 infection that complements clinical studies . The most appropriate therapeutic regimens for HIV therapy continue to remain the subject of disagreement and debate . In addition , many new therapeutic modalities aimed at achieving viral eradication , such as the HDAC inhibitors [46] , or therapeutic vaccination [53] are being proposed . However , testing new treatment regimens is a costly and time consuming task , and the logistic challenges and expense of running clinical trials to evaluate and compare treatments remain a major bottleneck to translational advances in HIV therapy . Mathematical models have proved of value in the past , but have suffered from omitting important biological processes , thus compromising their ability to accurately recapitulate clinical observations . Our model explicitly incorporates cell-to-cell transmission , and changing rates of cellular activation , two processes that are known to be a key feature of HIV infection . With increased sophistication , and hence ability to accurately model the known biological drivers of disease progression , mathematical models can become increasingly important in preclinical testing of modified or novel HIV therapies . The model developed here provides specific predictions which emerge from the close links between CD4+ T cell activation and cell-to-cell spread , and their combined contribution to both early and late phases of HIV-1 . These predictions emphasize the potential benefits of early or prophylactic treatment with antiretroviral therapies , and suggest that drugs with the ability to effectively block cell-to-cell spread may provide significant therapeutic benefit in long term management or eradication of HIV-1 infection .
Individuals acutely-infected with HIV-1 were recruited at the Mortimer Market Centre for Sexual Health and HIV Research ( London , UK ) . Subjects were mostly male Caucasians who presented with symptoms of acute retroviral illness . Patient viral loads and CD4+ T cell counts were measured longitudinally at serial time-points following infection using standard clinical tests . All subjects were offered anti-retroviral treatment at diagnosis . The subjects selected for inclusion in the study all chose not to receive anti-retroviral therapy in acute or early infection , and remained untreated until disease progression , evidenced by a substantial decline in their circulating CD4+ T cell count occurred . Patients provided written informed consent for study participation . Study approval was obtained from The National Health Service Camden and Islington Community Local Research Ethics Committee . There were 39 patients in total in the data from Mortimer Market Centre for Sexual Health and HIV Research ( London , UK ) . For this study , we focused our analysis only on patients with more than ten data points for both HIV load and CD4+ measurements ( 29 out of the 39 patients ) . We also excluded from the analysis a further 12 patients who showed no overall decrease in CD4+ count , or no increase in viral load at later timepoints . The focus of the model described above is to capture the “typical” characteristics of HIV infection , which include that CD4+ count falls in general , and viral load increases in general as infection progresses . It is widely accepted that in some patients ( for example the so-called “elite controllers” ) viral load remains low or undetectable and CD4+ count remains unchanged for long periods . The mechanisms responsible for these phenomena are still incompletely understood . The current model does not attempt to incorporate any such mechanism , and this group of patients was therefore not included in the study . Further elaboration of the current model to include additional features of viral control will be informative in helping to understand such patients . The identifiers for the remaining 17 patients are: MM1 , MM4 , MM8 , MM9 , MM12 , MM13 , MM23 , MM24 , MM27 , MM33 , MM39 , MM40 , MM42 , MM43 , MM45 , MM57 , MM60 . According to [28] , the median interval between seroconversion and randomization ( start of trial ) was 12 weeks . The exact average time of seroconversion for patients in the SPARTAC trial is not directly available from [28] . We assume it is 2 weeks after initial infection , as seroconversion normally happens within a few weeks after HIV-1 infection . Then the time of randomization ( start of trial ) can be estimated as 7 ( 2 + 12 ) = 98 days after infection . The treatment is estimated to start 3 days after randomization [28] , i . e . 98 + 3 = 101 days after infection . And the primary end points for patients in standard care , 12-week ART , and 48-week ART groups are respectively: 98 + 7 × 157 = 1197 days , 98 + 7 × 184 = 1386 days , and 98 + 7 × 222 = 1652 days after infection . Calculation results in this paper are obtained using LeoTask [54] , a parallel task running and results aggregation framework that we have developed for computational research . The framework and an executable programme that implements our HIV-1 model are both freely available at http://github . com/mleoking/leotaskapp .
|
The ability to spread using more than once mechanism , named hybrid spreading , is a ubiquitous feature of many real world epidemics including HIV and Hepatitis C virus infection in vivo , and computer worms spreading on the Internet . Hybrid spreading of HIV is well documented experimentally but its importance to HIV progression has been unclear . In this paper , we introduce a mathematical model of HIV dynamics that explicitly incorporates hybrid spreading . The model output shows excellent agreement to two sets of clinical data from a treatment naive cohort and from the Short Pulse Anti-Retroviral Therapy at Seroconversion trial . The model demonstrates that hybrid spreading is an essential feature of HIV progression , a result which has significant implications for future therapeutic strategies against HIV .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
Hybrid Spreading Mechanisms and T Cell Activation Shape the Dynamics of HIV-1 Infection
|
The Toxin Complex ( TC ) is a large multi-subunit toxin first characterized in the insect pathogens Photorhabdus and Xenorhabdus , but now seen in a range of pathogens , including those of humans . These complexes comprise three protein subunits , A , B and C which in the Xenorhabdus toxin are found in a 4∶1∶1 stoichiometry . Some TCs have been demonstrated to exhibit oral toxicity to insects and have the potential to be developed as a pest control technology . The lack of recognisable signal sequences in the three large component proteins hinders an understanding of their mode of secretion . Nevertheless , we have shown the Photorhabdus luminescens ( Pl ) Tcd complex has been shown to associate with the bacteria's surface , although some strains can also release it into the surrounding milieu . The large number of tc gene homologues in Pl make study of the export process difficult and as such we have developed and validated a heterologous Escherichia coli expression model to study the release of these important toxins . In addition to this model , we have used comparative genomics between a strain that releases high levels of Tcd into the supernatant and one that retains the toxin on its surface , to identify a protein responsible for enhancing secretion and release of these toxins . This protein is a putative lipase ( Pdl1 ) which is regulated by a small tightly linked antagonist protein ( Orf53 ) . The identification of homologues of these in other bacteria , linked to other virulence factor operons , such as type VI secretion systems , suggests that these genes represent a general and widespread mechanism for enhancing toxin release in Gram negative pathogens .
Photorhabdus are Gram-negative members of the Enterobacteriaceae that live in a symbiotic association with entomopathogenic nematodes which invade and kill insects . Upon insect invasion the nematode regurgitates the bacteria which grow within the insect open blood system releasing a plethora of toxins and insecticides , to kill the insect and protect the cadaver from invading microbes , scavengers and saprophytes [1] . One major class of secreted insecticidal toxins are the Toxin Complex ( TC ) proteins [2]–[3] . They constitute large multimeric protein complexes that have been shown to exhibit oral toxicity to a range of insects including crop pests [4]–[5] . These complexes consist of three protein subunit types; homologues of TcdA TcdB and TccC proteins , from here on referred to simply as A , B and C-subunits respectively [6] . These protein subunits themselves are large with the TcdA1 , TcdB1 and TccC5 being 2517 aa ( 283 kDa ) , 1477 aa ( 165 kDa ) and 939 aa ( 105 kDa ) respectively . Recent work on a Xenorhabdus nematophilus TC suggests that the A ( XptA2 ) , B ( XptB1 ) and C ( XptC1 ) subunits are in a 4∶1∶1 stoichiometry respectively [7]–[8] . There is now reasonable evidence to indicate that the A subunit forms a cage-like tetramer of around 1120 kDa which associates with a tightly bound 1∶1 sub-complex of B and C . The A subunits can bind the host cell membranes of insect brush border cells and form a pore facilitating the entry of the BC sub-complex [8]–[9] . The toxic activity of the complex resides in the C-terminal region of the C subunit . The TCs represent a potential target to augment the successful B . thuringiensis crystal toxin crop protection technology and have been the subject of significant investment by the agrochemical industry . The availability of a large number of bacterial pathogen genome sequences has revealed that TC toxins are not restricted to Photorhabdus and Xenorhabdus sp . , but are in fact widely distributed [2] . This includes Gram-negative human pathogens such as Yersinia and Burkholderia and a Gram-positive insect pathogen such as B . thuringiensis strain IBL200 ( accession NZ_ACNK01000119 ) . Although the role of these TC homologues in most pathogens remains obscure , recent work in Yersinia pseudotuberculosis has suggested that homologues of these toxins have been adapted to act upon the mammalian gut [10] . They have also been implicated in mammalian gut colonisation in at least one strain of Y . enterocolitica [11] . Some progress has been made in TC research including the reconstitution of TC function through heterologous expression of the three essential subunits , A , B and C , in E . coli [3] , [8] , [12]–[13] , expression of partial activity in transgenic plants [14] and the analysis of TC homologues from other species of bacteria [10] , [15] . In addition TC toxins have also been heterologously expressed in Enterobacteria species which associate with termites as a control strategy [16] . Most recently Lang et al demonstrated the mode of action of certain C-subunit C-terminal domains in the ADP-ribosylation of actin and RhoA [9] . Nevertheless , efforts in understanding the biological context of TC have been hampered by an incomplete understanding of how Gram-negative bacteria are able to assemble and secrete such large multimeric protein complexes . The tc gene homologues are encoded at several different loci in the Photorhabdus genome and in strain Pl W14 , two of these loci , tca and tcd were shown to be responsible for oral toxicity to Manduca sexta larvae [4] , [17] . The tcd locus is a large pathogenicity island ( pai ) containing multiple homologues of the A , B and C-subunit genes in tandem [18] . Previously we used RT-PCR and western blotting to demonstrate that Pl W14 expresses both the tca and tcd loci genes during insect infection [19]–[20] . Previous work has shown that the A and B+C-subunits are capable of exhibiting oral toxicity independently when expressed at high levels , yet nevertheless assemble into a far more potent large multimeric complex [8] , [21] . We noted that strains belonging to the species P . luminescens could be classified into two distinct biotypes , some produced Tcd which remained attached to the cell surface ( e . g . strain Pl TT01 ) , while others also released it into the surrounding medium ( e . g . ; Pl W14 ) ( Fig . 1AB ) . We hypothesized that there would be specific genetic factors facilitating this variation in TC deployment . We have used comparative genomics between a strain that releases Tcd and a strain that does not , together with a cosmid based E . coli heterologous expression system to investigate this hypothesis . Here we present evidence that toxin secretion is enhanced by a lipase ( Pdl1 ) , the activity of which is controlled by a small tightly linked antagonist protein ( Orf53 ) . Homologues of these genes can be seen to be linked to many virulence loci in other pathogens and we suggest that this work represents the first characterised example of an important and widespread secretion enhancement mechanism .
The c1AH10 was identified from a Pl W14 cosmid library [18] by aligning cosmid end sequences to the tcd pai sequence using SeqMan . The E . coli c1AH10 clone was tested for oral toxicity as described below . Insertional mutagenesis was performed with EZ::TN<TET1> transposon ( Epicentre ) to generate and characterise an insertion mutant sub-library as previously described [22] . Briefly , insertion mutant cosmid clone colonies were picked and DNA was prepared on a RoboPrep plasmid preparation robot ( MWG Biotech ) . The insertion site of the transposon in each mutant was determined by sequencing out from the transposon with a transposon specific primer using an ABI3700 nucleotide sequencer ( Applied Biosystems ) . Sequences were assembled onto the c1AH10 template sequence by using the LASERGENE software package ( DNASTAR , Madison , WI ) allowing the transposon insert sites to be determined . The pdl1 , orf54 and orf53 genes were amplified from P . luminescens strain W14 genomic DNA using rTth DNA polymerase ( Applied Biosystems ) . Polymerase chain reaction ( PCR ) conditions were 1 . 2–1 . 6 mM magnesium acetate , 2 mM each dNTP and 1 mM each primer . Thermocycling was performed as follows: 93°C for 30 s; 55°C for 30 s and 68°C for 3 min , for 30 cycles and a final 68°C incubation for 10 min . PCR primers , used for cloning into pET-28a ( + ) , pCDF-1b ( Novagen ) , and the arabinose inducible expression plasmid pBAD30 were designed to include unique restriction sites for subsequent cloning . The primer sequences ( 5′ to 3′ ) used for cloning the pdl1 , orf54 and orf53 genes into pBAD30 were as follows: For His-tagged cloning , pdl1 and orf54 were initially cloned into pET-28a ( + ) as His-fusion proteins and then subsequently PCR amplified from this template as a His-fusion for cloning into pBAD30 . The primer sequences ( 5′ to 3′ ) were as follows: W14PDL-XbaIf: ATTcTAgaGGAAAGAGTATCAATGAGC; 28a-r-XhoI-pdl: ATATATCTCGAGTACAGACAGTTCCTGT; 30-r-SphI-PET28: ATATATGCATGCTAGTTATTGCTCAGCGG; W14O53-XbaIF: ATTCTAGAATCTAAATGCCAACATGAG; 28a-r-XhoI-O53: ATCTCGAGCTTTATTTTCCAGTAGTC . Following PCR , the products were purified ( Millipore Montage PCR column as per instructions ) , cut with the appropriate restriction enzyme , and then re-purified prior to ligation and cloning . Plasmid DNA from pET-28a ( + ) , pBAD30 and pCDF-1b expression vectors was prepared ( Qiagen miniprep kit as per instructions ) and co-digested with the relevant restriction enzymes . Ligations were performed at a 3∶1 molar excess of insert to vector using the Promega T4 DNA ligase rapid ligation system . Aliquots of the ligation reaction containing pET-28a ( + ) or pBAD30 vectors were electroporated into Epicentre Transformax EC100 E . coli and recovered on Luria–Broth ( LB ) agar containing 100 µg/ml ampicillin . Ligation containing pCDF-1b vector was transformed into chemically competent BL21 E . coli ( Invitrogen ) and recovered on LB agar containing 50 µg/ml Streptomycin . Correct constructs were selected by restriction digest of DNA prepared from candidate clones and verified by subsequent sequencing and stored at −80°C in 15% glycerol . Positive clones were subsequently induced for protein expression . Glycerol stocks were used to inoculate 5 ml of fresh LB media supplemented with 0 . 2% glucose ( w/v ) and the appropriate antibiotic for selection . Bacteria were grown overnight at 37°C with shaking , 1 ml of this culture was then harvested and resuspended in 100 ml of the same media and incubated at 37°C until an OD600 of 0 . 7–0 . 9 was achieved . Cells were then harvested at room temperature by centrifugation at 4000 rpm for 10 min . The pellet was re-suspended in 100 ml of fresh LB , supplemented with the appropriate antibiotic and 0 . 2% ( w/v ) of the pBAD30 promoter inducer L-arabinose or 1 mM IPTG for pCDF-1b construct . Cells or trichloroacetic acid ( TCA ) precipitated and concentrated supernatants from overnight induced cultures were collected for analysis using SDS-PAGE to confirm expression . The c1AH10 derived cosmids containing transposon insertions into the pdl1 gene ( pdl1 KO1-mutant ) or in the pWEB vector backbone ( CVI-wt ) were isolated from the E . coli Pl W14 library clones and each transformed into chemically competent E . coli BL21 cells . To complement the mutant pdl1 gene , E . coli BL21 cells were co-transformed with both the pdl1 KO1-mutant cosmid and the pCDF-1b:pdl1 construct . Transformant clones were recovered on LB agar containing the appropriate antibiotics . Positive clones containing both constructs were confirmed both by colony PCR and by DNA isolation and identification of cosmid and plasmid DNA using gel electrophoresis . Single colonies were subcultured in 5 ml of LB with the appropriate antibiotics and were grown overnight at 28°C with shaking . Larger subcultures were inoculated with these overnight pre-cultures ( 1/100 v/v dilution ) and grown at 28°C with shaking for 72 h to be used in bioassays . Whole cultures , washed cells and supernatants were used in the M . sexta neonate feeding bioassays . Supernatants or washed cells were diluted in 1× phosphate-buffered saline ( PBS ) and applied to 1 cm3 disks of artificial wheat germ diet as previously described [13] . Treated food blocks were allowed to dry for 30 min and two neonate M . sexta larvae were placed on each food block before incubation at 25°C for 7 days . Larvae were then scored for mortality and weighed . Larval growth differences are then expressed either as the direct mean values or as relative weight gain ( RWG ) means normalised to control groups [13] . The control groups are typically E . coli culture conditioned medium . RWG means are calculated using the following formula: RWG mean = 1+ ( ( sample mean−control mean ) /control mean ) . P . luminescens W14 ( Pl W14 ) was electroporated with various expression constructs using a standard E . coli optimised protocol . Transformants were confirmed by restriction digest of plasmid DNA prepared using a Qiagen kit as per manufacturer's protocol . Pl W14 carrying the pdl1 or orf53 expression constructs were induced with arabinose and cultured for different time points ( 3 h , 1 day and 2 days ) . Cells were washed in 1× PBS , normalized to an optical density of 10 . 0 , and lysed by sonication ( 10 s on and 10 s off 45% power regime for 3 min ) . Unbroken cells were removed by low speed centrifugation . One millilitre of each cell-free lysate was fractionated into the soluble ( cytoplasm and periplasm ) and insoluble ( inner and outer membrane ) fractions by ultracentrifugation ( 28 000 r . p . m . for 2 h in a Beckman SW40Ti rotor ) . The insoluble fraction was resuspended in 1 ml of PBS to restore its original concentration relative to the soluble fraction proteins in the lysate . Supernatant samples were prepared and concentrated using TCA precipitation . Protein sample amounts and quality were initially assessed by standard Coomassie blue staining of SDS-PAGE gels . For western blotting , protein fractions were separated by 1 dimensional SDS-PAGE and western blotted onto Nitrocellulose using a Biorad semi-dry blotter . We probed these membranes using a standard protocol with the following antibodies: TcdB1 ( C terminal ) anti-peptide antibody ( raised against aa856- YSSSEEKPFSPPNDC-aa869 ) , monoclonal anti-poly-histidine antibody ( SIGMA ) , anti-β-Lactamase antibody ( Millipore ) and anti-SecG ( a kind gift from Prof . Hajime Tokuda , University of Morioka ) . Immune-reactive bands were visualized using alkaline phosphatase-conjugated anti-rabbit or anti-mouse secondary antibody ( SIGMA ) at 1∶5000 and developed using NBT-BCIP agent . Anti-β-Lactamase western blots were performed to ensure no contamination of soluble material in the membrane fraction and anti-SecG western blots ensured no contamination of the membrane material in the soluble fraction . Total RNA was extracted from normalised cell number samples ( OD600 of 0 . 5 ) of E . coli or W14 cells at different time points using RNeasy kit ( QIAGEN ) according to the manufacturer's protocol . In all cases , the RNA was initially quality controlled by performing a standard Taq PCR reaction , using the intended primer pairs to ensure no contaminating DNA . RT-PCR was performed using One-step RT-PCR kit ( QIAGEN ) . Thermocycling was performed as follows: 50°C for 30 min; 95°C for 15 min; 94°C for 30 s; 50°C for 30 s and 72°C for 1 min , for 25 cycles and a final 72°C incubation for 10 min . Primers , which were designed to include unique sequence of different target genes , were described as follows ( 5′-3′ ) : tccC5-F: CAGGCGGAACAGGTGATTAT; tccC5-R GAGTTGGATCTGCGGTCAAT; tcdA1-F: TTGAGAGCGTCAATGTCCTG; tcdA1-R; TATCCGCGGCTCTGTCTAGT; tcdB1-F: TGGAAGCCTCGATATCATCC; tcdB1-R: ATAGGCCAGTTCCAGTGGTG; orf53-F: AAATTACGTCTGGATGTGAAG; orf53-R: CCAGTAGTCTATCGTTTGGCG; pdl-F: GGGAACAATAAGCAGGGTGA; pdl-R: GGTGACGGCGATAACAACTT; tccC2-F: ATCGGGGTGTTCTCAGTACG; tccC2-R: TTCTGTTTGGCTGTTTGCTG ( i ) Plate assay: E . coli carrying pBAD30-pdl1 were plated onto LB agar supplemented with ampicillin ( 100 µg/ml ) and sheep red blood cells ( SIGMA ) . For induction of pdl1 expression 0 . 2% ( w/v ) arabinose was also included . Plates were incubated at 37°C overnight before visualisation . ( ii ) Liquid assay: These were performed as described elsewhere [23] . Briefly , washed cells were isolated from induced and un-induced pBAD30-pdl1 cultures and sonicated . For induction , 0 . 2% ( w/v ) arabinose was added to cultures during exponential growth phase . Subsequently , 50 µl of red blood cells were suspended in phosphate buffered saline ( PBS ) , added to 150 µl of either control buffers , induced or un-induced bacterial lysate before incubation at 37°C . After 24 hours , the whole red blood cells were removed by centrifugation at 5 rpm for 5 min , and the extent of haemoglobin release within the reaction supernatant was measured at an optical density of 540 nm in a spectrophotometer . Percentage haemolysis was calculated as [ ( OD540 treatment sample/OD540 total lysis ) ×100] . NCBI accession numbers for the proteins/genes described in this study are as follows: Pdl1 , AAL18491; Orf53 , AAL18489; TcdA1 , AAL18486; TcdB1 , AAL18487; TccC5 , AAO17210; Orf54 , AAL18490; Orf47 , AAL18483 and Orf48 , AAL18484 .
We compared M . sexta oral toxicity data with a phylogenetic analysis of the P . luminescens species [24]–[25] . Members of the clade exemplified by strain Pl W14 all produce orally toxic supernatants and cells , while those in the clade which include strain Pl TT01 exhibit orally toxic cells but not supernatants ( Fig . 1C ) . It is important to note here that we are talking about toxicity and not infection as P . luminescens cannot infect the model insect M . sexta via the oral route . Despite this , limited microarray analysis confirmed that all P . luminescens strains encode homologues of the oral toxins tcdA1 and tcdB1 genes [26] , suggesting other lineage specific genetic factors might control the release of the orally toxic Tcd complex off the cell and into the surrounding medium . Previous work has identified the tca and tcd locus in Pl W14 as responsible for oral toxicity to M . sexta [4] . With the availability of the Pl TT01 genome sequence [27] and the sequences of several tc loci of Pl W14 [20] we have been able to directly compare the tca and tcd loci . The tca locus has undergone genetic degradation in Pl TT01 and so cannot contribute to oral toxicity . Conversely the Pl TT01 and Pl W14 tcd pathogenicity islands are well conserved , although the Pl W14 locus also contains a number of additional genes absent from Pl TT01 ( Fig . 2A ) . We speculated that these genes are responsible for the release of Tcd into the surrounding supernatant in Pl W14 , and without them , the majority of the Tcd complex remains associated with the cell surface in Pl TT01 . Previous immuno-gold localisation studies using a Tca polyclonal antibody to probe thin sections of Pl W14 confirmed that TC cross-reactive antigens were indeed localised on the surface of the bacterial cells [19] . To study Tcd export we developed a heterologous model for expression and secretion in E . coli . Previously we PCR screened a Pl W14 cosmid library for clones that encompassed the tcdA1B1 genes [18] and tested them for the production of orally toxic supernatants . We confirmed that cosmid c1AH10 produced both supernatant and cell associated oral toxicity , consistent with the Pl W14 phenotype . This cosmid encodes intact copies of all three necessary subunit genes , A , B and C , ( tcdA1 tcdB1 and tccC5 ) , required for full toxicity ( Fig . 2B ) . Nevertheless the roles of other genes on this cosmid were not known . Two small Pl W14 specific islands are present on this cosmid . The first small island encodes a gene named pdl1 , the predicted protein product of which contains putative lipase and protease domains , in addition to two small self-similar tightly linked ORFs ( orf53 and orf54 ) each with type II secretion signal peptides but no other recognisable domains . The second island contains the genes orf47 , which also has a predicted esterase/lipase domain but is unrelated to pdl , and orf48 which again has a putative signal peptide leader sequence but no other recognizable domains . To validate this cosmid as a suitable model for studying Tcd expression and secretion we compared the transcription and translation of several genes on c1AH10 in E . coli with those in the original Pl W14 strain across the growth curve using RT-PCR ( Fig . S1 ) and western blot analysis ( Fig . S2B ) . RT-PCR analysis revealed that the A , B and C-subunit gene transcripts are present at all growth phases , but decline by 3 days . A similar trend in transcriptional levels is seen for both the cosmid and the W14 strains supporting its relevance as a model system . We note a peak in pdl1 mRNA at the late phase of growth ( OD = 2 , around 8 h for W14 ) especially in the cosmid clone which is consistent with the onset of toxin secretion . We designed an anti-peptide antibody against the C-terminus of the B-subunit ( TcdB1 ) and used this to track the translation and location of the Tcd complex . We do not see B-subunit translation in either strain Pl W14 or the cosmid clone until 24 hours ( Fig . S2A ) , which correlates with the appearance of the oral toxicity phenotype . The failure of the antibody to cross react with supernatant proteins from a W14 tcdAB KO strain [4] confirms the specificity of this antibody for the TcdB1 subunit . When Tcd is produced by Pl W14 and the cosmid clone , it is present in the membrane and supernatant fractions ( Fig . S2B ) . This correlates with whole-cell and supernatant associated oral toxicity of both the cosmid clone and Pl W14 . Nevertheless , as previously seen , the overall levels of Tcd production are low [28] . An in vitro transposon mutagenesis kit was used to create a library of transposon mutants of cosmid c1AH10 . Insertion knock-out mutants in all genes were identified by sequencing out from the transposon into the flanking cosmid sequence . The oral toxicity of cells and supernatants of this panel of cosmid mutants ( in E . coli ) was then examined by M . sexta oral bioassay . As expected this confirmed that all the three subunit genes were required for toxicity in E . coli ( at least under these “native” expression levels ) . In addition however we can see that insertion mutagenesis of pdl1 significantly lowered toxicity of the supernatant and to some extent the cells ( Fig . 3 ) . A panel of transposon mutants was transformed into Pl TT01 which is normally unable to release its native Tcd into the surrounding supernatant . Oral bioassay revealed that a cosmid in which the transposon was inserted into the cosmid vector backbone did indeed generate orally toxic supernatants ( Fig . S3 ) as it does in E . coli . Importantly the pdl1 cosmid mutant in the Pl TT01 background was again unable to produce orally toxic supernatants . Furthermore insertion into either of the Pl W14 C- or B-subunit genes also failed to produce toxic supernatants , suggesting they were essential to the Pdl1 mediated Tcd export enhancement . Interestingly , cosmid clones in which the A-subunit gene ( tcdA1 ) was interrupted were still able to produce orally toxic supernatants . This indicated that a Pl TT01 A-subunit homologue was able to trans-complement for the cosmid Pl W14 equivalent . Conversely , the inability of the Pl TT01 B- and C-subunit homologues to trans-complement the equivalent cosmid mutants revealed that the Pl W14 Pdl1 does show some substrate specificity . As further confirmation of this we cloned pdl1 alone into the arabinose inducible expression plasmid pBAD30 for over expression in Pl TT01 . Consistent with our hypothesis , this construct did not significantly increase oral toxicity of the supernatants ( data not shown ) . In order to understand the effect of Pdl1 on Tcd export we examined cell soluble and membrane fractions and culture supernatants for Tcd by western blotting using an anti B-subunit antibody . We compared these fractions from E . coli harbouring the wild-type ( Fig . 4 - CVI ) and pdl1 mutant ( Fig . 4 – pdl1 KO ) cosmids ( Fig . 4 ) . The wild-type cosmid strain secretes Tcd into the supernatant with very little remaining associated with the soluble fraction , which consists of cytoplasmic and periplasmic fractions . Conversely , the pdl1 mutant cosmid showed a reduction in the overall amount of Tcd secreted into the supernatant but an increase in amounts in the soluble and membrane fractions . This is despite a reduction in cell associated toxicity of the mutant ( Fig . 3 ) , suggesting that Pdl1 increases the efficiency of the normal secretion process . The soluble fraction shows the Tcd is accumulating in either in the cytoplasm and/or periplasm while the membrane fraction shows at least some is accumulating in either the inner or outer membranes . The role of Pdl1 in increasing overall export and release of Tcd was supported with a complementation assay in E . coli . We co-transformed E . coli BL21 with both the pdl1 knock out mutant cosmid and a pdl1 expression construct , pCDF-1b:pdl1 . We confirmed Pdl1 expression in these strains using SDS-PAGE and confirmed that trans-complementation restored supernatant toxicity close to levels seen in the “wild-type” parental cosmid strain encoding an intact copy of the pdl1 gene ( Fig . S4 ) . Expression of Pdl1 from pCDF-1b:pdl in the absence of the tcd cosmid showed no toxicity as expected . These experiments confirmed that the transposon insertion in the pdl1 KO1-mutant cosmid was affecting the pdl1 gene only . Interestingly we also saw a slight increase in cell associated toxicity in this trans-complemented strain , consistent with the cosmid mutation studies . This indicates that Pdl1 not only increases Tcd release into the surroundings but also enhances overall expression levels . C–terminal His tagged fusions of pdl1 and the adjacent orf53 were cloned into the arabinose inducible pBAD30 expression plasmid ( orf54 was not used as it appears to be translationally coupled to pdl1 ) . These constructs were transformed into Pl W14 , induced for 3 , 24 and 48 hours and supernatant , membrane and soluble cell fractions prepared . The location of the His-tagged Pdl1 and Orf53 proteins were examined using western blotting with an anti-his tag antibody . Pdl1 was found in the soluble and membrane fractions for up to 2 days , but was always absent from the supernatants ( Fig . S5A ) . Interestingly , the Orf53 protein was seen in the soluble and the membrane fractions at 3 h . The full length , unprocessed protein is seen in the soluble fraction while the majority of protein in the membrane fraction is a slightly shorter processed form , presumably with the type II signal leader removed ( Fig . S5B ) . By 24 h the protein levels have fallen and are absent by 48 h . The oral toxicity of cells and supernatants from these same samples was tested ( Fig . 5 ) . At 24 h , over-expression of pdl1 in Pl W14 increases the level of oral toxicity of both cells and supernatants relative the vector control . Conversely over-expression of orf53 lowers toxicity of the supernatant . Interestingly by 48 h the orf53 over-expression strain returns to the normal wild-type level of oral toxicity . This correlates to the disappearance of the Orf53 protein on the western blots by two days . Pdl1 over-expression continues to increase oral toxicity even at 48 h . This supports a model whereby Pdl1 increases Tcd export and Orf53 acts antagonistically to inhibit this effect . The demonstration that Pdl1 is not found in the supernatant fraction of Pl W14 cultures confirms that the increase in Tcd dependant supernatant toxicity is not the result of a continued physical association of Pdl1 with the secreted Tcd complex ( Fig . S5 ) . Protein alignments indicate that Pdl1 and Pdl2 contain putative protease and lipase like catalytic domains ( Fig . S6 ) . We therefore tested the possibility that Pdl1 might directly modify the Tcd complex in order to increase its specific activity . We mixed Tcd containing samples with lysate from the induced E . coli pBAD30-pdl1 expression strain and performed oral toxicity bioassays ( Fig . S7 ) . These experiments confirmed that Pdl1 had no effect on the activity of Tcd isolated from any of the cell compartments , suggesting it is not likely to proteolytically “activate” the complex . In order to test if Pdl1 has any measurable lipase activity we tested its ability to lyse sheep red blood cells in vitro . We did this using both standard blood-agar haemolysis plates and a liquid assay based on measuring haemoglobin release . Pdl1 was heterologously expressed using the E . coli pBAD30-pdl1 expression construct . To ensure any observed lysis was not due to the E . coli sheA dependant cryptic haemolytic activity [29] , the pBAD30-pdl1 construct was first electroporated into the E . coli sheA mutant CFP201 . Fig . 6 confirms that after overnight incubation , Pdl1 expressing cells show limited lysis of sheep erythrocytes consistent with lipase activity . Finally we investigated the effect of pdl1 over-expression and the co-expression of the tightly linked orf54 and orf53 upon E . coli . Pdl1 expression induced the release of specific protein species into the culture supernatant when compared to a vector only E . coli control ( Fig . S8 ) . Interestingly the most abundant of polypeptide appears to be a fragment derived from a RhsA-like protein of E . coli as well as other predicted membrane bound proteins such as the transporter MchF and the outer membrane lipoprotein PgaB . Interestingly when orf54 was co-expressed with pdl1 , the abundance of the various released polypeptides decreased , and was in most cases abolished when both orf54 and orf53 were both co-expressed . This further supports our findings with Tcd secretion that Pdl1 is able to influence protein export and that the tightly linked orf54 and orf53 gene products serve to repress the function of Pdl1 in a gene dose dependant manner .
Transcript and Western blot analysis revealed that the TC subunits are under post-transcriptional regulation , and that their translation is timed with secretion and release from the cell . The TC toxin subunits encode no recognised export signal sequences and show no similarity to either two partner secretion proteins , or known auto-transported proteins . Nevertheless these large proteins are secreted from the cell and assembled into a large complex without the need for cell lysis as confirmed by the viability of the E . coli cosmid clones . This suggests that either the TC subunits are auto-transported or that they may be secreted by a chromosomally encoded system conserved in both Photorhabdus and the E . coli laboratory strains used . The mechanism of TC subunit secretion is currently under investigation in our laboratory however it is clear that the exported TC becomes localised to the membrane fraction , and is accessible to the outside of the cell , potentially located on the outer membrane . The B and C subunit proteins can both be seen to contain tyrosine-aspartic acid ( YD ) repeat motifs which in other proteins are involved in binding carbohydrates [30] . It is possible that the YD repeats found in the B- and C-subunits are responsible for the wall association of the TC with membrane or extracellular polysaccharide components . Another possibility is that one or more of the subunits is a lipoprotein and is attached by a lipid anchor , although they do not exhibit the usual lipidation export signal peptide . Bacterial precursor lipoproteins can be translocated across the cytoplasmic membrane by either the Sec ( general secretory ) or Tat ( twin arginine transport ) pathways [31] and then lipid modified by a lipoprotein diacylglycerol transferase on a conserved cysteine . This cysteine is located in what is known as a lipobox motif at the end of the signal peptide [32] . The TC subunits contain no such leader peptides . Our findings show that the Pdl1 protein can enhance Tcd secretion and that products of the small tightly linked genes are able to repress this effect . Our experiments confirmed that both Pdl1 and Orf53 are located in the membrane fraction although the exact location is not clear . Unfortunately we were unable to prepare isolated inner and outer membrane fractions from Photorhabdus in order to precisely localise the tagged Pdl1 and Orf53 proteins . It should be noted that similar problems were encountered in experiments with the closely related bacteria Xenorhabdus nematophila [33] . The presence of a type II Sec-dependant secretion leader on Orf53 does suggest that it can at least reach the periplasmic space , and indeed our experiments show both intact and processed forms consistent with cleavage of the signal peptide . PSORT analysis of the protein sequence gives a strong prediction for its localisation in the periplasmic space ( certainty value = 0 . 930 ) . Results presented here suggest that the Pdl1 protein does not associate with or modify the Toxin Complex directly in order to increase its specific activity . Rather our data are consistent with a role in the enhancement of secretion . The over-expression of Pdl1 in Pl W14 increases the overall levels of both supernatant and cell contact dependant toxicity . One possible model infers that translation is coupled to export , with Pdl1 increasing the efficiency of secretion , leading to an overall increase in TC production . In this model , as cell surface binding sites become limited , then excess toxin is shed into the surrounding medium . Pdl1 may interact with the Tcd secretion mechanism on the cytoplasmic membrane during export . In this model , the Orf53 molecules would antagonise the activity of Pdl1 by interactions on the periplasmic face of the inner membrane . While we confirm lipase activity of Pdl1 , the predicted presence of a serine protease catalytic motif does suggest that while proteolytic activity may not be involved in “activation” of the TC , it could still play a role in enhancing secretion . Experiments in which cosmid mutants were transformed into strain TT01 suggest that the Pdl1 mediated release is dependent upon the BC sub-complex and is independent of the A-subunit . It should be noted that C-subunits belong to a much larger family that includes the Recombination Hot Spot genes ( Rhs ) of E . coli . These genes have an unusual structure with conserved N-termini , a highly conserved central core region and variable C-terminal “tails” [30] , [34] . In the case of the C-subunit proteins we know the variable C-terminal domains encode toxic effectors [9] . The N-terminal domain of RhsA has been shown to be involved in the biosynthesis and export of capsular polysaccharide in E . coli although the exact mechanism remains cryptic [35] . Interestingly , Rhs-core like genes are often tightly linked to vgrG genes which in some cases have been shown to be contact dependant cytotoxins exported via type VI secretion machinery , temping the speculation of an equivalent toxin secretion and anchoring role [36] . An examination of the full genome sequence of P . luminescens TT01 shows only a single pdl homologue unlinked to any tc genes . Interestingly this is however closely linked to a putative type III effector toxin gene , plu3163 ( HopT1-2 homologue ) . The emerging human pathogen P . asymbiotica can however be seen to contain several genomic islands ( GIs ) containing repeats of pdl-orf54 like genes . When cloned on cosmids , three of these pdl-GIs have been shown to be toxic to insects and nematodes [37] . Insertional mutagenesis of a cosmid containing the pdl-GI_2 region ( Fig . S9A ) confirmed that knock-out of either the putative toxin gene ( a vgrG homologue ) or two of the pdl homologues significantly reduced toxicity of the bacterial cultures [37] . When compared to genomic sequences in the public databases pdl-like genes may be seen in other pathogenic bacterial genera including Vibrio and Pseudomonas [38]–[39] . In many genomes we can see tight linkage between type VI secretion operons and pdl-orf54 homologue gene pairs , suggesting they also play a role in these secretion systems ( Fig . S9B ) . Taken together , these findings suggest that Pdl homologues may represent a novel general mechanism for enhancing secretion of toxins from Gram-negative bacteria .
|
Bacterial pathogens of insects deploy a range of toxins to combat the innate immune system and kill the host . There is significant interest in developing these toxins as candidates for crop protection strategies . To date , transgenic crops expressing Bacillus thuringiensis Cry toxins have been used to resist predation by pests . In order to minimize the risk of insect resistance development , current research in crop biotechnology comprises the design of new transgenic plants expressing toxins with different modes of action . The Toxin Complex ( TC ) gene family first identified in the insect pathogen Photorhabdus has received interest as an alternative . It remains obscure how Photorhabdus regulates , assembles , and secretes such a large toxin complex . We have identified a small lipase protein , Pdl1 , which enhances secretion and leads to the release of the Toxin complex off the bacterial surface . This is of wider significance because TC toxin homologues are also found in a range of human pathogens , such as Yersinia in which they have been implicated in human virulence . Furthermore homologues of pdl are also seen tightly linked to other virulence loci such as the type VI systems of Vibrio . We speculate that this Pdl mediated secretion enhancement system is a widespread and important mechanism used by Gram negative bacterial pathogens .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"gram",
"negative",
"microbial",
"pathogens",
"biology",
"microbiology",
"host-pathogen",
"interaction",
"bacterial",
"pathogens",
"pathogenesis"
] |
2012
|
Pdl1 Is a Putative Lipase that Enhances Photorhabdus Toxin Complex Secretion
|
Phylogenetically distinct Mycobacterium tuberculosis lineages differ in their phenotypes and pathogenicity . Consequently , understanding mycobacterial population structures phylogeographically is essential for design , interpretation and generalizability of clinical trials . Comprehensive efforts are lacking to date to establish the West African mycobacterial population structure on a sub-continental scale , which has diagnostic implications and can inform the design of clinical TB trials . We collated novel and published genotyping ( spoligotyping ) data and classified spoligotypes into mycobacterial lineages/families using TBLineage and Spotclust , followed by phylogeographic analyses using statistics ( logistic regression ) and lineage axis plot analysis in GenGIS , in which a phylogenetic tree constructed in MIRU-VNTRplus was analysed . Combining spoligotyping data from 16 previously published studies with novel data from The Gambia , we obtained a total of 3580 isolates from 12 countries and identified 6 lineages comprising 32 families . By using stringent analytical tools we demonstrate for the first time a significant phylogeographic separation between western and eastern West Africa not only of the two M . africanum ( West Africa 1 and 2 ) but also of several major M . tuberculosis sensu stricto families , such as LAM10 and Haarlem 3 . Moreover , in a longitudinal logistic regression analysis for grouped data we showed that M . africanum West Africa 2 remains a persistent health concern . Because of the geographical divide of the mycobacterial populations in West Africa , individual research findings from one country cannot be generalized across the whole region . The unequal geographical family distribution should be considered in placement and design of future clinical trials in West Africa .
West Africa consists of 15 countries with 245 million inhabitants ( S1A Fig ) , 13 of which belong to the world’s 42 countries with the lowest human development index [1] . Consequently , it faces great challenges in controlling infectious diseases , such as tuberculosis ( TB ) . Clinical trials investigating the local health needs are much needed to understand and tackle the TB epidemic in West Africa . The composition of the endemic mycobacterial population infecting human study subjects can have a major impact on TB clinical trial outcomes and should ideally be accounted for in the planning phase of any project [2] . Considering bacterial variation between study sites is also essential to estimate to what extent country-specific results can be generalised to the whole of West Africa . The MTBc can be divided into six major lineages , comprised of the Indo-Oceanic ( L1 ) , East-Asian ( L2 ) , Central Asian ( L3 ) , Euro-American lineages ( L4 ) and the two endemic African lineages M . africanum West Africa 1 ( MAF1 , L5 ) and M . africanum West Africa 2 ( MAF2 , L6 ) [3] . Although MAF1 seems to be disappearing in some countries , the longitudinal development of MAF2 is not known . Each of these phylogenetically distinct lineages can be further differentiated into mycobacterial families , such as , amongst others , the Latin-American-Mediterranean ( LAM ) or Haarlem families within the Euro-American lineage [3] . Interestingly and for reasons not understood , West Africa is the only region in the world in which all of the six major human lineages are present . This exceptional diversity necessitates future West African trials to be adjusted for this unique bacterial variability—even more than trials in other parts in the world . Therefore the scope of the present publication was to describe the geographical distribution and spatial variations of mycobacterial families across the region .
We searched Pubmed using terms “spoligotype” , “spoligotyping” with respective country names . Studies on pulmonary TB up to December 2014 were included , in which spoligotypes on all isolates were available . Individual spoligotypes designated as mixed infections were excluded . In case several publications analysed the same dataset , the most comprehensive collection was selected . M . bovis studies , conducted in high risk populations ( abattoir staff ) were excluded . To assign mycobacterial families to isolates , and to ensure comparability between different datasets , we re-analysed extracted spoligotype information using a standardized approach . Isolates were classified into families using the online platform “Spotclust” at the default settings . For M . africanum isolates , Spotclust identifies , but does not distinguish between MAF1 and 2 . Therefore “TBLineage” was further applied to M . africanum isolates previously identified by Spotclust [4] . Both Spotclust and TB Lineage are mathematical algorithms that were shown to reliably identify mycobacterial lineages and families based on respective signature spoligotype patterns . A detailed description of the algorithms and their performance is described elsewhere [4 , 5] . The lineage/family distribution per country/study site was plotted as chloropleth maps generated using QGIS 2 . 0 . 1 ( http://qgis . osgeo . org ) . To investigate geographical differences in mycobacterial families across West Africa we split West Africa into a Western and an Eastern region . Western countries include Gambia , Guinea-Bissau , Guinea , Sierra Leone , Ivory Coast , Mali , Senegal , while Eastern countries include Benin , Burkina Faso , Ghana , Niger and Nigeria ( S1 Fig ) . With region as response variable , the proportion of each family was tested univariately using logistic regression , with country fitted as a cluster to account for multiple studies per site . Families found in one region and not in the other cannot be modelled mathematically because the maximum likelihood for these families does not exist . We defined families with complete separation between regions as ‘perfect predictors’ . A two-sided p-value <0 . 05 was considered statistically significant and a two-sided p-value ≥0 . 05 & <0 . 10 was considered of borderline significance . No adjustment was made for multiple testing . All analyses were performed using Stata v12 . 1 ( StataCorp . 2011 . Stata Statistical Software: Release 12 . College Station , TX: StataCorp LP . ) . Phylogeographic analysis using linear axis analysis in GenGISvs2 . 2 . 2 was conducted [6] . The default GenGIS Africa map was used . A UPGMA phylogenetic tree was constructed from spoligotyping data ( S2 Fig ) using the publicly available MIRU-VNTRplus software [7] and uploaded into GenGIS allowing for the re-ordering of leaf nodes . A Linear axis plot ( 10 . 000 permutations ) was run at significance level p = 0 . 001 . Gambian isolates , collected within a TB Case Contact cohort , in which all cases of the Greater Banjul area are recruited [8] were spoligotyped . Genotyping was approved by the Gambian Government/MRC joint ethics committee . Longitudinal lineage data was modelled using logistic regression for grouped data . The outcome was the number of a particular lineage out of the total number of samples taken in each year . Both lineage and year were fitted as explanatory variables and interactions between the two explored . The multicollinearity between the lineages was avoided by excluding one lineage and fitting the model on the remaining .
Of 20 original research articles , four were excluded ( based on above criteria ) , with the remaining 16 covering 12 of 15 West African countries . In total we collected , extracted and ( re ) analysed spoligotype information of 3580 isolates , belonging to six major human lineages , of which the Euro-American lineage ( L4 ) , together with M . africanum lineages ( L5 and 6 ) were the main causes of pulmonary TB ( Table 1 ) . Thirty-two different mycobacterial families were identified , but 84% of all patients are infected by only eight major families ( Fig 1 ) . Common to most of the countries is the “ill-defined” T1 family . We also confirmed the previously described geographical distribution of two M . africanum lineages [24] . While MAF1 ( L5 ) has the highest presence in Nigeria/Benin , MAF2 ( L6 ) is mainly found in Gambia/Guinea-Bissau . Besides MAF2 as a major cause for TB , a variety of Euro-American families ( Haarlem 1 and 3 , LAM9 , amongst others ) are prevalent in western West Africa . This is in sharp contrast to eastern West Africa where , besides MAF1 , the great majority of TB infections is attributable to only one other dominant family LAM10 . A recently introduced family into West Africa is the Beijing family which lead to an outbreak in Cotonou , Benin [25] . The only other place with comparably high numbers of Beijing isolates is Dakar in Senegal . Both cities , Dakar and Cotonou have major international ports . To evaluate whether identified families are geographically equal , we divided West Africa into a Western and Eastern region ( S1B Fig ) . Univariate logistic regression analysis showed that the proportion of mycobacterial families can serve as predictors for the two regions . 13 out of 32 families were associated with one of the two regions ( see S1 Table ) . Amongst these were four of the eight major families: LAM10 ( perfect predictor at proportion ≥0 . 12 ) and MAF1 ( p = 0 . 08 ) as predictors for the East and Haarlem 3 ( p = 0 . 07 ) and MAF2 ( p = 0 . 09 ) for the West . To verify the geographic separation of these four major families , which cause 51% of all TB , we carried out an independent phylogeographic analysis using GenGIS software ( Fig 2 ) . We constructed an UPGMA tree based on 279 unique Haarlem 3 , MAF1/2 and LAM10 spoligotypes , which was superimposed onto geographic locations and mycobacterial family distributions of the study sites ( Fig 2A ) . In case of geographical separation , one expects significantly less crossings between the phylogenetic tree and the spoligotype distribution in the study sites than by mere chance . A linear axis analysis ( p<0 . 001 , 10 . 000 permutations ) identified several orientations of the tree’s geographical axis that resulted in less than the 9759 . 5 crossings expected by chance . Fig 2B demonstrates that geographical separation occurs at various geographical axis angles , with the least crossings ( 9144 ) at 228 . 1° ( Fig 2A ) . Although spoligotyping might have led to minor misclassifications of MAF1/MAF2 isolates in our phylogenetic analyses ( S2 Fig ) , we expect such misclassification to result in an unbiased underestimation of the observed geographical separation . 1164 consecutive TB patients were recruited between 2002–2010 . Logistic regression modelling revealed both a non-significant lineage by time interaction ( p = 0 . 38 ) and a non-significant time main effect ( p = 0 . 80 ) . Our analysis therefore indicated that the proportions of lineages are stable over time . The overall lineage percentages , in order of magnitude , are Euro-American 57 . 2% ( 95%CI 54 . 4%-60 . 0% ) , MAF2 35 . 4% ( 95%CI 32 . 7%-38 . 2% ) , Indo-Oceanic 4 . 3% ( 95%CI 3 . 3%-5 . 6% ) , East Asian ( Beijing ) 2 . 5% ( 95%CI 1 . 7%-3 . 6% ) , MAF1 1 . 0% ( 95%CI 0 . 4%-2 . 4% ) and East African Indian 0 . 8% ( 95%CI 0 . 2%-3 . 2% ) ( Fig 3 ) .
We confirmed that modern Euro-American strains are the predominant lineage followed by the two M . africanum lineages . Although the polyphyletic T1 family [26] is rather equally distributed across the whole region , we find geographical variations of other families . While western West Africa shows a high genetic diversity from a multitude of mycobacterial families , the MTBc of Eastern West Africa is mainly composed of two dominant families ( LAM10 and MAF1 ) . Although other West and Central African countries observed a replacement of MAF1 and MAF2 with modern strains [10 , 11 , 14 , 27] , our longitudinal analysis from The Gambia did not confirm these findings and MAF2 remains an important cause of TB in the country . The exact mechanism of how MAF2 can maintain a stable prevalence of 35% over the last decade within The Gambia ( despite a slower progression to disease when compared to M . tuberculosis [28] ) is not fully understood . Besides the known geographical divide of the two M . africanum lineages , we find for the first time geographical separation of major Euro-American families in West Africa . Due to this spatial variation previous research findings observed in one West African country/region are hardly generalizable to the sub-region . In addition , the unequal distribution has important implications for design of future trials . For instance , western West African countries with their high genetic diversity are appropriate settings for research that aims to test whether novel diagnostics or vaccine candidates work equally well against different MTBc families . In contrast , research on host genetics , benefitting from low diversity , would yield more robust results when conducted in eastern West Africa with predominant LAM10 and MAF1 families . To investigate the spreading of novel TB families , one can follow up on the geographical expansion of LAM10 or on recently introduced Beijing strains into Benin or Senegal . As first studies confirmed that the “ill-defined”T1 is not a monophyletic clade [26] , further research using more robust phylogenetic markers could focus on understanding the endemic MTBc composition T1-endemic countries . The presented phylogeography also has limitations: first , we combined genotypic information , independent from respective collection strategies ranging from convenience to systematic sampling . Therefore data presented are a cross-sectional compilation of genotyping information between 1986–2012 . Also , individual patient’s treatment history , whether they presented as new or retreatment cases , was not systematically collected and has not been accounted for . In order to avoid over-interpretation of results , we agree that comparing differing sampling strategies is challenging , and we therefore limited our discussion to proportions of families with larger isolate numbers . Lastly , the families themselves consist of a multitude of strains characterised by specific spoligotypes ( shared international types , SITs ) and we did not study whether the local expansion of a family was driven by one or several individual proliferating SIT within the family . Spoligotyping can be successfully used to assign the majority of mycobacterial isolates to one of the major mycobacterial lineages and their families [4] . We appreciate that classification of mycobacteria in West Africa would ideally be based on whole genome sequencing ( WGS ) data , however , limited bioinformatics capacity combined with financial and infrastructural constraints did not allow high-throughput sequencing in most resource-limited West African countries to date . By summarizing available and novel data , we showed significant geographical variation of the MTBc , which will impact on the overall outcome of clinical trials in any specific region . With the generated data researchers can consider the demonstrated spatial variation in the planning stage of respective future clinical TB trials .
|
Tuberculosis is caused by bacteria belonging to the Mycobacterium tuberculosis complex ( MTBc ) , which consists of seven major , phylogenetically distinct lineages and their families . West Africa is the only region in the world where , besides the common M . tuberculosis lineages , the two M . africanum lineages are endemic . We demonstrate that the composition of the mycobacterial population in the western part of West Africa significantly differs from the one in the eastern part . This documented variation will impact on generalizability and interpretation of clinical trials outcomes . Therefore future trial designs need to consider the geographical diversity of underlying mycobacterial populations .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"biogeography",
"taxonomy",
"ecology",
"and",
"environmental",
"sciences",
"medicine",
"and",
"health",
"sciences",
"population",
"dynamics",
"population",
"genetics",
"geographical",
"locations",
"tropical",
"diseases",
"bacterial",
"diseases",
"phylogenetics",
"data",
"management",
"phylogenetic",
"analysis",
"paleontology",
"molecular",
"biology",
"techniques",
"population",
"biology",
"bacteria",
"africa",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"geography",
"computer",
"and",
"information",
"sciences",
"tuberculosis",
"paleogeography",
"actinobacteria",
"phylogeography",
"evolutionary",
"systematics",
"molecular",
"biology",
"molecular",
"biology",
"assays",
"and",
"analysis",
"techniques",
"people",
"and",
"places",
"mycobacterium",
"tuberculosis",
"earth",
"sciences",
"genetics",
"biology",
"and",
"life",
"sciences",
"evolutionary",
"biology",
"organisms",
"geographic",
"distribution"
] |
2016
|
A Mycobacterial Perspective on Tuberculosis in West Africa: Significant Geographical Variation of M. africanum and Other M. tuberculosis Complex Lineages
|
Evolution is shaping the world around us . At the core of every evolutionary process is a population of reproducing individuals . The outcome of an evolutionary process depends on population structure . Here we provide a general formula for calculating evolutionary dynamics in a wide class of structured populations . This class includes the recently introduced “games in phenotype space” and “evolutionary set theory . ” There can be local interactions for determining the relative fitness of individuals , but we require global updating , which means all individuals compete uniformly for reproduction . We study the competition of two strategies in the context of an evolutionary game and determine which strategy is favored in the limit of weak selection . We derive an intuitive formula for the structure coefficient , σ , and provide a method for efficient numerical calculation .
Constant selection implies that the fitness of individuals does not depend on the composition of the population . In general , however , the success of individuals is affected by what others are doing . Then we are in the realm of game theory [1]–[3] or evolutionary game theory [4]–[8] . The latter is the study of frequency dependent selection; the fitness of individuals is typically assumed to be a linear function of the frequencies of strategies ( or phenotypes ) in the population . The population is trying to adapt on a dynamic fitness landscape; the changes in the fitness landscape are caused by the population that moves over it [9] . There is also a close relationship between evolutionary game theory and ecology [10]: the success of a species in an ecosystem depends on its own abundance and the abundance of other species . The classical approach to evolutionary game dynamics is based on deterministic differential equations describing infinitely large , well-mixed populations [6] , [11] . In a well-mixed population any two individuals interact equally likely . Some recent approaches consider stochastic evolutionary dynamics in populations of finite size [12] , [13] . Evolutionary game dynamics are also affected by population structure [14]–[22] . For example , a well-mixed population typically opposes evolution of cooperation , while a structured population can promote it . There is also a long standing tradition of studying spatial models in ecology [23]–[25] , population genetics [26] , [27] and inclusive fitness theory [28]–[30] . Evolutionary graph theory is an extension of spatial games , which are normally studied on regular lattices , to general graphs [31]–[34] . The graph determines who meets whom and reflects physical structure or social networks . The payoff of individuals is derived from local interactions with their neighbors on the graph . Moreover , individuals compete locally with their neighbors for reproduction . These two processes can also be described by separate graphs [35] . ‘Games in phenotype space’ [36] represent another type of spatial model for evolutionary dynamics , which is motivated by the idea of tag based cooperation [37]–[39] . In addition to behavioral strategies , individuals express other phenotypic features which serve as markers of identification . In one version of the model , individuals interact only with those who carry the same phenotypic marker . This approach can lead to a clustering in phenotype space , which can promote evolution of cooperation [36] . ‘Evolutionary set theory’ represents another type of spatial model [40] . Each individual can belong to several sets . At a particular time , some sets have many members , while others are empty . Individuals interact with others in the same set and thereby derive a payoff . Individuals update their set memberships and strategies by global comparison with others . Successful strategies spawn imitators , and successful sets attract more members . Therefore , the population structure is described by an ever changing , dynamical graph . Evolutionary dynamics in set structured populations can favor cooperators over defectors . In all three frameworks – evolutionary graph theory , games in phenotype space and evolutionary set theory – the fitness of individuals is a consequence of local interactions . In evolutionary graph theory there is also a local update rule: individuals learn from their neighbors on the graph or compete with nearby individuals for placing offspring . For evolutionary set theory , however , [40] assumes global updating: individuals can learn from all others in the population and adopt their strategies and set memberships . Global updating is also a feature of the model for games in phenotype space [36] . The approach that is presented in this paper requires global updating . Therefore , our result holds for evolutionary set theory and for games in phenotype space , but does not apply to evolutionary graph theory .
As a particular game we can study the evolution of cooperation . Consider the simplified Prisoner's Dilemma payoff matrix: ( 6 ) This means cooperators , , pay a cost for others to receive a benefit , . Defectors , , pay no cost and distribute no benefits . The game is a Prisoner's Dilemma if . As shown in [41] , if we use equation ( 2 ) we can always write the critical benefit-to-cost ratio as ( 7 ) provided . If the benefit-to-cost ratio exceeds this critical value , then cooperators are more abundant than defectors in the mutation-selection equilibrium of the stochastic process for weak selection . A higher corresponds to a lower benefit-to-cost ratio and is thus better for the evolution of cooperation . From eqs ( 3 ) and ( 7 ) we can write ( 8 ) This formula is very useful for finding the critical benefit-to-cost ratio numerically . Moreover , we can rewrite the critical benefit-to-cost ratio in terms of average number of interactions rather than total number of interactions as ( 9 ) These equations provide intuitive formulations of the critical benefit-to-cost ratio for processes with global updating . Our new formula for ( eq . 3 ) gives a simple numerical algorithm for calculating this quantity in any spatial process with global updating and constant birth or death rate . We simulate this process under neutral drift for many generations . For each state we evaluate , , and . We add up all products to get the numerator in eq ( 3 ) , and then we add up all products to get the denominator . The resulting can be used for any game given by the payoff matrix ( 1 ) to determine if strategy is more frequent than strategy in the limit of weak selection . In this section we use the simple numerical algorithm suggested by our formula ( 3 ) to find for evolutionary dynamics on sets [40] . In that paper , the authors compute an exact analytic formula for that depends on the parameters of their model . We compare our simulated estimates for with their theoretical values and find perfect agreement ( Figure 2 ) . Furthermore , we use our computational method to calculate in an extension of the original model . An analytic solution for this extended model has not yet been found . Thus our simulated estimates constitute the first “solution” of this extended model ( Figure 3 ) . The original set-structured model describes a population of individuals distributed over sets . Individuals interact with others who belong to the same set . Two individuals interact as many times as they have sets in common , and these interactions lead to payoffs from a game as described in general in Section 2 . Reproductive updating follows a Wright-Fisher process , where individuals are selected with replacement to seed the next generation . The more fit an individual , the more likely it is to be chosen as a parent . An offspring adopts the parent's strategy with probability , as described in Section 2 . The offspring adopts the parent's set memberships , but this inheritance is also subject to mutation; with probability , an offspring adopts a random list of set memberships . This updating process can be thought of as imitation-based dynamics where both strategies and set memberships are subject to selection [40] . To obtain exact analytical calculations , it is assumed that each individual belongs to exactly sets . In Figure 2 , we pick values for , and and plot as a function of the set mutation rate , . The continuous curves are based on the analytic formula for derived in [40] . The new numerical algorithm generates the data points . There is perfect agreement between these two methods . In Figure 3 , we consider a variant of this model . Instead of belonging to exactly sets , individuals now belong to at most sets . With probability , an offspring adopts a random list of at most memberships , the length of which is uniformly random . So far there exists no analytical solution for this model but we can use eq . ( 3 ) to compute numerically . We interpolate the numerical results with smooth curves . We observe that for low mutation , Fig . 3 ( A ) , the case gives a which is smaller than the case . Hence , for low mutation , allowing people to be in at most sets turns out to be worse for cooperation than restricting them to be in exactly sets . However , for high strategy mutation , Fig . 3 ( B ) , the for is greater than the one for . Hence , for high strategy mutation , allowing individuals to be in at most sets seems to be better for cooperation than restricting them to be in exactly sets . This suggests that there exists an intermediate strategy mutation rate where the two cases are similar .
It has been shown that evolutionary dynamics in a structured population can be described by a single parameter , , if we are merely interested in the question , which of the two competing strategies , or , is more abundant in the limit of weak selection [41] . Payoff matrix ( 1 ) describes the interaction between the two strategies and and the inequality specifies that is more abundant than in the mutation-selection equilibrium . In general the parameter can depend on the population structure ( which specifies who interacts with whom for accumulating payoff and for evolutionary updating ) , the population size and the mutation rates; but it does not depend on the entries of the payoff matrix . The parameter has been explicitly calculated for a number of models including games on graphs , games in phenotype space , games in set structured populations and a simple model of multi-level selection [42] . Here we provide a general formula for the factor , which holds for the case of global updating . Global updating means that all members of the population compete globally ( as opposed to locally ) for reproduction . For example , global updating arises in the following way: one individual reproduces and another random individual dies ( in order to maintain constant population size ) ; the offspring of the first individual might inherit ( up to mutation ) the strategy and the ‘location’ of the parent . Global updating is a feature of models for games in phenotype space [36] and for games on sets [40] . Our main result , eq ( 3 ) , provides both an intuitive description of what the factor is and an efficient way for numerical computation .
|
At the center of any evolutionary process is a population of reproducing individuals . The structure of this population can greatly affect the outcome of evolution . If the fitness of an individual is determined by its interactions with others , then we are in the world of evolutionary game theory . The population structure specifies who interacts with whom . We derive a simple formula that holds for a wide class of such evolutionary processes . This formula provides an efficient computational method for studying evolutionary dynamics in structured populations .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion"
] |
[
"evolutionary",
"biology/evolutionary",
"ecology",
"evolutionary",
"biology/human",
"evolution",
"mathematics",
"computational",
"biology/evolutionary",
"modeling",
"evolutionary",
"biology"
] |
2009
|
Calculating Evolutionary Dynamics in Structured Populations
|
At the Drosophila NMJ , BMP signaling is critical for synapse growth and homeostasis . Signaling by the BMP7 homolog , Gbb , in motor neurons triggers a canonical pathway—which modulates transcription of BMP target genes , and a noncanonical pathway—which connects local BMP/BMP receptor complexes with the cytoskeleton . Here we describe a novel noncanonical BMP pathway characterized by the accumulation of the pathway effector , the phosphorylated Smad ( pMad ) , at synaptic sites . Using genetic epistasis , histology , super resolution microscopy , and electrophysiology approaches we demonstrate that this novel pathway is genetically distinguishable from all other known BMP signaling cascades . This novel pathway does not require Gbb , but depends on presynaptic BMP receptors and specific postsynaptic glutamate receptor subtypes , the type-A receptors . Synaptic pMad is coordinated to BMP’s role in the transcriptional control of target genes by shared pathway components , but it has no role in the regulation of NMJ growth . Instead , selective disruption of presynaptic pMad accumulation reduces the postsynaptic levels of type-A receptors , revealing a positive feedback loop which appears to function to stabilize active type-A receptors at synaptic sites . Thus , BMP pathway may monitor synapse activity then function to adjust synapse growth and maturation during development .
Bone morphogenetic proteins ( BMPs ) modulate a wide variety of cellular processes via canonical and noncanonical signaling pathways [1–3] . BMP signaling is initiated when extracellular dimeric BMP ligands bind to a heterotetrameric complex of Ser/Thr kinases , known as type I and type II BMP receptors ( BMPR ) . Following ligand binding , the type II receptor phosphorylates and activates the type I receptor , which in turn phosphorylates the intracellular R-Smad transducer ( Smad1 , 5 or 8 in vertebrates , and Mad in Drosophila ) [4 , 5] . Phosphorylated Smads ( pSmads ) associate with Co-Smads , and translocate into the nucleus where , in conjunction with other transcription factors , they regulate expression of target genes . Activated BMPRs can also signal independently of Smads through noncanonical pathways , which include various types of mitogen-activated protein kinase ( MAPK ) , LIM ( Lin-11/Isl-1/Mec-3 gene products ) kinase , phosphatidylinositol 3-kinase/Akt ( PI3K/Akt ) , and Rho-like small GTPases [1 , 6 , 7] . Intriguingly , pSmads also accumulate at the cell membrane in at least two instances: ( a ) at tight junctions during neural tube closure [8] , and ( b ) at the Drosophila neuromuscular junction [9 , 10] . If Smads are not involved in noncanonical BMP pathways and pSmads translocate to the nucleus in response to canonical BMP signaling , then how do pSmads accumulate at membrane locations and why ? During neural tube closure , pSmad1/5/8 binds to apical polarity complexes and mediates stabilization of BMP/BMPR complexes at tight junctions [8]; prolonged BMP blockade disrupts the tight junctions and affects epithelial organization [11] . Local pMad accumulation at the fly NMJ requires specific glutamate receptor subtypes [12] , but the nature and biological relevance of the pMad-positive puncta remain obscure . At the Drosophila NMJ , BMP signaling controls NMJ growth and promotes synapse homeostasis [13–17] . In the absence of BMP signaling , individual synapses form but the NMJs remain small , with fewer boutons , and exhibit numerous structural and functional defects . Ultrastructural studies indicate that BMP pathway mutants have enlarged active zones , with frequent detachments between the pre- and postsynaptic membranes [14–16] . These mutant NMJs have significantly reduced evoked potentials and lack the ability to induce homeostatic compensatory responses . It is generally thought that BMP signaling fulfills these functions via canonical and noncanonical pathways triggered by Glass-bottom boat ( Gbb ) , a BMP7 homolog , which binds to presynaptic BMPRII , Wishful thinking ( Wit ) , and BMPRIs , Thickveins ( Tkv ) and Saxophone ( Sax ) . The canonical pathway activates presynaptic transcriptional programs with distinct roles in the structural and functional development of the NMJ [18 , 19] . For example , the BMP pathway effector Trio , a Rac GEF , can rescue the NMJ growth in BMP pathway mutants , but does not influence synapse physiology , whereas Target of Wit ( Twit ) can partially restore the mini frequency in wit mutants but has no effect on NMJ growth . Besides the canonical BMP signaling , Gbb and the BMP type II receptor Wit signal through the effector protein LIMK1 to regulate synapse stability and addition of new boutons with increased synaptic activity [20 , 21] . LIMK1 is not required for Mad-mediated NMJ growth; instead , LIMK1 regulates the presynaptic actin dynamics partly by controlling the activity of the actin depolymerizing protein Cofilin . BMP signaling is perturbed in mutants that affect endocytosis , endosomal sorting and retrograde transport , which may disrupt the proper subcellular distribution and transport of BMP/BMPRs signaling complexes to the motor neuron soma [10 , 22 , 23] . During development , an early and transient BMP signal is both necessary and sufficient for NMJ growth and activity-dependent synaptic plasticity , whereas the control of NMJ function starts early , during late embryonic stages , and requires continuous BMP signaling throughout development [24] . We have recently discovered that pMad accumulates at synaptic terminals in response to specific glutamate receptor subtypes , the type-A receptors [12] . The fly NMJ ionotropic glutamate receptors ( iGluRs ) are heterotetrameric complexes composed of three essential subunits–GluRIIC , GluRIID and GluRIIE–and either GluRIIA ( type-A receptors ) or GluRIIB ( type-B ) [25–29] . The two receptors have identical single-channel conductances , but type-B receptors desensitize nearly ten times faster that type-A and have reduced quantal size ( the postsynaptic response to the fusion of single synaptic vesicles ) ( reviewed in [30] ) . GluRIIA and GluRIIB are similarly abundant at the larval NMJ , but they utilize distinct mechanisms for targeting and stabilization at synaptic sites . Furthermore , GluRIIA competes with GluRIIB for the limiting essential subunits , GluRIIC , -D and -E , so an increase in synaptic GluRIIA induces a decrease in GluRIIB [27] . Studies on mutant GluRIIA variants indicate that channel properties also influence trafficking and synaptic distribution of GluRIIA [26 , 31] . GluRIIA is both necessary and sufficient for the experience-dependent strengthening and growth of the NMJ [32] . pMad accumulates at synaptic terminals at the onset of synaptogenesis and mirrors postsynaptic GluRIIA throughout the NMJ development [12] . Here we examine the regulation and function of synaptic pMad during NMJ development . We found that synaptic pMad is generated in the presynaptic compartment via a novel BMP signaling pathway that is genetically distinguishable from the canonical BMP signaling and the Wit/LIMK1 noncanonical pathway . This novel pathway does not require Gbb , but depends on presynaptic Wit and Sax and postsynaptic GluRIIA receptors . Using genetic epistasis experiments , we demonstrate that synaptic pMad has no role in the regulation of NMJ growth . Instead , selective disruption of presynaptic pMad accumulation reduced postsynaptic GluRIIA levels , revealing a positive feedback loop which appears to function to stabilize active type-A receptors at synaptic sites .
Previous light microscopy studies argued that pMad localizes to both pre- and postsynaptic compartments [9 , 41] . However , loss of wit effectively eliminates the synaptic pMad signals [12 , 42] . As Wit is known to function and to be predominantly expressed in the motor neurons , it was inferred that synaptic pMad was presynaptic . Since 3D-SIM resolution is not sufficient to address this issue , we set up a tissue specific rescue experiment using a mad deficiency chromosome and a strong hypomorphic mad allele , mad12 , which produces a truncated Mad without the last 39 residues , including the site of BMP-dependent phosphorylation [43] . As expected , both synaptic and nuclear pMad signals were largely absent from these mad mutants ( mad12/Df ) , which die as translucent third instar larvae , with almost no fat body ( Fig 2 ) . Expression of Mad-GFP in pre-synaptic motor neurons of mad mutants did not rescue the adult viability but was sufficient to restore both nuclear and synaptic pMad levels during larval stages . In fact , the nuclear pMad levels were greatly increased ( by 685 ± 88% , n = 9 ) in the motor neuron nuclei of rescued animals as compared with wild-type controls . The synaptic pMad levels increased from 13 ± 2% at mad mutant NMJs to 58 ± 6% ( p<0 . 0001 , n = 20 ) in the rescued larvae relative to controls . In contrast , expressing Mad-GFP in the postsynaptic muscle did not restore synaptic pMad . These larvae had severely enlarged boutons , marked by GFP-positive signals , indicating that Mad-GFP accumulates at postsynaptic locations , but their synaptic pMad levels remained similar to those measured at mad mutant NMJs . Nuclear pMad was also largely absent from these rescued larvae , except for a small subset of neurons which express the 24B-Gal4 line used here . The fact that Mad expressed in motor neurons , but not in muscle , restored the synaptic pMad at mad mutant NMJs unambiguously demonstrates that synaptic pMad resides primarily in the presynaptic compartment . Elevated levels of synaptic pMad were previously correlated with synaptic overgrowth , in particular with the presence of supernumerary/satellite boutons observed in many endocytic mutants [44] . For example , synaptic pMad and the number of satellite boutons are elevated in the absence of Nervous wreck ( Nwk ) , an adaptor protein which appears to link Tkv , the type I BMPR , with the endocytic machinery [45] . To test whether synaptic pMad directly influences the formation of satellite boutons we set up a series of genetic epistasis experiments . We found that local pMad signals were completely abolished in nwk; IIA double mutants ( Fig 3A and 3B ) . This is consistent with our previous finding that synaptic pMad is absolutely dependent on postsynaptic GluRIIA . However , loss of synaptic pMad in the nwk; IIA double mutants did not alleviate the aberrant morphology observed in nwk mutants; these NMJs remained overgrown with a high number of satellite boutons , similar to mutations in nwk alone ( Fig 3A and 3C ) . Thus , synaptic pMad does not influence the nwk-dependent NMJ overgrowth . Drosophila NMJs grow exuberantly and are greatly expanded in the absence of Highwire ( Hiw ) , a conserved E3 ubiquitin ligase that limits synaptic growth [17 , 46 , 47] . It has been shown that BMP signaling mutants suppress the excessive NMJ growth of hiw mutants [17] . We found that NMJs in hiw; IIA double mutants selectively lose synaptic pMad signals , but they remain overgrown , and resemble the hiw mutant NMJs ( Fig 3A , 3B and 3C ) . These findings indicate that synaptic pMad is not required for the NMJ overgrowth . Previous studies demonstrate that synaptic pMad signals are selectively lost in importin-ß11 mutants ( impß11 ) but could be restored by excess presynaptic BMPRs or by blocking retrograde transport in the motor neurons [42] . Since Nwk limits the retrograde BMP signaling partly by controlling Tkv turnover [45] , we asked whether nwk is epistatic to impß11 . We found that the local pMad levels at impß11; nwk double mutant NMJs were restored to normal levels ( Fig 3B–3D ) . However , the NMJ morphology of impß11;nwk double mutants resembled the characteristic nwk mutant phenotype , i . e . overgrown NMJs with numerous satellite boutons . These results indicate that nwk is epistatic to impß11 in controlling the NMJ morphology and growth , but impß11 and nwk appear to influence the local pMad accumulation through distinct pathways . Intriguingly , overexpression of Mad cannot rescue the loss of synaptic pMad at impß11 NMJs [42] . How can excess BMPRs restore the pMad signals at impß11 NMJs while excess Mad cannot ? Since synaptic pMad presumably marks local BMP/BMPR active complexes , excess neuronal BMPRs may restore the synaptic BMP/BMPR pool and thus rescue the accumulation of pMad at impß11 mutant NMJs , whereas in the absence of synaptic BMP/BMPR complexes , Mad could not be phosphorylated locally even when in excess . We have previously shown that GluRIIA muscle expression increases local pMad due to increased synaptic type-A receptors [12] . Interestingly , GluRIIA muscle expression efficiently rescued the pMad levels at impß11 NMJs , although it did not alleviate their growth defects ( Fig 3B–3D ) . This result has two implications: First , it demonstrates that GluRIIA is sufficient for presynaptic pMad accumulation . Second , since presynaptic BMPRs also restore local pMad at impß11 NMJs [42] , then postsynaptic GluRIIA and presynaptic BMPRs likely function together to trigger pMad accumulation at synaptic terminals . Impß11 presumably limits the retrograde transport of BMP/BMPR complexes and thus further stabilizes local pMad . Additional endocytic components including Spichthyin , Endophilin , Spinster and Liquid facets , the Drosophila homolog of Epsin1 , limit the local pMad pool [23 , 48] . We found that removal of GluRIIA in any of the endocytic mutants tested induced complete loss of synaptic pMad . Thus , GluRIIA is both required and sufficient for the synaptic accumulation of pMad . While genetic manipulations of postsynaptic GluRIIA receptors induced proportional changes in the level of synaptic pMad , such manipulations had no detectable effect on the nuclear pMad [12] . This implies that genetically distinct pathways regulate the nuclear and local pMad pools . Smad levels and activities are tightly controlled by posttranslational regulation [49 , 50] . To test whether the levels of Mad are limiting for synaptic pMad accumulation , we examined the effect of excess Mad . Overexpression of Mad-GFP in motor neurons , but not in muscles , produced a dramatic increase of pMad in motor neuron nuclei ( Fig 4A ) . However , the synaptic pMad remained unchanged irrespective of excess Mad in the pre- and/or postsynaptic compartments ( Fig 4B and 4C ) . No change in local pMad was observed for various tagged and non-tagged Mad transgenes ( see below ) . These Mad transgenes were functional as indicated by their ability to induce increased expression of twit , a BMP transcriptional target ( Mad-Myc shown in Fig 4D ) . Twit levels were directly monitored via a MiMIC insertion line , twitMI06552 [51] , which generates a Twit-GFP chimera: Twit-GFP was reduced in the absence of wit , and was strongly increased when Mad was overexpressed in motor neurons ( Fig 4D ) . Overexpression of Mad-GFP in GluRIIA mutants induced high levels of nuclear pMad , but pMad was undetectable at synaptic locations ( Fig 4E ) . Lack of synaptic pMad cannot be explained by a deficit in Mad-GFP axonal trafficking and/or local translation since high levels of Mad-GFP accumulated at synaptic terminals in control and GluRIIA mutants ( Fig 4F ) . Instead , our data demonstrate that synaptic pMad is absolutely dependent on postsynaptic GluRIIA receptors and is not limited by the net levels of Mad . While Gbb is absolutely required for the nuclear accumulation of pMad and transcriptional regulation of BMP target genes , we found that Gbb is dispensable for local pMad accumulation ( Fig 5A and 5B ) . In gbb mutant animals , the mean pMad level per NMJ was similar in intensity to control animals ( w1118 ) ( quantified in S1 Fig ) . Significant levels of synaptic pMad were observed in animals transheterozygous for different combinations of gbb null alleles , ruling out the contribution of the genetic background to this unexpected result . In contrast , synaptic pMad was completely lost at wit mutant NMJs ( Fig 5B and [12] ) . Super resolution imaging of gbb mutant boutons showed enlarged active zones ( Fig 5C ) , consistent with ultrastructural defects reported for BMP pathway mutants [14 , 17] . In particular , a striking feature of BMP mutant synapses is increased accumulation of presynaptic T-bar material , some of it detached from the presynaptic membrane , and the appearance of active zone profiles with two or more T-bars , presumably due to perturbed Brp recruitment [24 , 32] . Indeed , at gbb mutant boutons we found many synapses decorated by multiple or enlarged Brp-positive rings . These enlarged Brp domains juxtaposed large Neto-positive fields that appear to include both enlarged ( arrows ) and multiple ( arrowheads ) PSDs ( Fig 5D and 5E and S4 and S5 Movies ) . Nonetheless , pMad was present at all gbb mutant active zones—adjacent to Brp and partially overlapping with Neto . These data indicate that Gbb and canonical BMP signaling regulate the assembly and/or maintenance of presynaptic T-bars , likely via BMP transcriptional targets . However , Gbb is not essential for the accumulation of pMad at active zones . Previous studies have reported an absence of synaptic pMad in gbb mutants , or residual pMad signals due to leaky transgenes [24] . We considered whether environmental conditions could account for difference in synaptic pMad levels in our studies versus others , especially since gbb has been implicated in regulating nutrient storage and energy homeostasis [52] . It has been found empirically that more gbb mutant larvae progress to third instar when raised on a yeast-rich diet . We found that rearing gbb mutants exclusively on yeast paste induced a drastic reduction of local pMad ( S1 Fig ) . Furthermore , wild-type animals reared on yeast paste also showed a dramatic reduction of synaptic pMad ( up to 5 fold ) but normal nuclear pMad . The loss of synaptic pMad could be caused by reduced NMJ activity and limited locomotion as these animals no longer roam for food , or may reflect a response to the nature of food and/or other factors . Nonetheless , our results underscore the importance of rearing conditions when examining BMP signaling at the larval NMJ . Similar to controls , knockdown of GluRIIA in the gbb mutant background drastically diminished the synaptic pMad levels , indicating that local pMad requires GluRIIA , albeit not Gbb ( Fig 6A ) . Furthermore the local pMad was lost at gbb; wit double mutant NMJs ( Fig 6B ) , demonstrating that Wit is absolutely required for the accumulation of synaptic pMad . Together these data show that Gbb , but not Wit or GluRIIA , is dispensable for the synaptic pMad accumulation . Other ( s ) ligand may promote accumulation of synaptic BMPR complexes that phosphorylate Mad locally . It was previously shown that glia-derived Maverick ( Mav ) is secreted in the synaptic cleft and signals to the postsynaptic muscle via the BMPR type II Punt ( Put ) to modulate NMJ development partly by controlling Gbb expression [41] . Intriguingly , knockdown of Mav in glia also triggered a dramatic reduction of synaptic pMad , raising the possibility that Mav may also signal to the presynaptic neuron to influence pMad accumulation . To test this scenario , we first confirmed that depletion of glia-derived Mav using the available TRiP lines reduced the synaptic pMad levels ( Fig 6C ) . We found that RNAi-mediated reduction of Mav in the glia induced a significant decrease of synaptic pMad at control NMJs as well as in a gbb mutant background . A similar reduction was observed by knockdown of Put , the Mav receptor , in the striated muscle , suggesting that Mav signaling to the muscle may explain its effect on synaptic pMad ( see below ) . In Drosophila , two type I receptors , Tkv and Sax , transduce the BMP-type signals . At the NMJ , both Tkv and Sax are required for nuclear pMad accumulation and NMJ growth [17] . Furthermore , excess activated Tkv in the motor neurons induces increased synaptic pMad [45] , and excess activated Tkv and Sax restores the pMad signals at impß11 NMJs [42] . We found that Sax is also required for synaptic pMad accumulation ( Fig 7A ) . sax null third instar mutants ( sax4/Df ) showed dramatically reduced levels of synaptic pMad and practically no detectable nuclear pMad signals above background . Thus , similar to Wit , Sax appears to be required for both nuclear and synaptic pMad accumulation . In flies , as well as vertebrates , Wit has a large intracellular domain that binds and signals through LIM kinase 1 ( LIMK1 ) to regulate synapse stability and activity-dependent synaptic growth [20 , 21] . To test whether LIMK1 influences the local pMad accumulation we examined the NMJs of witΔC genomic animals , which lack the LIMK1-binding intracellular part of Wit [15] . Synaptic pMad was restored at wit mutant NMJs by either control or witΔC genomic transgenes ( Fig 7D and 7E ) . Thus , Wit is essential for both nuclear and synaptic pMad accumulation , but neither of these functions requires its interaction with LIMK1 [14 , 15] . In contrast , Gbb is required for nuclear pMad and the transcriptional control of BMP target genes , but appears to be dispensable for pMad accumulation at active zones . This implies that the differences between the wit and gbb mutant NMJs should reflect , at least in part , a role for synaptic pMad during NMJ development . Mutations in either wit or gbb induce severe deficits in NMJ growth and function , but only wit mutant NMJs have reduced mini amplitude , or quantal size , the postsynaptic response to the spontaneous fusion of a single synaptic vesicle [14 , 19] . In contrast , gbb mutants have relatively normal quantal size [16 , 53] . Previous studies established that a key determinant of quantal size is the dose of synaptic GluRIIA versus GluRIIB ( IIA/IIB ) [25 , 26] . Therefore , the difference in quantal size in wit and gbb mutants may arise from different IIA/IIB synaptic composition . We tested this prediction by examining the relative GluRIIA and GluRIIB signal intensities at mutant NMJs . We found that the GluRIIA and GluRIIB synaptic signals were reduced by more than 30% ( to 67 ± 8% and 65 ± 9% respectively of controls , n = 19 ) at gbb mutant NMJs , but the IIA/IIB ratio remained normal ( Fig 8A and 8B ) . In contrast , the GluRIIA and GluRIIB signals were severely and unequally reduced in wit mutants , such as that the IIA/IIB ratio was reduced to 47 ± 7% , n = 18 . A similar , asymmetrical reduction of GluRIIA and GluRIIB synaptic signals and a decreased IIA/IIB ratio was also apparent at mad mutant NMJs ( Fig 8A and [54] ) . This is consistent with a small reduction in quantal size observed in mad hypomorphs or dominant negative allelic combinations [17 , 24] . Interestingly , a subtle reduction in the IIA/IIB ratio was also observed at imp mutant NMJs , consistent with the previously described 20% decrease in quantal size [42] . Overall , the differences in GluRIIA levels were smaller than those observed for synaptic pMad ( Figs 3 and 7 and [12 , 42] ) . This may be due to GluRIIA receptors present at synaptic sites but in configurations that cannot trigger local pMad accumulation [12 , 55] . In contrast , depletion of Mav in the glia produced reduction of both GluRIIA and GluRIIB subunits but did not significantly alter the postsynaptic IIA/IIB ratio ( Fig 8C and 8D ) . The same result was found when Put was depleted in the striated muscle . This severe reduction in the synaptic distribution of both GluRIIA and GluRIIB subunits is reminiscent of Activin-type signaling [54] . It has been shown that motor neuron-derived Activin signals via Baboon , the type I TGF-β receptor , and Smox/dSmad2 , the pathway effector , to regulate the expression of GluRIIA and GluRIIB in the muscle . In the absence of Activin pathway components , GluRIIA and GluRIIB transcripts are dramatically reduced , and GluRIIA synaptic distribution is further diminished via post-translational mechanisms . The similarities between Mav/Put and Act/Babo/Smox signaling pathways suggest that these pathways may share common TGF-β receptors ( i . e . Put ) and converge , at least in part , onto subsets of transcriptional targets , such as GluRIIA and GluRIIB . While these data do not exclude a more direct role for Mav ( or Act ) in the synaptic accumulation of pMad , these pathways likely influence local pMad indirectly , via postsynaptic GluRIIA , which is both required and sufficient for synaptic pMad accumulation ( Fig 3 ) . At the Drosophila NMJ , the type-A receptors are the first to arrive at nascent synapses , followed by the type-B receptors , which mark more mature synapses [56] . This ordered incorporation of iGluR subtypes is modulated in part by Neurexin ( Nrx ) and Neuroligin1 ( Nlg1 ) , a pair of conserved adhesion molecules that form trans-synaptic complexes which stabilize synaptic contacts and organize receptor fields [57] . Lack of Nrx or Nlg1 perturbs the dynamics of iGluRs recruitment and stabilization at the fly NMJ and causes morphological and physiological defects [58–60] . We found that the IIA/IIB ratio was increased by 25% at nrx mutant NMJs ( Fig 9A and 9B ) . This change is consistent with the increased quantal size reported for the nrx mutants [60] . Importantly , synaptic but not nuclear pMad levels were increased by 50% in third instar nrx larvae ( Fig 9C and 9D ) . A similar , but less dramatic increase of synaptic pMad was found in nlg1 mutants . In these mutants pMad immunolabeling produced high background , with patches of pMad-positive areas outside of the motor neuron nuclei and of the synaptic terminals; such patches were not included in our quantifications . As expected , IIA; nrx double mutants showed a complete loss of synaptic pMad without any change in the nuclear pMad levels ( Fig 9E ) . The morphology of IIA; nrx double mutant NMJs resembled the nrx mutants , with fewer and larger boutons , grouped closer together [60] . Thus , the iGluR subtypes and local pMad do not influence the NMJ morphology of nrx mutants . This result is similar to that observed for nwk or hiw mutants ( Fig 3 ) . Unlike nwk , which limits Tkv endocytosis and restores local pMad at imp; nwk double mutant NMJs , loss of Nrx could not rescue the synaptic pMad in imp; nrx mutants ( Fig 9E and 9F ) . This suggests that Nrx influences the local pMad indirectly , perhaps by limiting postsynaptic GluRIIA . Together our data indicate that postsynaptic GluRIIA and presynaptic BMPRs are key determinants for the accumulation of pMad at active zones . Furthermore , BMP signaling modulators ( i . e . Nwk and Imp ) , acting in the presynaptic compartment , control the levels of synaptic pMad . Intriguingly , increased synaptic pMad ( such as in nrx mutants ) correlates with increased IIA/IIB ratio and increased quantal size , while loss of synaptic pMad ( in imp , wit and mad mutants ) correlates with a decreased IIA/IIB ratio and reduced quantal size . In contrast , the presence of synaptic pMad even in a transcriptionally impaired BMP mutant ( i . e . gbb ) ensured relatively normal IIA/IIB ratio and quantal size . This tight correlation suggests a feedback mechanism whereby active postsynaptic GluRIIA receptors induce the accumulation of pMad at active zones , which in turn promotes the stabilization of GluRIIA receptors at postsynaptic sites . In this scenario , selective disruption of synaptic pMad should “destabilize” the GluRIIA receptors and cause decreased IIA/IIB ratio . To test for such a positive feedback loop we have to disrupt the local pMad accumulation without affecting the canonical BMP signaling and transcription of BMP target genes . This precludes the use of any BMP signaling components or known BMP modulators , as any such manipulations will affect both local and transcriptional functions of BMP pathway . Previous studies demonstrate that Mad phosphorylation at S25 by Nemo kinase , a MAPK-related kinase , promotes nuclear export of pMad in heterologous cells [61] , whereas nemo mutant NMJs have increased local pMad [62] . Nemo does not appear to interfere with the ability of BMP/BMPR complexes to phosphorylate Mad at its C-terminal residues; instead , Nemo influences the subcellular distribution of Mad irrespective of its BMP-dependent phosphorylation status . Interestingly , nemo mutants have reduced synaptic pMad and neuronal overexpression of activated Tkv , but not Mad , could rescue this deficit , suggesting that BMPRs become limiting in nemo mutants [62] . Since the levels of BMPR are tightly controlled [45] , and lack of S25 phosphorylation ( in nemo mutants ) increases the synaptic pMad and promotes the pMad-BMPRs association , Nemo-dependent phosphorylation may provide a means for regulating the stability of pMad-BMP/BMPR complexes . This predicts that S25 and ( Nemo-phosphorylated ) pS25 will have opposing effects on the formation and stabilization of Mad-BMP/BMPR complexes . We reasoned that overexpression of a Nemo-phosphomimetic Mad variant ( S25D ) in the motor neurons should not affect the nuclear pMad pool since excess MadS25D will be efficiently exported from the motor neuron nuclei . However , at active zones , excess MadS25D will compete with the endogenous Mad for BMPR-mediated phosphorylation , but will presumably dissociate from the presynaptic BMP/BMPR complexes , thus diminishing the local pMad accumulation . Indeed , we found that neuronal overexpression of a MadS25D transgene induced a minimal increase in the nuclear pMad pool ( 20% more than controls , n = 10 ) but significantly reduced the synaptic pMad levels ( to 55% of control , n = 24 ) ( Fig 10A , 10B and 10C ) . By comparison , neuronal overexpression of wild-type Mad , or the phospho-mutant variant MadS25A , induced a 4-fold increase in nuclear pMad and no change in the synaptic pMad levels ( Figs 4 and S2 ) . More importantly , neuronal expression of MadS25D caused a drastic reduction in the postsynaptic GluRIIA levels ( Fig 8C and 8D ) . At the same time the GluRIIB levels were increased , likely because of the competition between GluRIIA and GluRIIB for the essential iGluR subunits [27] . The excess neuronal MadS25D did not influence the NMJ morphology and growth and did not affect the Twit-GFP levels , indicating normal transcriptional regulation in response to the canonical BMP signaling . The number of synaptic contacts and the iGluR receptor fields , as visualized by juxtaposed Brp and GluRIIC signals , also appeared normal ( S3 Fig ) . The only change we could detect in animals with excess presynaptic MadS25D/reduced synaptic pMad was a reduced IIA/IIB ratio . To examine whether this change in the distribution of iGluR subtypes influences NMJ function , we performed electrophysiology recordings of spontaneous junction currents and potentials from muscle 6 of control and third instar larvae expressing various Mad transgenes in motor neurons ( Figs 10E–10I and S4 ) . Consistent with the reduced IIA/IIB ratio observed , neuronal overexpression of MadS25D , but not MadS25A , caused a 22% reduction in mEJC amplitude ( 0 . 58 ± 0 . 03 nA for MadS25D vs control 0 . 74 ± 0 . 04 nA , or MadS25A 0 . 73 ± 0 . 03 nA , p < 0 . 01 ) . Furthermore , the decay time constant was decreased when MadS25D was overexpressed in the motor neurons ( to 4 . 85 ± 0 . 14 ms for MadS25D , comparing to control 6 . 61 ± 0 . 74 ms , and MadS25A 5 . 66 ± 0 . 40 ms , respectively; p < 0 . 05 ) . Since GluRIIB-containing receptors desensitize much faster that the GluRIIA [26] , these data are consistent with the observed shift towards more GluRIIB postsynaptic receptors at excess MadS25D . Overexpression of the phospho-mutant variant MadS25A had no significant effect on the mEJC amplitude and decay constant , but produced a strong reduction of mEJC frequency . This may be due to excess nuclear MadS25A and perturbed expression of BMP target genes , including twit , which encodes a modulator of mini frequency . Similarly , neuronal overexpression of MadS25D , but not MadS25A , caused a 25% reduction in mEJP amplitude ( *p = 0 . 020 , S4 Fig ) . We found no change in the resting potential and input resistance in these larvae . As in the case of GluRIIA mutants , the amplitude of evoked junctional potentials remained normal at NMJs with excess neuronal MadS25D ( p = 0 . 555 , S4 Fig ) , demonstrating a compensatory increase in quantal content , the number of vesicles released in response to each action potential . We estimated the quantal content by dividing the mean EJP amplitude to mean mEJP and found an 80% increase in quantal content ( *p = 0 . 032 , S4 Fig ) at NMJs with excess MadS25D . This indicates a robust presynaptic response to the MadS25D-induced reduction of IIA/IIB ratio . These findings suggest that diminished synaptic pMad in the motor neurons causes a reduction of postsynaptic GluRIIA and induces a change in the synaptic accumulation of iGluR subtypes towards more type-B receptors . The reduction of IIA/IIB ratio , evident in histology as well as electrophysiology experiments , induced a significant increase in quantal content compared with control . Such compensatory response in presynaptic transmitter release is characteristic of low levels of postsynaptic GluRIIA [26] . Thus , synaptic pMad accumulates in response to active GluRIIA and , in turn , appears to stabilize the type-A receptors at synaptic sites . This positive feedback could shape the synaptic composition for iGluRs as a function of type-A receptor activity .
At the Drosophila NMJ , BMP signaling controls NMJ growth and promotes synapse homeostasis [14–16 , 63] . BMP fulfills all these functions via canonical and noncanonical pathways . Canonical BMP signaling activates presynaptic transcriptional programs with distinct roles in the structural and functional development of the NMJ [18 , 19] . For example , the BMP pathway effector Trio can rescue NMJ growth in BMP pathway mutants , but does not influence synapse physiology , whereas Twit can partially restore the mini frequency but has no effect on NMJ growth . It has been shown that both muscle and neuron derived Gbb are required for the structural and functional integrity of NMJ , and multiple mechanisms that regulate Gbb expression , secretion and extracellular availability have been described [41 , 53 , 64 , 65] . Binding of Gbb to its receptors also triggers a noncanonical , Mad-independent pathway that requires the C-terminal domain of Wit . This domain is conserved among Drosophila Wit and vertebrate BMPRII and functions to recruit and activate cytoskeletal regulators such as LIMK1 [66 , 67] . In flies , Wit-mediated activation of LIMK1 mediates synapse stability and enables rapid , activity-dependent synaptic growth [20 , 21] . In this study we uncovered a novel , noncanonical BMP pathway that triggers accumulation of presynaptic pMad in response to postsynaptic GluRIIA receptors . This pathway requires Wit and Sax , suggesting that various BMP pathways compete for shared components . Super resolution imaging mapped the pMad domains at active zones , in close proximity to the presynaptic membrane . These domains concentrate the pMad immunoreactivities into thin discs that reside mostly within individual synapse boundaries . The size and shape of pMad domains suggest that pMad could associate with membrane-anchored complexes at the active zone . Since BMP signals are generally short lived [33 , 34 , 68] , these pMad domains likely represent pMad that , upon phosphorylation , remains associated with the BMP/BMPR kinase complexes at synaptic sites . Alternatively , pMad may accumulate in synaptic aggregates that protect it from dephosphorylation . While we cannot exclude the second possibility , several lines of evidence support the first scenario . First , excess Mad cannot increase the levels of synaptic pMad ( Fig 4 ) . Second , neuronal expression of activated Tkv/Sax but not Mad can restore the synaptic pMad at impß11 mutant NMJs [42] . Finally , during neural tube closure , junctional pSmad1/5/8 and its association with PAR complexes depend on BMPs [8] . Previous studies indicate a reduction of synaptic pMad signals in response to muscle-specific Mad RNAi [9 , 41] . We too have observed such a reduction ( S5 Fig ) . In addition , we found a significant decrease of postsynaptic IIA/IIB ratio in Mad-depleted muscles: GluRIIA and GluRIIB synaptic levels were reduced to 49% and respectively 78% of control ( n = 21 ) . Since GluRIIA is key to the synaptic pMad accumulation we suspect that the muscle Mad RNAi phenotype is due to perturbation in synaptic GluRIIA levels , perhaps by interference with the Activin signaling pathway ( see below ) . How are the BMP/BMPR complexes stabilized at synaptic sites ? Studies on single receptors demonstrate that the confined mobility of BMPRI on the plasma membrane is key to stabilize BMP/BMPR complexes and differentially stimulate canonical versus noncanonical signaling [69] . Direct interactions between phosphorylated Smad5 and the Par3-Par6-aPKC polarity complex occur at the apical junctions [8] . Similarly , synaptic pMad , which remains associated with BMP/BMPR complexes , may engage in interactions that restrict the mobility of BMP/BMPR complexes on the presynaptic membrane . Nemo-mediated phosphorylation of Mad-S25 could disrupt the pMad/BMPR association and expose the BMP/BMPR complexes , so they could dissociate and/or be internalized . The heteromeric BMPR complexes are transient; ligand binding greatly increases their lifespan and stability [70] . Albeit Gbb is not essential for synaptic pMad , it may act redundantly with other ligands to stabilize BMP/BMPR local complexes . Several ligands secreted in the synaptic cleft have been shown to bind and signal via BMPRII; they include glia secreted Maverik [41 , 71] , Myoglianin , which could be secreted from muscle and/or glia [72 , 73] , and Activins [74] . However , these ligands also appear to signal via a canonical Activin pathway , which regulates the postsynaptic GluRIIA/GluRIIB abundance at the Drosophila NMJ [54] . Alterations in the Activin signaling pathway drastically alter the synaptic recruitment of both iGluR subtypes , in particular the GluRIIA , which controls synaptic pMad , making it difficult to identify the nature and the directionality of the signaling molecule ( s ) involved in the synaptic pMad accumulation . Interestingly , all of these ligands are substrates for BMP-1/Tolloid-type enzymes , which control their activity and distribution [75] . Treatments that induce long-term stimulation up-regulate a BMP-1/Tolloid homolog in Aplysia neurons [76] . An intriguing aspect of this novel BMP pathway is the dependence on active postsynaptic GluRIIA , which is both required and sufficient for pMad accumulation at active zones . Since pMad and BMP/BMPR complexes cluster at synaptic sites ( Fig 1 and [45] ) , we speculate that trans-synaptic complexes may couple postsynaptic type-A glutamate receptors with presynaptic BMP/BMPRs . The synaptic cleft is 200 Å; the iGluR tetramer expands 135 Å in the synaptic cleft [77] , and the BMP/BMPR complexes ~55 Å [78 , 79] . The iGluRs auxiliary subunit Neto has extracellular CUB and LDLa domains predicted to expand 120–130 Å in the synaptic cleft , based on related structures . CUB domains are BMP binding motifs [80] that may localize BMP activities and/or facilitate ligand binding to BMPRs . In this model , Neto provides the link between postsynaptic GluRIIA and presynaptic BMP/BMPR complexes . During receptors gating cycle , the iGluRs undergo corkscrew motions that shorten the channels as revealed by cryo-electron microscopy [81] . Such movements may influence the stability of trans-synaptic complexes and allow synaptic pMad to function as a sensor for GluRIIA activity [12] . While more components of this novel pathway remain to be determined , it is clear that this pathway does not contribute to NMJ growth and instead has a critical role in synapse maturation . Unlike canonical BMP signaling , loss of local pMad induces minor reductions in bouton number [12 , 25 , 82] and does not rescue the NMJ overgrowth of endocytosis mutants ( Fig 3 ) . Local pMad accumulates independently of Wit-mediated LIMK1 activation and does not appear to influence synapse stabilization; in fact , nrx mutants have synapse adhesion defects [60] but show increased synaptic pMad levels ( Figs 7 and 9 ) . The striking correlation between synaptic pMad levels and GluRIIA activity , together with previous findings that GluRIIA activity and gating behavior directly impacts receptor mobility and synaptic stabilization [26 , 31] suggest a positive feedback mechanism in which active GluRIIA receptors induce stabilization of BMP/BMPR complexes at synaptic sites which , in turn , promote stabilization of type-A receptors at PSDs . In this scenario , presynaptic pMad marks active BMP/BMPR complexes and acts to maintain the local BMP/BMPR complexes in large clusters that evade endocytosis . Selective disruption of local pMad via a neuronal dominant-negative MadS25D presumably destabilizes the large presynaptic BMP/BMPR clusters and causes a significant reduction in the IIA/IIB ratio and quantal size ( Fig 10 ) . This positive feedback couples synaptic activity with synapse development and is controlled by ( 1 ) active GluRIIA receptors , ( 2 ) presynaptic BMP receptors , Wit , Sax , and likely Tkv , ( 3 ) mechanisms regulating BMPR heteromers assembly , endocytosis and turnover , and ( 4 ) the ability of pMad to remain associated with its own kinase upon phosphorylation . Perturbations of any of these components trigger variations in local pMad levels accompanied by changes in the IIA/IIB ratio and/or quantal size . For example , nemo mutants have increased synaptic pMad levels and increased mEJCs [62] , while imp mutants have decreased synaptic pMad levels and decreased mEJPs ( Fig 3 and [42] ) . The assembly and function of these putative trans-synaptic complexes , in particular ligand availability , should be influenced by the composition and organization of the synaptic cleft . Indeed , local pMad and quantal size are increased in mutants lacking heparan sulfate 6-O-endosulfatase ( sulf1 ) , or 6-O-sulfotransferase ( hs6st ) [65] . Since this Mad-dependent , noncanonical pathway shares components with the other BMP signaling pathways , the balance among different BMP pathways may coordinate the NMJ development and function . The complexity of BMP signaling at the Drosophila NMJ is reminiscent of the neurotrophin-regulated signaling in vertebrate systems ( reviewed in [83] ) . Neurotrophins were first identified as neuronal survival factors . Like BMPs , they are secreted as pro-proteins that must be processed to form mature ligands . The active dimers bind to transmembrane kinase receptors and induce their activation through trans-phosphorylation . Neurotrophin/receptor complexes are internalized and transported along axons to the cell soma [84]; signaling in the cell soma controls gene expression and promotes neuronal differentiation and growth . In addition , local neurotrophin signaling regulates growth cone motility , enhances the presynaptic release of neurotransmitter and mediates activity-dependent synapse formation and maturation ( reviewed in [85] ) . At the Drosophila NMJ , several neurotrophins have been implicated in neuron survival , axon guidance and synapse growth [86–88] . It will be interesting to test for the crosstalk between neurotrophin and BMP signaling at these synapses . To our knowledge , the novel noncanonical BMP pathway reported here is the first example of a BMP pathway triggered by selective neurotransmitter receptors and influencing receptor distribution at PSDs . We expect that some of these functions will apply to mammalian glutamatergic synapses: First , as indicated in the Allen Brain Atlas , glutamate receptors and Neto proteins are widely expressed in mammalian brain structures where BMPs , BMPRs and Smads are expressed . Second , BMPs have been shown to rapidly potentiate glutamate-mediated currents in human retina neurons , presumably via a noncanonical pathway [89] . Finally , mice lacking Chordin , a BMP antagonist , have enhanced paired-pulse facilitation and LTP and show improved learning in a water maze test [90] . Such changes could not be explained by Smad-dependent transcriptional responses and were not accompanied by structural alterations in synapse morphology . Instead , presynaptic noncanonical BMP pathway may influence the activity of postsynaptic glutamate receptors by modulating their synaptic distribution and stability .
Drosophila stocks used in this study are as follows: GluRIIASP16 , Df ( 2L ) clh4 , and UAS-GluRIIA [25] ( from A . DiAntonio , Washington University ) ; hiwND8 [46]; impß1124 and impß1170 [42]; twitMI06552 [51]; mad12 [91]; mad deficiency Df ( 2L ) C28 [92]; UAS-Mad-GFP [9] ( from M . Gonzalez-Gaitan , University of Geneva ) ; UAS-Mad-Myc [62]; UAS-T7-MadS25A [61]; gbb1 and gbb2 [93]; sax4 [94]; witΔC . genomic , witA12 and witB11 [15] ( from M . O’Connor , University of Minnesota ) ; nwk1 , nwkγ3 [95] ( from K . O’Connor-Giles , University of Wisconsin ) ; nrx273 and nlg1Δ46 [96] ( from B . Mozer , NIH ) ; BG380-Gal4 [97]; elav-Gal4 ( BL-8760 ) ; 24B-Gal4 ( BL-1716 ) ; G14-Gal4 and MHC-Gal4 ( from C . Goodman , University of California at Berkeley ) . For UAS-T7-MadS25D , the T7-tagged MadS25D [61] was cloned into pUAST and transgenic lines were generated by germline transformation ( BestGene ) . For RNAi-mediated knockout we used UAS-put-RNAi and UAS-mad-RNAi ( ID 848 and respectively 12635 , Vienna Drosophila RNAi Center ) and TRiP lines generated by the Transgenic RNAi Project , Harvard Medical School , GluRIIA ( P[TRiP . JF02647]attP2 ) , and mav ( P[TRiP . HMS01125]attP2 and P[TRiP . GL01025]attP40 ) . The flies were reared on Jazz-Mix food ( Fisher Scientific ) . To control for larvae crowding , 8–10 females were crossed with 5–7 males per vial and were passed to fresh vials every 3 days . For rearing on yeast , embryos were collected on grape agar plates for 24 hours , incubated at 25°C for 24 hours and then 50 first instar larvae of appropriate genotype were transferred to a vial of standard fly food containing a 20 mm2 paper saturated with 20% ( w/v ) active dry baker’s yeast in water . Larvae remained on paper and did not burrow into food . Fresh yeast solution was added daily to keep paper saturated . Larvae were kept on yeast at 25°C until reaching third instar stage . Larvae were dissected as described previously in ice-cooled Ca2+-free HL-3 solution [98 , 99] . The samples were fixed in either 4% formaldehyde ( Polysciences , Inc . ) for 25 min or in Bouin’s fixative ( Bio-Rad ) for 3 min and washed in PBS containing 0 . 5% Triton X-100 . Primary antibodies from Developmental Studies Hybridoma Bank were used at the following dilutions: mouse anti-GluRIIA ( MH2B ) , 1:200; rat anti-Elav ( 7E8A10 ) , 1:200; mouse anti-Bruchpilot ( Brp ) ( Nc82 ) , 1:200 . Other primary antibodies were as follows: rabbit anti-phosphorylated Mothers against decapentaplegic ( pMad ) , 1:500 , ( a gift from Carl Heldin ) [100]; rabbit anti-pSmad3 , 1:500 , ( Epitomics , [10] ) ; FITC- , rhodamine- , and Cy5- conjugated goat anti-HRP , 1:1000 ( Jackson ImmunoResearch Laboratories , Inc . ) ; rabbit anti-GFP , 1:250 ( Abcam ) ; rat anti-Neto , 1:1000 [37]; Cy5- conjugated goat anti-HRP , 1:1000 ( Jackson ImmunoResearch Laboratories , Inc . ) . The rabbit polyclonal anti-GluRIIB and anti-GluRIIC were generated as previously described [26] against synthetic peptides ASSAKKKKKTRRIEK , and QGSGSSSGSNNAGRGEKEARV respectively ( Pacific Immunology Corp ) . Alexa Fluor 488- , Alexa Fluor 568- , and Alexa Fluor 647- conjugated secondary antibodies ( Molecular Probes ) were used at 1:400 . Larval filets were mounted in ProLong Gold and brains were mounted in SlowFade Gold ( Invitrogen ) . Samples of different genotypes were processed simultaneously and imaged under identical confocal settings using laser scanning confocal microscopes ( CarlZeiss LSM780 ) . Boutons were counted using anti-HRP immunoreactivities . All quantifications were performed while blinded to genotype . The numbers of samples analyzed are indicated inside the bars . All the pMad data quantified here were obtained using the anti-pMad serum from Carl Heldin . For these analyses , the samples were fixed with either 4% formaldehyde or Bouin fixative as described above , washed extensively in PBS containing 0 . 5% Triton X-100 , marked by distinct cuts per genotype and pooled together in the same tube , and , without any blocking agent , incubated overnight at 4°C with anti-pMad 1:500 plus other relevant primary antibodies . The samples were then washed with PBS containing 0 . 5% Triton X-100 , incubated with secondary antibodies ( 1:200 ) and no blocking agent for either 2 hours at room temperature or overnight at 4°C . After extensive washes with PBS containing 0 . 5% Triton X-100 , the samples were mounted as above , then imaged and quantified together . For motor neuron nuclei , confocal regions of interest ( ROIs ) were determined with Imaris software ( Bitplane ) by using the “spots” feature to automatically identify motor nuclei using Elav immunoreactivity . Spots were verified manually and mean center intensity for all nuclei in a given sample was recorded . This procedure was repeated for all samples of a given genotype and the mean was used for comparison between genotypes . For NMJ signal quantifications , mean signal intensity within the ROI encompassing the synaptic area was normalized to HRP signal . To determine Brp intensity per puncta , individual puncta within the ROI were manually counted and total Brp intensity was divided by the number of puncta . Student’s t test was performed using Sigma Plot ( Systat ) to evaluate statistical significance . All graphs represent mean value of all samples of the given genotype ± SEM . Samples were prepared as described for immunohistology and mounted using #1 . 5 cover glasses ( Cat . 12-541-B , Fisher Scientific ) . 5 phases and 3 rotations of 3D SIM images were captured using a Zeiss Elyra microscope . The interval for all z stacks was 110nm . Channels were aligned using parameters obtained from calibration measurements with 100 nm TetraSpeck beads . Zeiss SIM images were taken with a 100X 1 . 46 NA oil objective and a PCO edge sCMOS camera ( 16 bit images ) . Laser power and exposure time were optimized to use a large portion of the camera’s dynamic range while minimizing bleaching . As a part of the reconstruction processing using the Zeiss Zen software , Wiener filtering was carefully optimized to maximize resolution and minimize artifacts . The estimated resolution after reconstruction was ~100 nm lateral and ~250 nm axial . Using the Zeiss Zen software , we generated intensity profiles across structures of interest and exported the table containing the fluorescence intensity as a function of distance . To measure distances we calculated the distances between intensity peaks . Surface rendering was performed using Imaris software . Recordings were performed on muscle 6 , segment A3 of third instar larvae as previously reported [29] . Briefly , wandering third instar larvae were dissected in ice-cold , calcium-free physiological HL-3 saline [98] , and immersed in HL-3 containing Ca2+ before being shifted to the recording chamber . The calcium-free HL-3 saline contains ( in mM ) : 70 NaCl , 5 KCl , 20 MgCl2 , 10 HCO3 , 5 trehalose , 115 sucrose , 5 HEPES , pH adjusted to 7 . 2 at room temperature . The recording solution was HL-3 with either 0 . 4 or 0 . 5 mM CaCl2 as described in the text . Intracellular electrodes ( borosilicate glass capillaries of 1 mm diameter ) were filled with 3 M KCl and resistances ranged from 12 to 25 MΩ . Recordings were done at room temperature from muscle cells with an initial membrane potential between –50 and –70 mV , and input resistances of ≥ 4 MΩ . For mEJCs recording the muscle cells were clamped to –80 mV . To calculate mean amplitudes and frequency of mEJCs or mEJP , 100–150 events from each muscle were measured and averaged using the Mini Analysis program ( Synaptosoft ) . Minis with a slow rise and falling time arising from neighboring electrically coupled muscle cells were excluded from analysis [101 , 102] . To measure the decay time constant of mEJCs , 20–30 clear representative events from each recording were averaged and fit by a single exponential function . For evoked EJP recordings , the nerve roots were cut near the exiting site of the ventral nerve cord so that the motor nerve could be picked up by a suction electrode . Following motor nerve stimulation with a suction electrode ( 100 μs , 5 V ) , evoked EJPs were recorded . Four to six EJPs evoked by low frequency of stimulation ( 0 . 1 Hz ) were averaged . Quantal content was calculated by dividing the mean EJP by the mean mEJP . Since data were recorded in low calcium saline ( 0 . 4 mM Ca2+ ) no correction was made for nonlinear summation . Statistical analysis used KaleidaGraph 4 . 5 ( Synergy Software ) . Electrical signals were recorded with an Axoclamp 2B amplifier ( Axon Instruments ) . The signals were filtered at 1 kHz and digitized at 10 kHz by using an analog-digital converter ( Digidata 1440A ) and pCLAMP software ( version 10 . 0 , Axon Instruments ) . Data are presented as mean ±SEM . One-way ANOVA followed by a Tukey’s post hoc test was used to assess statistically significant differences among genotypes . Differences were considered significant at p < 0 . 05 .
|
Synaptic activity and synapse development are intimately linked , but our understanding of the coupling mechanisms remains limited . Anterograde and retrograde signals together with trans-synaptic complexes enable intercellular communications . How synapse activity status is monitored and relayed across the synaptic cleft remains poorly understood . The Drosophila NMJ is a very powerful genetic system to study synapse development . BMP signaling modulates NMJ growth via a canonical , Smad-dependent pathway , but also synapse stability , via a noncanonical , Smad-independent pathway . Here we describe a novel , noncanonical BMP pathway , which is genetically distinguishable from all other known BMP pathways . This pathway does not contribute to NMJ growth and instead influences synapse formation and maturation in an activity-dependent manner . Specifically , phosphorylated Smad ( pMad in flies ) accumulates at active zone in response to active postsynaptic type-A glutamate receptors , a specific receptor subtype . In turn , synaptic pMad functions to promote the recruitment of type-A receptors at synaptic sites . This positive feedback loop provides a molecular switch controlling which flavor of glutamate receptors will be stabilized at synaptic locations as a function of synapse status . Since BMP signaling also controls NMJ growth and stability , BMP pathway offers an exquisite means to monitor the status of synapse activity and coordinate NMJ growth with synapse maturation and stabilization .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"phosphorylation",
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"nervous",
"system",
"electrophysiology",
"neuroscience",
"animals",
"motor",
"neurons",
"animal",
"models",
"organisms",
"developmental",
"biology",
"drosophila",
"melanogaster",
"model",
"organisms",
"bioassays",
"and",
"physiological",
"analysis",
"molecular",
"biology",
"techniques",
"muscle",
"electrophysiology",
"drosophila",
"research",
"and",
"analysis",
"methods",
"animal",
"cells",
"proteins",
"hyperexpression",
"techniques",
"electrophysiological",
"techniques",
"molecular",
"biology",
"insects",
"molecular",
"biology",
"assays",
"and",
"analysis",
"techniques",
"arthropoda",
"gene",
"expression",
"and",
"vector",
"techniques",
"biochemistry",
"signal",
"transduction",
"larvae",
"cell",
"biology",
"anatomy",
"synapses",
"post-translational",
"modification",
"physiology",
"neurons",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"metamorphosis",
"cell",
"signaling",
"neurophysiology",
"bmp",
"signaling"
] |
2016
|
A Novel, Noncanonical BMP Pathway Modulates Synapse Maturation at the Drosophila Neuromuscular Junction
|
Loops in proteins are flexible regions connecting regular secondary structures . They are often involved in protein functions through interacting with other molecules . The irregularity and flexibility of loops make their structures difficult to determine experimentally and challenging to model computationally . Conformation sampling and energy evaluation are the two key components in loop modeling . We have developed a new method for loop conformation sampling and prediction based on a chain growth sequential Monte Carlo sampling strategy , called Distance-guided Sequential chain-Growth Monte Carlo ( DiSGro ) . With an energy function designed specifically for loops , our method can efficiently generate high quality loop conformations with low energy that are enriched with near-native loop structures . The average minimum global backbone RMSD for 1 , 000 conformations of 12-residue loops is Å , with a lowest energy RMSD of Å , and an average ensemble RMSD of Å . A novel geometric criterion is applied to speed up calculations . The computational cost of generating 1 , 000 conformations for each of the x loops in a benchmark dataset is only about cpu minutes for 12-residue loops , compared to ca cpu minutes using the FALCm method . Test results on benchmark datasets show that DiSGro performs comparably or better than previous successful methods , while requiring far less computing time . DiSGro is especially effective in modeling longer loops ( – residues ) .
Protein loops connect regular secondary structures and are flexible regions on protein surface . They often play important functional roles in recognition and binding of small molecules or other proteins [1]–[3] . The flexibility and irregularity of loops make their structures difficult to resolve experimentally [4] . They are also challenging to model computationally [5] , [6] . Prediction of loop conformations is an important problem and has received considerable attention [5]–[27] . Among existing methods for loop prediction , template-free methods build loop structures de novo through conformational search [5]–[7] , [9] , [10] , [13] , [14] , [17] , [18] , [21] , [23] , [28] . Template-based methods build loops by using loop fragments extracted from known protein structures in the Protein Data Bank [11] , [19] , [27] . Recent advances in template-free loop modeling have enabled prediction of structures of long loops with impressive accuracy when crystal contacts or protein family specific information such as that of GPCR family is taken into account [14] , [23] , [25] . Loop modeling can be considered as a miniaturized protein folding problem . However , several factors make it much more challenging than folding small peptides . First , a loop conformation needs to connect two fixed ends with desired bond lengths and angles [8] , [12] . Generating quality loop conformations satisfying this geometric constraint is nontrivial . Second , the complex interactions between atoms in a loop and those in its surrounding make the energy landscape around near-native loop conformations quite rugged . Water molecules , which are often implicitly modeled in most loop sampling methods , may contribute significantly to the energetics of loops . Hydrogen bonding networks around loops are usually more complex and difficult to model than those in regular secondary structures . Third , since loops are located on the surface of proteins , conformational entropy may also play more prominent roles in the stability of near-native loop conformations [29] , [30] . Approaches based on energy optimization , which ignore backbone and/or side chain conformational entropies , may be biased toward those overly compact non-native structures . Despite extensive studies in the past and significant progress made in recent years , both conformational sampling and energy evaluation remain challenging problems , especially for long loops ( e . g . , ) . In this paper , we propose a novel method for loop sampling , called Distance-guided Sequential chain-Growth Monte Carlo ( DiSGro ) . Based on the principle of chain growth [15] , [31] , [32] , [34] , [35] , the strategy of sampling through sequentially growing protein chains allows efficient exploration of conformational space [15] , [34]–[37] . For example , the Fragment Regrowth via Energy-guided Sequential Sampling ( FRESS ) method outperformed previous methods on folding benchmark HP sequences [15] , [33] . In addition to HP model [15] , sequential chain-growth sampling has been used to study protein packing and void formation [35] , side chain entropy [29] , [38] , near-native protein structure sampling [30] , conformation sampling from contact maps [39] , reconstruction of transition state ensemble of protein folding [40] , RNA loop entropy calculation [37] , and structure prediction of pseudo-knotted RNA molecules [41] . In this study , we first derive empirical distributions of end-to-end distances of loops of different lengths , as well as empirical distributions of backbone dihedral angles of different residue types from a loop database constructed from known protein structures . An empirical distance guidance function is then employed to bias the growth of loop fragments towards the -terminal end of the loop . The backbone dihedral angle distributions are used to sample energetically favorable dihedral angles , which lead to improved exploration of low energy loop conformations . Computational cost is reduced by excluding atoms from energy calculation using REsidue-residue Distance Cutoff and ELLipsoid criterion , called Redcell . Sampled loop conformations , all free of steric clashes , can be scored and ranked efficiently using an atom-based distance-dependent empirical potential function specifically designed for loops . Our paper is organized as follows . We first present results for structure prediction using five different test data sets . We show that DiSGro has significant advantages in generating native-like loops . Accurate loops can be constructed by using DiSGro combined with a specifically designed atom-based distance-dependent empirical potential function . Our method is also computationally more efficient compared to previous methods [8] , [9] , [18] , [22] , [42] . We describe our model and the DISGRO sampling method in detail at the end .
We use five data sets as our test sets . Test Set 1 contains loops at lengths four , eight , and twelve , for a total of loops from PDB structures , which were described in Table 2 of zRef . [8] . Test Set 2 consists of eight , eleven , and twelve-residue loops from Table C1 of Ref . [42] . Several loop structures were removed as they were nine-residue loops but mislabeled as eight-residue loops: ( 1awd , 55–63; 1byb , 246–254; and 1ptf , 10–18 ) . Altogether , there are eight-residue loops . Test Set 3 is a subset of that of [5] , which was used in the RAPPER and FALCm studies [10] , [22] . Details of this set can be found in the “Fiser Benchmark Set” section of Ref . [10] . Test Set 4 is taken from Table A1–A6 of Ref . [42] . Test Set 5 contains fourteen , fifteen , sixteen and seventeen-residue loops from Table 3 of Ref . [23] . Test Set 1 and 2 are used for testing the capability of DiSGro and other methods in generating native-like loops . Test Set 3 , 4 , and 5 are used for assessing the accuracy of predicted loops based on selection from energy evaluation using our atom-based distance-dependent empirical potential function . Our results are reported as global backbone RMSD , calculated using the N , , C and O atoms of the backbone . To evaluate our method for producing native-like loop conformations , we use Test Set 1 and 2 . We generate loops for each of the loop structures in Test Set 1 at length , , and residues , respectively . We compare our results with those obtained by CCD [8] , CSJD [12] , SOS [18] , and FALCm [22] . The minimum RMSD among sampled loops generated by DiSGro are listed in Table 1 , along with results from the four other methods . Accurate loops of longer length are more difficult to generate . For loops with residues , DiSGro generates more accurate loops than other methods . Our method has a mean of Å for the minimum RMSD , compared to Å for FALCm , the next best method in the group [22] . The minimum RMSD of nine of the ten -residue loops have RMSD Å , while five loops of the ten generated by FALCm have RMSD Å . Compared to the CCD , CSJD , and SOS methods , our loops have significantly smaller minimum RMSD ( Å vs , , and Å , respectively , Table 1 ) . The average minimum global backbone RMSD for -residue loops can be further improved when we increase the sample size of generated loop conformations . The minimum global RMSD is improved to Å , Å , and Å when the sample size is increased to 20 , 000 , 100 , 000 , and 1 , 000 , 000 , respectively . Further improvement would likely require flexible bond lengths and angles . For loops with residues , DiSGro has an average minimum RMSD value smaller than the CCD , CSJD , and SOS methods ( Å vs Å , Å , and Å , respectively , Table 1 ) . In eight of the ten 8-residue loops , DiSGro achieves sub-angstrom accuracy ( RMSD Å ) , although the mean of minimum RMSD of -residue loops is slightly larger than that from FALCm ( Å vs Å ) . For loops with -residue , the mean of the minimum RMSD ( Å ) by DiSGro is significantly smaller than those by the CSJD and the CCD methods ( Å and Å , respectively ) , and is similar to those by the SOS and FALCm methods ( Å and Å , respectively ) . Noticeably , three of the ten loops have RMSD Å , indicating our sampling method has good accuracy for short loop modeling . These loops can be generated rapidly . The computing time per conformation averaged over 5 , 000 conformations for , , and -residues is , , and using a single AMD Opteron processor of . In addition to improved average minimum RMSD , DiSGro seems to take less time than CCD ( , , and on an AMD 1800+ MP processor for the , , and -residue loops ) , and is as efficient as SOS ( , , and for the , , and -residue loops on an AMD 1800+ MP processor ) . Reducing the number of trial states in DiSGro can further reduce the computing time , with some trade-off in sampling accuracy . For example , when we take , the computing time per conformation averaged over 5 , 000 conformations for , , and -residues is only , , and , respectively , with the average minimum RMSDs comparable to those from SOS's ( Å vs Å , Å vs Å , and Å vs Å for the , , and -residue loops , respectively ) . Although the CSJD loop closure method has faster computing time ( , , and on AMD 1800+ MP processor ) , the speed of DiSGro is adequate in practical applications . We compare DiSGro in generating near-native loops with Wriggling [43] , Random Tweak [44] , Direct Tweak [42] , [45] , [45] , and PLOP-build [13] using Test Set 2 . The minimum RMSD among loops generated by DiSGro are listed in Table 2 , along with results from the other methods obtained from Table 2 in Ref . [42] . Direct Tweak and from the LoopBuilder method and our DiSGro have better accuracy in sampling than Wriggling , Random Tweak , and PLOP-build methods . For loops with 11 and 12-residues , these three methods are the only ones that can generate near-native loop structures with minimal RMSD values below Å . Among these , DiSGro outperforms in generating loops at all three lengths: the average minimal RMSD ( ) is Å vs . Å for length , Å vs . Å for length , and Å vs . Å for length , respectively . Compared to the Direct Tweak sampling method , DiSGro has improved for -residue loops ( Å vs Å ) , slightly improved for -residue loops ( Å vs Å ) and inferior for -residue loops ( Å vs Å ) . Overall , these results show that DiSGro are very effective in sampling near-native loop conformations , especially when modeling longer loops of length 11 and 12 . Our DiSGro method can generate accurate loops and has significant advantages for longer loops compared to previous methods . Using RMSD values calculated from three backbone atoms N , , and C for all loop lengths lead to the same conclusion . To assess the accuracy of loops selected by our specifically designed atom-based distance-dependent empirical potential function , we test DiSGro using Test Set 3 and follow the approach of reference [22] for ease of comparison . Because of the high content of secondary structures , these loops are very challenging to model . In the study of [22] , backbone conformations with the best scores evaluated by DFIRE potential function [46] were retained after screening generated backbone conformations for each loop . Loop closure and steric clash removal were not enforced to the conformations . We follow the same procedure , except the DFIRE potential function is replaced by our atom-based distance-dependent empirical potential function . The ensemble of the selected backbone conformations are then subjected to the procedure of side-chain construction as described in the Section “Side-chain modeling and steric clash removal” . The loop conformations with full side-chains are then scored and ranked by the atom-based distance-dependent empirical potential function . Our results are summarized in Table 3 . We measure the average minimum backbone RMSD , the average ensemble RMSD , and the average RMSD of the lowest energy conformations of the 1 , 000 loop ensemble with the same length . Overall , DiSGro performs significantly better than FALCm and RAPPER in , and for all loop lengths . Compared to FALCm , DiSGro shows significant advantages in on sampling long loops of – residues . Our method has of Å compared to Å for -residue loops , Å compared to Å for -residue loops , and Å compared to Å for -residue loops , respectively . For example , as can be seen in Figure 1 , the lowest energy loop ( red ) of a 12-residue loop in the protein 1scs ( residues 199–210 ) has a Å RMSD to the native structure ( white ) . The generated top five lowest energy loops are all very close to the native loop , yet are diverse among themselves . DiSGro also generates loops with smaller compared to FALCm in loops with length ranging from to , indicating DiSGro can generate ensemble of loop conformations with enriched near native conformations . Furthermore DiSGro achieves better modeling accuracy using the atom-based distance-dependent empirical potential function . Compared to FALCm , DiSGro has a of Å vs Å for -residue loops , Å vs Å for -residue loops , Å vs Å for -residue loops , Å vs Å for -residue loops , and Å vs Å for -residue loops , respectively . DiSGro is also much faster than other methods . The reported typical computational cost of FALCm is cpu minutes for – residue loops on a Linux server of a 2-core Intel Xeon processor [47] . The computation cost for DiSGro method is only and 10 cpu minutes for and 12–residue loops on a single AMD Opteron processor , respectively . In addition , FALCm has a size restriction , and it only works with proteins with residues . In contrast , the overall protein size has no effect on the computational efficiency of DiSGro since the numbers of atoms for energy calculation that are retained by the ellipsoid criterion are bounded . The LOOPER method is an accurate and efficient loop modeling method using a minimal conformational sampling method combined with energy minimization [17] . The test set used in the LOOPER study is the original Fiser data set without removal of any loops . Therefore , it is different from Test Set 3 used in the RAPPER and FALCm studies [10] , [22] . For ease of comparison , we compare DiSGro to the LOOPER using the test set with –-residue loops from [17] . Our results are summarized in Table 4 . We denote and as the mean and median of backbone RMSD of the lowest energy conformations with the same loop length . Similarly , we use , and to denote the mean and median RMSD values of all-heavy atoms . DiSGro shows improved prediction accuracy compared to LOOPER in both backbone and all-heavy atom RMSD . For the loops of length 12 , is Å compared to Å , while the median is Å compared to Å . It also has better all-heavy atom RMSD of Å/ Å ( mean/median ) , compared to Å/ Å for -residue loops , Å/ Å compared to Å/ Å for -residue loops , and Å/ Å compared to Å/ Å for -residue loops . It is worth noting that DiSGro outperforms LOOPER in speed as well . For a loop with residues , the time cost of DiSGro is minutes using a CPU versus cpu minutes using a processor according to Figure 7 in the LOOPER paper [17] . Prior publications also allowed us to compare results in loop structure predictions based on energy discrimination using Test Set 4 with results obtained using the LoopBuilder method [42] . Following [42] , we generated closed loop conformations for eight-residue loops , for nine-residue loops , for ten , eleven , and twelve-residue loops , and for thirteen-residue loops . Energy calculations are carried out using our atom-based distance-dependent empirical potential function . The average RMSD of the lowest energy conformations , , are then compared between these two methods . The results are summarized in Table 5 . Compared to LoopBuilder , DiSGro has better : Å vs Å for -residue loops , Å vs Å for -residue loops , Å vs Å for -residue loops , Å vs Å for -residue loops , and Å vs Å for -residue loops , respectively . DiSGro has inferior performance in selecting for -residue loops ( Å vs Å ) . The average time using LoopBuilder for twelve-residue loops was around 4 . 5 hours or 270 minutes , while the computational time using DiSGro is around 10 minutes . Overall , DiSGro has equal or slightly better performance than LoopBuilder in average prediction accuracy of loop structures with far less computing time . To test the feasibility of DiSGro in modeling longer loops with length , we use the Fiser -residue loops data set to generate and select low energy loop conformations . conformations with low energy are obtained . The mean of minimum backbone RMSD of loops with -residue is Å , and the median is Å . The mean/median of the backbone RMSD , and all heavy atom RMSD of the lowest energy conformations are Å/ Å and Å/ Å , respectively ( Table 6 ) . With extensive conformational sampling using molecular mechanics force field , the Protein Local Optimization Program ( PLOP ) can predict highly accurate loops [13] , [14] , [23] . We tested DiSGro using Test Set 5 consisting of loops with length – and compared results with those using PLOP . Here the sampling and scoring processes were similar to those used in Test Set 3 , except 100 , 000 backbone conformations were generated . We measured the average minimum backbone RMSD and the average RMSD of the lowest energy conformations . Our results are summarized in Table 7 . Loops predicted by the PLOP method have smaller compared to DiSGro [23] , although DiSGro samples well and gives small of Å for -residue loops , Å for -residue loops , Å for -residue loops , and Å for -residue loops . For loops of length , the of Å is less than the reported Å using PLOP , although it is unclear whether the of loops generated by PLOP is less than Å . Overall , DiSGro is capable of successfully generating high quality near-native long loops , up to length 17 . The accuracy of of loops generated by DiSGro may be further improved by using a more effective scoring function . We also compared the computational costs of the two methods . The average computing time for DiSGro is , , , and hours for loops of lengths , , , and using a single core AMD Opteron processor , respectively , which is more than two orders of magnitude less than the time required for the PLOP method ( , , , and hours for loops of length , , , and residues , respectively ) . We used a REsidue-residue Distance Cutoff and ELLipsoid criterion ( Redcell ) to improve the computational efficiency . To assess the effectiveness of this approach , we carry out a test using a set of 140 proteins ( see discussion of the tuning set in Materials and Methods ) . We compared the time cost of energy calculation of generating a single loop , with and without this procedure . When the procedure is applied , we only calculate the pairwise atom-atom distance energy between atoms in loop residues and other atoms within the ellipsoid . When the procedure is not applied , we calculate energy function between atoms in loop residues and all other atoms in the rest of the protein . The computational cost of energy calculations for sampling single loops with and -residues are shown in Figure 2A and Figure 2B , respectively . From Figure 1 , we can see that significant improvement in computational cost is achieved . The average time cost using our procedure is reduced from to for sampling -residue loops , and to for -residue loops . In addition , this approach makes the time cost of energy calculations independent of the protein size ( Figure 2A and Figure 2B ) , whereas the computing time without applying this procedure increases linearly with the protein size . The improvement is especially significant for large proteins . For example , to generate a -residue loop in a protein with residues , the computing time is improved from to , which is more than -fold speed-up . Detailed examination indicates that both distance cutoff and the ellipsoid criterion contribute to the computational efficiency . Furthermore , the full Redcell procedure has improved efficiency over using either “Ellipsoid Criterion Only” or “Cutoff Criterion Only” . The computing time for generating a -residue loops is when the full Redcell procedure is applied , compared to , and , when only the ellipsoid criterion and only the distance-threshold are used , respectively ( Figure 2C ) . Furthermore , there is no loss of accuracy in energy evaluation . Overall , Redcell improves the computational cost by excluding many atoms from collision detections and energy calculations , with significant reduction in computation time , especially for large proteins .
In this study , we presented a novel method Distance-guided Sequential chain-Growth Monte Carlo ( DiSGro ) for generating protein loop conformations and predicting loop structures . Ensembles of near-native loop conformations can be efficiently generated using the DiSGro method . DiSGro has better average minimum backbone RMSD , , compared to other loop sampling methods . For example , is Å for 12-residue loops when using DiSGro , while the corresponding values are Å , Å , Å , and Å when using the CCD , CSJD , SOS , and the FALCm method . DiSGro also performs well in identifying native-like conformations using atom-based distance-dependent empirical potential function . In comparison with other similar loop modeling methods , DiSGro demonstrated improved modeling accuracy , in terms of an average RMSD of the lowest energy conformations for the more challenging task of sampling longer loops of – residues . For example , DiSGro outperforms FALCm [22] ( Å vs Å ) and LOOPER [17] ( Å vs Å ) in predicting -residue loops , while taking less computing time ( minutes vs minutes for FALCm and minutes for LOOPER . Compared to LoopBuilder [42] , DiSGro also has better : For -residue loops , the is Å using DiSGro , but is Å when using the Loop Builder . The average computing time is also faster when using DiSGro: it takes about minutes to predict structures of -residue loops and minutes for -residue loops . DiSGro also works well for short loops , although this may be largely a reflection of the underlying analytical closure method [12] . There are a number of directions for further improvement . DiSGro can be further improved by adding fragments of peptides when growing loops instead of adding individual residues . Fragment-based approach has been widely used in protein structure prediction [48]–[51] and specifically in loop structure prediction [21] . It is straightforward to apply the strategy described in this study for fragment-based growth , and it will likely lead to improved sampling efficiency further and enable longer loops to be modeled . Furthermore , the energy function employed here can be further improved by optimization such as those obtained by training with challenging decoy loops using nonlinear kernel [52] , and/or using rapid iterations through a physical convergence function [53] , [54] . In addition , DiSGro is compatible with different loop closure methods [8] , [12] , [22] , and experimenting with other closure strategy may also lead to further improvement . An efficient loop sampling method such as DiSGro can help to improve overall modeling of loop structures . Currently , the hierarchical approach of the Protein Local Optimization Program ( PLOP ) [13] , [14] , [23] gives excellent accuracy in protein loop modeling , but requires significant computational time . The average time cost of modeling a -residue loop is about 4–5 days [23] . Kinematic closure ( KIC ) method can also make very accurate predictions of -residue loops [21] . However , KIC also requires substantial computation , with about CPU hours on a single Opteron processor for predicting 12-residue loops [21] . As suggested earlier by Spassov et al [17] , an efficient loop modeling method combined with energy minimization may overcome the obstacle of high computational cost . By generating high quality initial structures using DiSGro , near native conformations of loops can be used as candidates for further refinement .
All heavy atoms in the backbone and side chain of a protein loop are explicitly modeled . The bond lengths and angles are taken from standard values specific to residue and atom type [55] . The backbone dihedral angles and side chain dihedral angles constitute all the degrees of freedom ( DOFs ) in our model . In order to efficiently generate adequate number of native-like loop conformations , we have developed a Distance-guided Sequential chain-Growth Monte Carlo ( DiSGro ) method . Let the loop to be modeled begins at residue and ends at residue . The sequence of the positions of backbone heavy atoms from atom of residue to ( ) atom of residue are unknown and need to be generated . We assume that the backbone atoms before and after this fragment are known . Coordinates of side chain atoms are also unknown and need to be generated if the coordinates of the atoms they are attached to are unknown . At each step of the chain growth process , we generate three consecutive backbone atoms continuing from the backbone atom sampled at the previous step . At the -th growth step ( ) , the three backbone atoms are atom of residue , atom of residue , and atom of residue ( Figure 3 ) . The coordinates of the three atoms , , and , are denoted as , , and , respectively . The dihedral angles that determine the coordinate of atoms are sampled from a normal distribution with mean and standard deviation . In the next section , we describe in detail in sampling of the dihedral angles , which determine the coordinates of the and the atoms . To reduce computational cost of calculating atom-atom distances in energy evaluation , we use a procedure , REsidue-residue Distance Cutoff and ELLipsoid criterion ( Redcell ) to reduce computational time . Side chains are built upon completion of backbone sampling of a loop . For the -th residue of type , we denote the degrees of freedom ( DOFs ) for its side chain as . DOFs of side chain residues depend on the residue types , e . g . Arg has four dihedral angles ( ) , with ( ) . Val only has one dihedral angle ( ) , with ( ) . Each DOFs is discretized into bins of , and only bins with non-zero entries for all loop residues in the loop database are retained . We sample trial states of side chains from the empirical distribution obtained from the loop database . One of trials is then chosen according to the probability calculated by the empirical potential . Denote the side chain fragment for the -th residue as , we select following the probability distribution:where is the interaction energy of the newly added side chain fragment with the remaining part of the protein , and is the effective temperature . When there are steric clashes between side chains , we rotate the side-chain atoms along the axis for all residue types except Pro . For Pro , we use the axis for rotation . We consider two atoms to be in steric clash if the ratio of their distance to the sum of their van der Waals radii is less than [13] . To evaluate the energy of loops , we develop a simple atom-based distance-dependent empirical potential function , following well-established practices [46] , [52] , [60]–[66] . Empirical energy functions developed from databases have been shown to be very effective in protein structure prediction , decoy discrimination , and protein-ligand interactions [54] , [63] , [64] , [67]–[71] . As our interest is modeling the loop regions , the atomic distance-dependent empirical potential is built from loop structures collected in the PDB [72] . Instead of using detailed atom types associated with the amino acids , we group all heavy atoms into groups , similar to the approach used in Rosetta [50] . The side-chain atom types comprise six carbon types , six nitrogen types , three oxygen types , and one sulfur type . The backbone types are N , , C , and O . This simplified scheme helps to alleviate the problem of sparsity of observed data for certain parameter values . For an atom in the loop region of atom type and an atom of atom type , regardless whether is in the loop region , the distance-dependent interaction energy is calculated as : ( 10 ) where denotes the interaction energy between a specific atom pair at distance , and are the observed probability of this distance-dependent interaction from the loop database and the expected probability from a random model , respectively . The observed probability is calculated as: ( 11 ) where is the observed count of pairs found in the loop structures with the distance falling in the predefined bins . We use a total of bins for , ranging from Å to Å , with the bin width set to Å . ranging from Å to Å is treated as one bin . Here , where is the number of loops in our loop database , is the observed number of pairs at the distance of in the -th loop . is the observed total number of all atom pairs in the loop database regardless of the atom types and distance , namely , . The expected random distance-dependent probability of this pair is calculated based on sampled loop conformations , called decoys . It is calculated as: ( 12 ) where is the expected number of ( ) pairs averaged over all decoy loop conformations of all target loops in the loop database . Here is the number of pairs at distance in the -th generated loop conformations for the -th loop . is the number of decoys generated for a loop , which is set to . is the number of loops in our loop database . is the total number of all atom pairs in the reference state , . We have made the source code of DiSGro available for download . The URL is at: tanto . bioengr . uic . edu/DiSGro/ .
|
Loops in proteins are flexible regions connecting regular secondary structures . They are often involved in protein functions through interacting with other molecules . The irregularity and flexibility of loops make their structures difficult to determine experimentally and challenging to model computationally . Despite significant progress made in the past in loop modeling , current methods still cannot generate near-native loop conformations rapidly . In this study , we develop a fast chain-growth method for loop modeling , called Distance-guided Sequential chain-Growth Monte Carlo ( DiSGro ) , to efficiently generate high quality near-native loop conformations . The generated loops can be used directly for downstream applications or as candidates for further refinement .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biochemistry",
"proteins",
"protein",
"folding",
"protein",
"structure",
"biology",
"and",
"life",
"sciences",
"computational",
"biology",
"biophysics"
] |
2014
|
Fast Protein Loop Sampling and Structure Prediction Using Distance-Guided Sequential Chain-Growth Monte Carlo Method
|
Viruses impose diverse and dynamic challenges on host defenses . Diversifying selection of codons and gene copy number variation are two hallmarks of genetic innovation in antiviral genes engaged in host-virus genetic conflicts . The myxovirus resistance ( Mx ) genes encode interferon-inducible GTPases that constitute a major arm of the cell-autonomous defense against viral infection . Unlike the broad antiviral activity of MxA , primate MxB was recently shown to specifically inhibit lentiviruses including HIV-1 . We carried out detailed evolutionary analyses to investigate whether genetic conflict with lentiviruses has shaped MxB evolution in primates . We found strong evidence for diversifying selection in the MxB N-terminal tail , which contains molecular determinants of MxB anti-lentivirus specificity . However , we found no overlap between previously-mapped residues that dictate lentiviral restriction and those that have evolved under diversifying selection . Instead , our findings are consistent with MxB having a long-standing and important role in the interferon response to viral infection against a broader range of pathogens than is currently appreciated . Despite its critical role in host innate immunity , we also uncovered multiple functional losses of MxB during mammalian evolution , either by pseudogenization or by gene conversion from MxA genes . Thus , although the majority of mammalian genomes encode two Mx genes , this apparent stasis masks the dramatic effects that recombination and diversifying selection have played in shaping the evolutionary history of Mx genes . Discrepancies between our study and previous publications highlight the need to account for recombination in analyses of positive selection , as well as the importance of using sequence datasets with appropriate depth of divergence . Our study also illustrates that evolutionary analyses of antiviral gene families are critical towards understanding molecular principles that govern host-virus interactions and species-specific susceptibility to viral infection .
Ancient , pathogenic viruses have played a major role in shaping the extant host innate immune repertoire . Understanding how pathogen-driven evolution has shaped host-virus interfaces can reveal insights into the molecular basis of cross-species transmission , including human susceptibility to zoonoses [1] . The genetic signature of diversifying ( positive ) selection distinguishes many host antiviral genes , indicating their involvement in a long-standing genetic conflict with viral pathogens . Such genetic conflict has also driven gene copy number expansion in several mammalian antiviral genes , allowing further diversification of pathogen defense . For instance , the TRIM5 antiviral gene is present in one copy in primates but has expanded to 6–7 copies in mice and other mammals [2 , 3] . Similarly , primates encode seven members of the APOBEC3 antiviral gene family whereas mouse genomes encode only one [4 , 5] . Unlike the TRIM5 , APOBEC3 or other antiviral gene families , the copy number of myxovirus resistance ( Mx ) gene appears to be relatively static , with two copies in both primate and mouse genomes . Mx proteins are interferon-inducible dynamin-like large GTPases . They are composed of a highly conserved GTPase domain ( GD ) , which is connected to a helical stalk by a hinge-like bundle-signaling element ( BSE ) [6] . Previous work has shown that human MxA and both murine Mx1 and Mx2 proteins have broad and potent activity against a diverse range of RNA and DNA viruses [7 , 8] . In contrast , the antiviral activity of human MxB appears to be much more narrow , only recently having been shown to restrict HIV-1 and other primate lentiviruses [9–12] . In this study we employed a detailed evolutionary approach to address the basis for the apparent stasis of Mx gene copy number and the discrepancy in antiviral breadth of Mx homologs . Using maximum-likelihood approaches , we found strong evidence of diversifying selection in the N-terminal region of primate MxB genes in contrast to the previously observed diversifying selection in loop L4 of MxA [13] . Surprisingly , signatures of MxB diversifying selection do not overlap with previously-mapped lentiviral-restriction determinants . We therefore conclude that simian lentiviruses have not driven the rapid evolution of primate MxB . Our analysis instead suggests that MxB plays a central and conserved role in the interferon response to a broader range of pathogens than is currently appreciated . Extending our analysis to other mammalian Mx genes , we find that multiple , lineage-specific exchanges have occurred between Mx paralogs throughout mammalian evolution . These gene conversion events have led to both the preservation of key enzymatic and structural features of Mx GTPases , as well as the acquisition of new antiviral specificity via the complete conversion of MxB-like genes to a MxA-like state . Our findings highlight the impact of diversifying selection and gene conversion on the functional repertoires of antiviral gene families .
We wished to investigate whether the primate MxB gene has been subject to pathogen-driven diversifying selection . A previous analysis reported that the primate MxB gene had not evolved under diversifying selection , even though it did find evidence for positive selection for some individual sites ( see below ) [14] . In contrast , our previous analysis of primate MxA found strong evidence of positive selection at both the gene and codon level [13] . Although this variance could reflect genuine differences in selective pressures that have acted on the two paralogs , we also considered the possibility that lower sampling of MxB sequences in the previous report may have led to reduced power to detect selection [15] . We , therefore , cloned and sequenced MxB from 21 hominoid , Old World monkey and New World monkey species for a total of 32 MxB sequences after including sequences from public databases . Maximum likelihood tests were implemented using the PAML [16] and HyPhy [17] suite of programs to detect whether rates of non-synonymous changes ( dN ) exceeded synonymous changes ( dS ) ( dN/dS > 1 implies positive selection ) . Recombination can yield false signatures of positive selection [18] . We used a genetic algorithm for recombination detection ( GARD ) [19 , 20] to show that Mx genes did indeed undergo recombination during their evolutionary history . We therefore carried out selection analyses only on MxB gene segments for which evolutionary history was determined to be uniform by GARD ( Fig 1A ) . For all analyzed GARD segments , we found strong evidence for positive selection in primate MxB ( M7 vs . M8 , P < 0 . 001 ) ( Fig 1A ) . Our findings support a role for MxB in a long-standing and recurrent host-virus conflict during primate evolution . Our analyses also revealed six codons in primate MxB that showed strong evidence of diversifying selection ( model M8 , Bayes Empirical Bayes ( BEB ) posterior probability ( PP ) >0 . 95 ) ( Fig 1A and 1B ) . Four of these six rapidly evolving residues are located in the disordered N-terminus of MxB ( amino acids 1–83 ) ( Fig 1B ) . The MxB N-terminus determines both its localization to the nuclear pore as well as its antiviral specificity against simian lentiviruses [14 , 21–24] . The positive selection in MxB contrasts with the positive selection in MxA , which is concentrated in the loop L4 ( S1B Fig ) [13] . Although our previous analysis also found evidence of positive selection in the MxA N-terminus , we identified no sites in the loop L4 of MxB in primates that had a high posterior probability of having evolved under diversifying selection ( Fig 1 , S1A Fig ) . Our findings appear to be at odds with a recent analysis of MxB evolution in mammals , which concluded that positive selection is centered on the MxB L4 [25] . However , this previous analysis was based on a broad range of mammals and did not account for the possibility of MxA-MxB recombination , which might have confounded the analysis ( see below ) . Despite the high similarity of MxA and MxB GTPases [26] , our finding that diversifying selection is centered on distinct surfaces ( L4 versus N-terminus ) , together with their divergent cellular localization ( MxA is cytoplasmic whereas MxB is nuclear ) [22 , 27 , 28] , suggests that different pathogens have uniquely shaped MxA and MxB antiviral surfaces during primate evolution . The N-terminus of MxB has been shown to be essential for its antiviral activity against lentiviruses [14 , 21–24 , 29] . We considered whether simian lentiviruses could be responsible for driving positive selection in primate MxB . If so , we would expect that the amino acids that govern anti-lentiviral specificity of MxB would also be the amino acids that are under diversifying selection as has been previously observed for APOBEC3G and TRIM5 [30–33] . We therefore compared the residues that evolved under diversifying selection with two regions in the MxB N-terminus previously identified as molecular determinants of MxB anti-lentivirus activity ( Fig 1C ) , a triple-arginine motif ( RRR11-13 ) , and residues 37–39 . To our surprise , we found no overlap between either of these molecular determinants and positively selected sites . Mutation of the RRR11-13 motif abolishes MxB anti-lentiviral activity [21] . Moreover , anti-HIV-1 activity can be conferred on the highly diverged canine MxB by restoring the RRR11-13 motif [21] . We find that the triple arginine motif arose in the common ancestor of simian and prosimian primates and has since been strictly conserved , except for in New World monkeys in which it has degenerated multiple times ( Fig 1C ) . Residues 37–39 are known to dictate MxB's differential activity against different lentiviruses . Specifically , the group O HIV-1 chimeric CMO2 . 5 strain is sensitive to human but not African green monkey ( AGM ) MxB . Similarly , the HIV-1 P207S capsid mutant is sensitive to rhesus macaque but not AGM MxB . Functional differences between AGM and rhesus macaque MxB map to N-terminal residues 37–39 , especially residue 37 [14] . We find no evidence that recurrent diversifying selection has acted on either the RRR motif or residues 37–39 ( Fig 1 ) . Formally , target recognition of the lentiviral capsid may span a broader region of the MxB N-terminus although alanine-scanning mutagenesis of these sites did not appear to disrupt anti-HIV-1 activity [21] . Given that Old World monkeys are the primary primate lineage infected by lentiviruses , we also restricted our analysis to Old World monkey sequences; analysis of only these species might therefore be expected to better reveal lentiviral-driven selection . Again , we found positive selection in MxB but not in residues shown to confer lentiviral specificity ( Fig 1 ) . Likewise , if we exclude Old World monkeys and analyze only hominoids and New World monkeys , we still observe positive selection ( Fig 1 ) . Analysis of New World monkeys alone is also weakly suggestive of positive selection in the GARD 2 segment ( Fig 1 ) , but our analysis ( only 7 NWM species ) lacks the statistical power for confident interpretation of this result; a denser sampling of this clade is needed . Thus , despite the occurrence of intense diversifying selection in the MxB N-terminus , our results strongly imply that the only known antiviral activity of MxB , i . e . , towards primate lentiviruses , cannot explain the rapid evolution of primate MxB . Our analysis of positive selection in MxB is discrepant with two previous findings , which reported that ( A ) MxB has not evolved under positive selection in primates [14] , and that ( B ) the sites under positive selection in mammalian MxB are predominately located in loop L4 [25] . Here , we explore the sources of these discrepancies , which may help inform future analyses of antiviral genes . Busnadiego et al . [14] performed three NSsites tests to detect positive selection in primate MxB , one of which was significant . However , the single significant analysis ( model 0 vs . 3 ) tests variability in the ω ratio among sites and does not constitute a test of positive selection [34] . In contrast , both analyses that are designed to detect positive selection ( i . e . , model 1 vs . 2 , or 7 vs . 8 ) were not significant . However , it should be noted that Busnadiego et al . based their analysis on 12 MxB sequences , of which 3 were macaque species and 5 were hominoids , making this dataset relatively shallow . Previous work has demonstrated that species representation strongly influences the robustness and inference of positive selection analyses [15] . In addition , although ten sites were reported to be under positive selection by both REL ( posterior probability > 0 . 9 ) and M3 ( probability > 0 . 99 ) ; ( the former is a valid test of positive selection ) , false positives may arise from the analysis of shallow datasets [35] . Thus , the difference in our ability to detect positive selection in primate MxB is the result of the increased number and diversity of MxB sequences in our analysis wherein we included 32 primate species dispersed throughout the phylogenetic tree . Sironi et al . [25] also identified positive selection in MxB based on an analysis of 29 eutherian mammals . However , while our analysis identifies the MxB N-terminus as the “hotspot” of diversifying selection , they identified the MxB L4 ( Fig 1 ) . It is formally possible that primates have experienced a unique evolutionary history relative to other eutherian mammal lineages , which might explain the incongruence between our results . However , the conclusion that MxB L4 has been a target of selection in mammals should be tempered for two reasons: i ) Large divergences , such as those found across 29 mammals , are prone to dS saturation leading to the underestimation of dS ( i . e . , false positives ) [34] . ii ) The failure to account for recombination can also lead to false positives [18] . Recombination was not detected in their dataset despite our finding of recurrent recombination in a similarly sampled analysis ( see below ) . Although we do not know the exact set of sequences assayed in Sironi et al . , we note that 7 of 31 species listed in their S1 Table encode pseudogenized or gene converted copies of MxB . Our analyses highlight the merit of dense sampling within mammalian orders compared to broad sampling across orders , which is better suited to identify conserved features [36] . Many antiviral gene families have undergone dynamic , lineage-specific changes in copy number , presumably as a mechanism for gaining new antiviral specificities without losing existing functions [37] . In contrast to other antiviral gene families , previous studies have found that Mx gene copy number is relatively static [25 , 38] . However , this apparent stasis may be misleading . For instance , a recent report found that Mx genes have been lost in Odontoceti cetaceans ( toothed whales ) [39] . To more comprehensively determine the evolutionary dynamics of Mx genes in mammals , we performed phylogenetic analyses of mammalian Mx paralogs from at least one representative of all sequenced mammalian orders . We found shared synteny of the Mx locus throughout terrestrial vertebrates ( Fig 2A ) . Both human and mouse genomes encode two Mx genes , but as previously described [38] , rodents have lost the MxB-like gene and instead encode two MxA orthologs ( Mx1 and Mx2 ) . Platypus genomes encode two Mx genes that both appear ancestral to the eutherian MxA and MxB lineages; we term them here MxAB1 and MxAB2 ( Fig 2 ) . We were unable to identify any Mx genes in any of three marsupial genome sequences ( tammar wallaby , opossum , Tasmanian devil ) and suggest that Mx gene ( s ) were lost from the common marsupial ancestor . We therefore infer that MxA and MxB genes diverged at or just prior to the origin of the eutherian mammal lineage . Within eutherian mammals , our phylogenetic analyses revealed surprisingly poor resolution; many nodes have less than 50% bootstrap support and some discordance at well-supported nodes of the mammalian phylogeny ( Fig 2B ) [40] . These discrepancies are not entirely unexpected for rapidly evolving antiviral genes , and likely reflect complex evolutionary histories of Mx genes in multiple mammalian lineages ( see below ) . Nevertheless , we were able to conclude that most eutherian mammalian genomes encode both MxA-like ( orthologous to human MxA ) and MxB-like ( orthologous to human MxB ) genes . Further , elephants encode two closely related intact MxA genes in addition to MxB; this MxA duplication appears to have occurred since divergence from the sister order Macroscelidea ( e . g . , elephant shrew ) ( Fig 2 ) . With the exception of the loss of both MxA and MxB genes in toothed whales [39] , we found evidence for an intact MxA gene in all surveyed eutherian mammal species . In contrast , we found that MxB has been lost at least three additional , independent times in Rodentia , Felidae and Xenarthra ( Fig 2 and S2 Fig ) . Therefore , MxB loss has been tolerated on multiple occasions during eutherian mammal evolution despite the fact that its importance as an antiviral factor has been experimentally demonstrated . Phylogenetic and synteny analyses allowed us to propose a hypothetical scenario for the loss of the MxB gene in mouse . We found that the ancestral configuration of the mammalian Mx locus was Bace2 ( + ) ;Fam3B ( + ) ;MxB ( + ) ;MxA ( + ) ;Tmprss2 ( - ) ( Fig 2A ) . In the rabbit genome ( Lagomorpha , an outgroup to Rodentia ) , an inversion occurred in the Mx locus such that MxB and Fam3B have opposite orientations relative to the ancestral locus ( Fig 2A ) . Distinct rearrangements appear to have taken place in Rodentia , represented by squirrel and mouse genomes , leading to a Bace2 ( + ) ;Mx1 ( - ) ;Fam3B ( - ) ;Mx2 ( + ) ;Tmprss2 ( - ) configuration ( Fig 2A ) . Intriguingly , squirrel Mx2 is an MxB-derived pseudogene , whereas mouse Mx2 is an MxA-derived intact gene ( S2 Fig ) . Based on this , we propose that mouse Mx2 originated as a result of complete gene conversion by Mx1 ( MxA-like ) ( Fig 2A ) , possibly preceded by loss of the ancestral MxB-like gene in rodents . We further investigated gene conversion of rodent Mx genes using GARD [19 , 20] to identify putative recombination breakpoints in a multiple sequence alignment . We found that recurrent gene conversion has occurred between Mx genes throughout rodent evolution such that various Mx gene segments have distinct , incongruous evolutionary histories ( Fig 3A and S3 Fig ) . We used PHYML to generate bootstrapped phylogenies of each segment identified by GARD; this analysis confirmed that different segments have different phylogenies , with key nodes indicative of recombination supported by strong bootstrap values . For example , a phylogeny based on GARD segment A indicates that for the majority of rodent species the Mx1 and Mx2 genes are more closely related to each other than to the orthologous gene in related species ( Fig 3B , white circles ) . We also use mVISTA [41] to determine that gene conversion tracts can extend into intronic sequences ( S3 Fig ) . The phylogenetic grouping of mouse-like rodents also suggests that more ancestral conversion events have occurred . We estimate that at least eight independent gene conversion events have occurred in this region alone between Mx1 and Mx2 genes in the eight surveyed rodent species ( indicated at specific nodes in Fig 3B ) . However , given that our sampling is limited to sequenced genomes , this is likely an underestimate of the degree of gene conversion/recombination during rodent Mx evolution ( S3 Fig ) . Thus , a high frequency of gene conversion events has scrambled the phylogenetic relationships between rodent Mx genes . We next extended our survey of possible gene conversion between Mx genes beyond rodents to other mammalian genomes . We found at least two Mx gene segments have distinct evolutionary histories among eutherian mammalian Mx paralogs ( Fig 4 ) . In contrast to the nearly complete gene conversion/recombination between rodent Mx paralogs ( Fig 3 and S3 Fig ) , in other mammals recombination is more localized ( S4 Fig ) . For instance , the regions that coincide with the second coding exon of human MxA and MxB are remarkably similar to each other ( S4 Fig ) . These exons correspond to the Mx bundle-signaling-element BSE α1B ( amino acids 84–116 in human MxA ) and GTPase domain α1G ( amino acids 117–147 ) ( S1 and S4 Figs ) . α1G contains the highly conserved P-loop ( G1 ) , which is an essential GTP-binding element that interacts with the α- and β-phosphates of bound nucleotide [42] . Similarly , we also found that within primate , carnivore and ungulate mammals , the BSE-GTPase domain-encoding segments ( GARD segment A , Fig 4 ) from MxA and MxB cluster with each other , instead of by species; independent phylogenetic analyses support GARD's finding of recombination with high bootstrap support . Based on these observations , we conclude that there has been recurrent gene conversion/recombination between the BSE-GTPase domain-encoding exons of MxA and MxB genes that occurred early after the separation of the different mammalian orders ( Fig 4B , black circles ) . As in rodents , there have been additional recent exchanges in some mammalian lineages ( Fig 4B , white circles ) . We note that this frequent intergenic recombination is the likely cause of the apparent poor resolution of the Mx gene phylogeny in mammals ( Fig 2B ) . Therefore , gene conversion of the BSE-GTPase domains has contributed to the genetic exchange of critical functional elements between the two Mx genes over different evolutionary timeframes .
Molecular arms races between simian lentiviruses and their infected primate hosts have been ongoing for at least 5 million years [31 , 32] . Evidence for lentivirus-driven evolution has been identified in other restriction factors [32] indicating that ancient , simian lentiviruses have imposed a dominant selective pressure on primate antiviral genes . Despite the coincidence of rapid evolution being centered on the MxB N-terminal tail , we find no evidence of diversifying selection having acted on either the RRR11-13 lentiviral-restrictive motif or residue 37 , the two known molecular determinants of MxB anti-lentivirus activity . Previously , comprehensive triple-alanine-scanning mutagenesis of MxB’s N-terminal domain ( amino acids 1–91 ) effectively ruled out any other determinants of antiviral activity against HIV-1 [21] . We therefore fail to find evidence that primate lentiviruses have driven diversifying selection in primate MxB . However , since only a subset of primate lentiviruses have been directly tested for MxB restriction , it is possible that some of the diversifying selection we have mapped might correspond to either primate lentiviruses that have not been tested , or to ancient lentiviruses that no longer exist . On the other hand , since MxB anti-lentivirus activity does not require GTP binding or hydrolysis [9 , 10] , if lentiviruses were the driving target of primate MxB we might have expected MxB GTPase motifs to have degenerated . Contrary to this expectation , we find that all GTP binding motifs are strictly conserved in all surveyed mammals that encode an intact MxB gene ( S1C Fig ) . Preservation of the MxB GTPase domain may either imply a house-keeping function as previously suggested [43] , or an antiviral requirement for GTPase activity against other ( non-lentiviral ) pathogens , as is seen with MxA [44 , 45] . The recurrent loss of MxB in various mammalian lineages suggests that an important housekeeping role for MxB is either unlikely or would have to be highly lineage-specific . Despite the ability of MxB to inhibit primate lentiviruses , its evolution is inconsistent with a recurrent “arms race” scenario with that group of pathogens . The absence of overlap in MxB diversifying selection with previously-mapped lentiviral restriction determinants contrasts with other restriction factors such as TRIM5α and TRIMcyp , in which genetic innovation directly correlates with capsid-binding and antiviral restriction [33 , 46–50] . It is possible that the selective pressure exerted on lentiviruses by MxB in vivo may be weak compared to other restriction factors , thereby reducing the likelihood or intensity of an MxB-capsid arms race . In single round or spreading infections , MxB mediates 5–20 fold restriction [9 , 10] , lower than the >100-fold restriction demonstrated for APOBEC3G , TRIM5α , and TETHERIN [12 , 51–53] . Another possibility is that the site of MxB restriction in capsid does not allow for viral escape . This possibility seems unlikely since MxB-resistant capsid mutants can be readily selected in vitro [14 , 54] . On the other hand , though MxB-escape mutant viruses might arise readily , it is possible that they are rendered more susceptible to interaction with other capsid-binding host restriction factors ( e . g . , TRIM5alpha , TRIMcyp ) . If this is true , MxB may play a crucial albeit indirect role in constraining capsid evolution as part of a multifaceted interferon response . Finally , it is formally possible that the molecular determinants of MxB anti-lentivirus activity do not mediate direct binding to lentiviral capsid but instead mediate its interaction with a cellular cofactor , leading to purifying rather than diversifying selection . Our findings are reminiscent of previous studies in which we found diversifying selection of the primate Bst-2/Tetherin restriction factor at nef- but not vpu-interacting sites [55] . While such analyses cannot irrefutably prove that nef drove diversifying selection of primate Tetherin , it strongly argues that vpu did not . Using a similar rationale , we conclude that a different lineage ( s ) of viruses , distinct from currently known simian lentiviruses , have shaped MxB evolution in primates . We considered whether LINE-1 retrotransposons may have driven MxB evolution based on the recent discovery that they are also restricted by MxB [56] . However , LINE-1 restriction is independent of MxB's N-terminal tail [56] , so it is unlikely that a conflict with retroelements drove diversifying selection in the N-terminus of primate MxB . Instead , we hypothesize that the evolutionary signatures of diversifying selection in MxB N-terminus , together with the evolutionary constraint acting on its GTPase domain , implicates its action against a widespread family ( or families ) of as-yet-identified pathogens . Interestingly , the signature of diversifying selection distinguishes the mode of MxA and MxB antiviral specificity . Just as positive selection in MxA is centered in the loop L4 , which mediates its target recognition [13] , we predict that individual changes in the MxB disordered N-terminus may also define MxB antiviral specificity . MxA ( cytoplasmically localized ) and MxB ( nuclear pore localized ) proteins restrict different classes of viruses [38] , which likely explains their stable retention as paralogs in most mammals . In light of this fact , it is surprising that gene conversion has resulted in the loss of MxB in rodents , followed by retention of an additional MxA-like gene in some . Interestingly , mouse Mx1 and Mx2 appear to have subfunctionalized , diverging in their antiviral range by localizing to different cellular compartments ( cytoplasmic and nuclear ) [22 , 27 , 28] . Therefore , the conversion of MxB to MxA in some rodents allowed for the refinement of MxA-like antiviral activity against cytoplasmic and nuclear replicating viruses , even at the expense of ancestral MxB-like activity . Although mouse Mx1 is nuclear , its localization is distinct from MxB ( human MxB localizes to the nuclear pore whereas mouse Mx1 forms discrete nuclear foci ) . Furthermore , the evolutionary acquisition of a C-terminal NLS on Mx1 is unique to mouse-like rodents [57] . Unlike rodents , however , pseudogenization of MxB in other lineages ( e . g . armadillo , felids and squirrel ) does not appear to be have been driven by MxA diversification . Instead , these may reflect cases in which MxB-targeted pathogens went extinct , relaxing selective pressure to maintain MxB activity . A similar loss of constraint may also explain the dual loss of both MxA and MxB in toothed whales [39] . Rodents appear to be distinct from other mammals in their Mx gene conversion profiles . In other eutherian mammals , including primates , gene conversion between MxA and MxB appears to be largely restricted to a region encoding parts of the BSE and GTPase domains . Such localized gene conversion could be the result of an uncharacterized recombination hotspot . However , it is also possible that this is the indirect result of selective constraint . For example , a swap of the BSE-GTPase domain is least likely to deleteriously impact the antiviral repertoire of the Mx paralogs . As a result , gene conversions spanning the BSE-GTPase domain , but not other domains , might be more easily tolerated . In contrast to this ‘tolerated conversion’ model , it is also possible that the frequent gene conversion of the BSE-GTPase domain is instead directly favored by selection . We speculate that gene conversion may serve as an adaptive mechanism ( akin to diversifying selection ) to escape antiviral antagonism . For instance , if the MxB BSE-GTPase domain were the direct target of viral antagonism , replacing it with the diverged but functionally equivalent MxA BSE-GTPase domain might be a rapid means to escape antagonism without compromising function . In support of this latter idea , there are residues under diversifying selection in the GTPase domain of Mx proteins in primates [13] , which might be the result of a genetic conflict to escape viral antagonists . Many antiviral gene families have undergone extensive , lineage-specific copy number variation , reflecting distinct bouts of pathogen-driven evolution and highlighting genetic variation in the antiviral repertoires of even closely related species . At first glance , Mx antiviral genes appear to belie the highly dynamic nature of antiviral gene copy number variation . Instead , closer examination reveals the highly dynamic evolution of Mx antiviral genes , with both diversifying selection having driven changes in presumed viral specificity domains and recombination homogenizing the catalytic domains of Mx proteins . Due to both these evolutionary forces , there is little likelihood of Mx ‘orthologs’ maintaining functionally analogous antiviral repertoires . At the very least , our analysis raises the need for caution in functional assignments of Mx genes based on the established “Mx1" and "Mx2” nomenclature , especially for rodents . Indeed , similar whole or partial gene conversion events are likely to have shaped many mammalian antiviral multigene families in which paralogs are present in genetic proximity to each other [2 , 4 , 58] .
Primate fibroblasts were purchased from the Coriell Cell Repository . Cells were cultured in DMEM ( Gibco ) supplemented with 10% FBS and 5% penicillin/streptomycin . RNA was isolated from cultured primate fibroblasts using the RNeasy kit ( Qiagen ) . RT-PCR to obtain primate MxB cDNA sequences was performed using HiFi one-step RT-PCR ( Invitrogen ) with the following primers: Hominoid F 5’—ATGTCTAAGGCCCACAAGCCTTGG—3’ R 5’—GTGGATCTCTTTGCTGGAGAATTGACAGAGTG—3’ Old World monkey F 5’—ATGTCTAAGGCCCACAAGTCTTGGC—3’ R 5’—RTGGATCTCTTCGCTGGAGAATTGACAGAG—3’ New World monkey F 5’—ATGTCTAAGGCCCACAKGTCTTGGCC—3’ R 5’—CTCTGTAAATTCTCCAGTGAAGRGATCCACGACTACAAAGACGACGACAAATGA—3’ Gel extracted products were directly sequenced ( Sanger method ) . Contigs were assembled using CodonCode Aligner . Detailed information on primate species and cell lines used in this study can be found in S2 Table . Sequences have been deposited in GenBank ( accession numbers KT698228-KT698252 ) . Mx genes were retrieved from publically available databases ( non-redundant nucleotide collection , reference genomic sequences , high-throughput genomic sequences and whole-genome shotgun contigs ) using BLASTN or TBLASTN from at least one representative of all sequenced mammalian orders ( S1 Table ) . After including sequences obtained through RT-PCR , multiple sequence alignments were conducted using CLUSTALW or MUSCLE , and adjusted manually . To determine the evolutionary relatedness of Mx genes , maximum-likelihood phylogenies ( 1000 replicates ) were constructed using PhyML under the GTR substitution model either locally or through the phylogeny . fr website [59 , 60] . Analysis of Mx locus synteny was evaluated in Ensembl using the Comparative Genomics Alignment tool for standard reference genomes or by retrieving genomic sequences and using BLASTN analyses to determine order of Mx homologs and flanking genes . We examined alignments for evidence of recombination using the GARD algorithm [20] from the HyPhy package [17] , run on local computers . Briefly , for each in-frame alignment to be analyzed , we first used HyPhy's NucModelCompare to suggest the best-fitting nucleotide substitution model ( using as input a maximum-likelihood tree generated from the entire alignment using PhyML [59] and the GTR model ) . We then supplied the alignment and best-fitting substitution model to GARD , using the general discrete model of site-to-site rate variation with 3 rate classes . Breakpoints assigned by GARD were independently assessed by generating bootstrapped ( n = 1000 ) maximum-likelihood phylogenies in PhyML . Maximum-likelihood tests were performed with CODEML implemented with the PAML software suite [16] . Input trees were generated from each alignment using PHYML [59] with the GTR+I+G nucleotide substitution model; using alignment-specific trees is more appropriate than using the species tree in this case , given that recombination has scrambled the orthologous relationships of these genes . Mx coding sequence alignments were fit to NS sites models that disallow ( M7 or M8A ) or allow ( M8 ) ω > 1 . Models were compared using a chi-squared test ( degrees of freedom = 2 ) on twice the difference of likelihood values to derive P values reported in Fig 1A and S1A Fig ) . Analyses were robust to varying codon frequency models ( F3x4 and F61 ) ; results from analyses using model F3x4 , which we empirically find to be more conservative , are shown in all figures . The model 7 vs 8 comparison was also robust to the initial omega value used ( 0 . 4 , 1 or 1 . 5 ) ; model 8a cannot be run with other initial omega values . In cases where a significant difference ( P < 0 . 01 ) between M7 versus M8 ( or M8A versus M8 ) was detected , the Bayes Empirical Bayes ( BEB ) analysis was used to identify codons with ω > 1 ( reporting values with posterior probability ≥ 0 . 95 ) . The depiction of positively selected sites on crystal structure representations was carried out in PyMol ( pymol . org ) . We also used the REL algorithm of the HyPhy package to detect selected sites [17]—we uploaded each alignment to the DataMonkey website , used the model selection tool to select the most likely evolutionary model , and used that model and the alignment as input to the REL algorithm .
|
Evolutionary analyses have the potential to reveal not only biochemical details about host-virus arms-races but also the nature of the pathogens that drove them . Primate MxB was recently shown to restrict the replication of primate lentiviruses , including HIV-1 . However , we find that positive selection in primate MxB is incongruent with known molecular determinants of lentiviral restriction . This suggests that MxB has antiviral activity against a broader range of viruses than is currently appreciated . We also identified multiple losses of MxB in mammals , as well as rampant recombination between Mx paralogs , which has distorted gene orthology . Our study illustrates the importance of evolution-guided functional analyses of antiviral gene families .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
Evolutionary Analyses Suggest a Function of MxB Immunity Proteins Beyond Lentivirus Restriction
|
Changes of synaptic connections between neurons are thought to be the physiological basis of learning . These changes can be gated by neuromodulators that encode the presence of reward . We study a family of reward-modulated synaptic learning rules for spiking neurons on a learning task in continuous space inspired by the Morris Water maze . The synaptic update rule modifies the release probability of synaptic transmission and depends on the timing of presynaptic spike arrival , postsynaptic action potentials , as well as the membrane potential of the postsynaptic neuron . The family of learning rules includes an optimal rule derived from policy gradient methods as well as reward modulated Hebbian learning . The synaptic update rule is implemented in a population of spiking neurons using a network architecture that combines feedforward input with lateral connections . Actions are represented by a population of hypothetical action cells with strong mexican-hat connectivity and are read out at theta frequency . We show that in this architecture , a standard policy gradient rule fails to solve the Morris watermaze task , whereas a variant with a Hebbian bias can learn the task within 20 trials , consistent with experiments . This result does not depend on implementation details such as the size of the neuronal populations . Our theoretical approach shows how learning new behaviors can be linked to reward-modulated plasticity at the level of single synapses and makes predictions about the voltage and spike-timing dependence of synaptic plasticity and the influence of neuromodulators such as dopamine . It is an important step towards connecting formal theories of reinforcement learning with neuronal and synaptic properties .
We consider a Spike Response Model neuron with index that receives input from other neurons . The input spike from neuron arrives at time at a synapses onto neuron and causes there an excitatory ( or inhibitory ) postsynaptic potential ( EPSP or IPSP ) of time course and amplitude . The EPSPs and IPSPs of all incoming spikes are added to the membrane potential of neuron . Spikes are generated stochastically with an instantaneous rate ( or stochastic intensity ) ( 5 ) where is a positive function that increases with the membrane potential , see also Eq . ( 24 ) . Immediately after a spike of neuron at time , the neuron enters into a state of relative refractoriness , which is implemented by a hyperpolarizing spike afterpotential . Thus the total membrane potential of the Spike Response Model neuron is [20]: ( 6 ) where is the resting potential , is the set of presynaptic spikes , is the set of postsynaptic spikes up to time . Using this neuron model , we can calculate the probability that neuron generates a specific spike train with firing times during a trial of duration [34] , see Methods , Eq . ( 25 ) . Some of the spikes of neurons occur just before a reward is delivered , others not . The aim of learning is to change the synaptic weights so that the probability of receiving a reward increases . We consider learning rules of the form ( 7 ) where is the learning rate ( controlling the amplitude of weight updates ) , the moment when the animal hits the target or the wall , is the positive reward for finding the target , the ( negative ) reward for bumping into a wall and b a reward baseline , for instance an estimate of the positive reward based on past experience . The eligibility trace evolves according ( 8 ) where is the spike train of the postsynaptic neuron , the Dirac function , the eligibility trace time constant , a parameter with units of time , and the derivative of the function . Because of the parameter , the learning equations ( 9 ) and ( 8 ) define a family of learning rules , rather than one single instance of a rule . The parameter is a specific feature of our model which allows to turn the model from a strict policy gradient method ( , [33] , [34] see methods ) to a naive Hebbian model ( , see below the discussion of the postsynaptic factor ) . Thus we are able to link and compare these conceptually different rules via the modification of . We note that for small firing rates , Eq . ( 9 ) approximates the optimal policy gradient rule of [33] , [34] , while for larger firing rates , it enhances the Hebbian component of the rule . For , the term in the square brackets goes to so that for learning is driven by the Hebbian correlation term . In the main body of the simulation results , we pick a fixed value of which implies that we use a policy gradient method with a Hebbian bias . The estimate of the positive reward is calculated as a running mean updated at the end of the trial according the following equation: , with being the number of the trial and being the reward at the end of the trial ( 1 or 0 ) and the width of the averaging window . We will now show that Eqs . ( 7 ) and ( 8 ) can be interpreted as a three-factor learning rule for spiking neurons , within the general framework outlined in the introduction . The learning rules discussed in the previous subsection ( with , and ) are tested on a simulated Morris Watermaze task with variable start condition , a task known to involve hippocampus [51] . Hippocampus is represented as a population of place cells , with place cells centers organized on a rectangular grid . These model place cells project onto ‘action’ cells , putatively placed in the nucleus accumbens . The population of action cells represents the next action to be chosen by the model rat and is organized in a ring-like topology with lateral connectivity of the Mexican-hat type; see Figure 2 . We perform simulations of a model rat navigating in a square maze of , with a constant speed of 20cm/s . The rat performs a number of trials , with each trial consisting of an attempt to find the goal within a time limit of 90 seconds . At the beginning of each trial , the rat is placed near one of the walls of the maze . Actions are chosen at theta frequency ( every 200ms ) . Between two action choices , the simulated rat moves by about 4cm . The rewarded position ( target ) is at a random position near the central region of the maze and remains fixed at the same position within a set of trials whereas the initial position of the rat varies , as in the experimental paradigm [51] , [65] , [66] . Positive reward is only given if the rat reaches its target and negative reward if it hits the wall . Thus , synaptic modifications take place either at the time the rat reaches the platform , , or at the time the rat hits a wall , . For an overview of the algorithm see Figure 4 . When a new set of trials starts , the positions of both the rat and the goal are reinitialized as well as the synaptic release of all plastic synapses in the model . Thus each new set of trials corresponds to a different animal .
In the introduction we mentioned two classes of theoretical reinforcement learning algorithms , that is , temporal difference ( TD ) learning methods on one side [1] , [43] and policy gradient methods on the other side [39] , [40] . Our model task and model architecture would allow to test both types of algorithm in the form of a three-factor rule ( see [45] , [52]–[54] for examples of a TD algorithm for this task ) . One major difference between the TD algorithms and the algorithm in this paper lies in how the global factor encodes neuromodulatory feedback about the reward . In the case of TD-learning , the global factor expresses the difference between the reward received and the expected reward ( where the expected reward is calculated from the temporal difference between reward expectations of subsequent states [1] ) , whereas in the case of the gradient learning algorithm of this paper the global factor correspond to reward itself , possibly after subtraction of a baseline . Here we used a variant of the idea of a baseline , since we subtracted the mean reward averaged over order previous trials , see also [41] . Subtracting the expected reward should help rapid re-learning in case of the change of the learning task ( e . g . , by moving the escape platform to a different location ) [67] . Similar to TD learning the global factor can be interpreted in this case as reward minus expected reward . In contrast to TD learning , the expected reward arises from a running average , rather than a difference in reward expectation across different states as in spike-based TD algorithms [37] , [45] . Experiments on dopaminergic neurons suggest that the phasic dopamine signal indeed encodes a TD-like error signal [22] although other interpretations of the dopamine signal [68] and the involvement of other neuromodulators is also possible [69] . Our spike-based navigation model features a continuous description of state and action . Unlike traditional TD models with discrete state and action space , increasing the number of neurons while keeping the width of place fields and the width of lateral interactions between action cells constant ) does not change the performance of our model . In addition , the model provides insight in studying decision making in the context of navigation . We hypothesized that activity is modulated at theta frequency . Note that we implemented an extreme situation where the action choice is taken at the end of each theta cycle . However , it is easily possible to have the rat take an action as soon as the activity profile is formed . The time necessary to create an activity profile determines then a minimal time for deciding a new action . If this is so , then our model predicts that the time it takes to choose the next action is much faster after learning than before learning , because activity profiles are more rapidly formed with strong feedforward input - as it would occur after learning . To test the potential of our spike-based reinforcement rule , we have applied it to a biologically relevant navigation problem , i . e . , the Morris water maze task with variable start condition [51] . Our model which is based on a simplified model of place cells and action cells reproduces behavioral data of real rats in terms of escape latency versus learning time . The model consists of about 700 spiking neurons , in two layers and includes both feedforward and lateral connections . In the first trial , the model rat moves in a random trajectory and finds the hidden platform by exploration . Across several trials , approach paths towards the platform are reinforced , so that the escape latency is reduced . A positive reward is delivered when the model rat reaches the target location . In the model , we also use negative reward at the boundaries of the maze so that the simulated rat will learn to avoid the walls . This aspect does not reflect the fact that , normally , during development ( or even because of reflexes present at birth ) we could assume that the rat already knows how to avoid obstacles prior to the start of the watermaze task . However , since we did not want to include into the model prior knowledge about obstacle avoidance , we let the simulated rat ‘discover’ the effect of the walls . Since our model assumes the existence of place cells , we must assume , however , that the rat has had some pre-exposure to the environment long enough to establish place fields . Experiments have shown that place fields are established during a first exploration of the environment , so that during the learning task , they can be considered as given . Moreover , typical experiments require prior habituation of the animal to the environment , so that place cells may be formed . A model where place cells are learned from visual input and path integration is also possible [53] . While in our model place cells can be easily linked to cells in hippocampus , a direct identification of the action cells with the biological substrate is more problematic . In rodents , navigation in water maze task involves two competing pathways [70]–[72] . The first one is involved in taxon navigation ( e . g . , approaching a visible target , which could be achieved with stimulus-response habits [73] also called response learning [71] ) and associates visual input directly with motor actions . It is independent of hippocampus and the action choice for this navigation strategy can presumably be linked to the the dorsal striatum of the basal ganglia ( caudate-putamen in the rat ) . The second one is concerned with locale navigation ( also called place learning [71] or cognitive map [74] ) and this is the relevant pathway in the context of the present model . It relies on hippocampus [51] , [70] , [71] where the activity of place cells presumably encodes the location of simulated animal . The choice of motor actions is presumably encoded in the nucleus accumbens ( NA ) of the ventral striatum where our hypothetical action cells could be located . The Mexican hat connectivity between action cells is a simplification of a more complex wiring scheme , where excitatory neurons project to inhibitory neurons , which in turn inhibit other action cells that encode for “different” directions , see for example a biologically plausible winner-take-all [75] . However , to reduce the connectivity in our network , we chose to simulate the equivalent but simpler Mexican hat scheme . One limitation of the model is that learning only takes place in the presence of a reward signal with the consequence that learning can only occur in a limited radius around a reward . The radius is related to the time scale of the eligibility trace , governed by the time scale . In a large environment where at a fixed speed it takes much longer than to traverse the environment , information about the target falls off exponentially with a spatial scale . In our case we would encounter this limit only if the environment were scaled by a factor significantly larger than two . In a TD framework , the situation would be different: even without an eligibility trace , information about the presence of the reward can slowly diffuse across the landscape of estimated reward expectation values where is the position , even beyond the radius discussed above . This slow diffusion of reward information is possible because the update is not proportional to the reward itself , but to a factor where gives the difference between the reward estimation at location and that of the previous location and is the discount factor . An implementation of a TD learning structure in spiking neurons is possible using the actor-critic scheme [37] , [45] . If a TD algorithm is implemented in discrete time with time steps , and if the rat runs as before at a constant speed , the distance travelled between two time steps is . After convergence , the value function decreases exponentially with the distance from the target on a lenght scale . ( In other words , once the exponentially decaying dependence is reached , the in the update rule vanishes ) . A comparison with the result in the previous paragraph shows that the time scale of the eligibility trace in our model plays a role similar to in the TD model . The role of the eligibility trace has been extensively discussed in [35]; in our interpretation the eligibility trace is implemented in the synapse and its time constant corresponds to the decay time of some biochemical substance . The parameter is an ad-hoc parameter that allows us to vary the behavior of the learning rule from pure Hebbian to optimal in the sense of policy gradient theory . We do not wish to explicitly associate it with a biological substrate , but in our model it would be closely related to the voltage dependence of LTD . Recently , the influence of neuromodulators on spike-timing dependent synaptic plasticity has been investigated in a small number of studies [31] , [76] . These studies show that dopamine acts on the temporal profile of STDP , rather than a simple scaling of STDP . This result is in contrast to some of the assumptions of standard reward-modulated STDP [35] , [36] , but also in disagreement with policy gradient rules [33] , [34] , [38] and the learning rule discussed in this paper . For plasticity in the cortico-striatal synapse [31] , but not for glutamatergic synapses in hippocampal neurons [76] , dopamine is necessary for synaptic plasticity . In other words , learning is gated by the presence of dopamine . The plasticity rule in the cortico-striatal synapse is in that respect similar to the reward-gated plasticity rules in the present paper . Interestingly , the striatum is potentially involved in action selection . It should be noted that in standard cortical STDP experiments [77] , [78] the level of dopamine and other neuromodulators is not explicitly controlled and a background level of dopamine cannot be excluded . Therefore , it is unclear whether cortical STDP is unsupervised or shows a , possibly weak , dependence upon neuromodulators . An important parameter in our family of learning rules is the parameter , that tunes the learning rate such that for neurons that fire at high learning rates LTD is reduced . To see this , consider an instantaneous firing rate . Then the term converges to . Hence , the decrease of the eligibility trace in the absence of spikes is limited . Note that because of high rates correspond to large depolarizations of the membrane potential . For , the term vanishes , and the membrane potential no longer enters the update of the eligibility trace . In this case the eligibility trace pick up Hebbian correlations between EPSPs caused by presynaptic spike arrival and postsynpatic firing . The case corresponds to the learning rule derived from the reward maximization as shown in the methods section , i . e . , . For the two postsynaptic terms , i . e . , spike firing and voltage dependence cancel each other on average , because spikes are generated with the stochastic intensity , hence where angular brackets denote expectation values . However , a specific realisation of a spike train ( e . g . , one with more spikes than expected ) may lead to a reward whereas another one ( with less spikes than expected ) does not . In this case only the rewarded one is learned , making it more likely that the same spike train is reproduced again for the same input [34] . In fact , a large class of learning rules for conditioning can be explained as a reinforcement of the covariance between reward and a noise-induced variation of the output [79] . There are three reasons why the standard policy gradient rule with derived from reward maximization is not applicable in our scenario . ( i ) Large learning rate . The learning rule derived from reward optimization is a batch rule , i . e . , it assumes averaging across several realisations and many inputs . For the transition to the online rule we had to assume a very small learning rate so as to make the learning self-averaging . If learning is slow , then thousands of trials are needed before the weights change significantly , so that online and batch have nearly the same effect . In order to explain biological learning paradigms , we need , however , to achieve learning after as few as ten trials . If we work with a large learning rate , then terms of the form that average away in the batch rule , can make a big contribution in the eligibility trace of each single trial and can cause weight changes that are not causally linked to the reward . Thus the eligibility trace encodes noise , rather than relevant correlations . With small learning rate , these correlations would average away ( and only those systematically linked to the reward would survive ) , but with a big learning rate these changes act like a diffusion process . Moreover , the effect of the diffusion increases with the number of spikes in the decision window and therefore is highest for neurons having a large firing rate . Large firing rates appear in particular after learning for neurons inside the activity bump , because strong lateral input is added to strong feedforward input . Hence the eligibility trace is most noisy in the center of the bump , as shown in Figure 8 B . ( ii ) Decision by firing rates , not by spikes . The close relation between reward-maximisation by policy gradient rules and supervised learning shows that the spike-based rule with is optimal to learn a specific spatio-temporal spike pattern [34] . However , what counts for the action choice in our simulations is the firing rate accumulated over 200ms . To understand the importance of this distinction let us consider two Poisson neurons coding for actions ‘left’ and ‘right’ , respectively . The action ‘right’ is the rewarded one . Suppose the neurons receive inputs that drives the neurons coding for ‘left’ at an intensity and the other at . Suppose , because of intrinsic noise , the neuron coding for ‘left’ fires 2 spikes in a decision interval of , while the neuron coding for ‘right’ fires 9 spikes in the same time interval . If actions are chosen according to maximal firing rates , the neuron coding for right wins , the system performs the ‘right’ action and receives reward . However , the term is negative for the neuron coding for ‘right’ and ‘positive’ for the neuron coding for ‘left’ . Hence , after reward is received action ‘right’ is weakened , while action ‘left’ is reinforced , in contradiction to the fact that action ‘right’ is the correct one that should be reinforced . To put it differently , action neurons have to learn that ( a ) precise spike timing is irrelevant and that ( b ) even the absolute rates are irrelevant because all that matters is the firing rate relative to those of the other neurons . Since the policy gradient rule is desigend to learn precise spatio-temporal spike patterns , it is not ideally suited for our paradigm . In contrast , reward-modulated Hebbian learning just make the neurons that fired at high rate ( and influenced the action ) fire at even higher rates . In the specific task we are considering this happens to be a viable strategy . ( iii ) Populations of neurons , not single neurons . Furthermore , because of the formation of an activity bump and the readout by a population vector the decision about actions is taken by a population of neurons rather than individual neurons . Learning in populations suffers from the problem that firing of individual neurons may differ from the majority vote that led to the actions , so that giving appropriate feedback is nontrivial [80] . Figure 8 illustrates the detrimental interaction of points ( i ) – ( iii ) for the standard policy gradient rule . We focus on a presynaptic neuron which codes for the current location of the rat so that synapses from to all action neurons are active . The instantaneous firing rate represents the activity bump ( Figure 8 A ) . Despite the fact that the term has an expectation value of zero , the term gives a non-neglibible contribution in each trial , see also Figure 1 C – as it should be since policy gradient rules need to exploit fluctuations . However , we would like to emphasize two aspects . First , the standard deviation of grows with time , similar to a diffusion process . Second the diffusion constant increases with the instantaneous rate . Therefore the deviation from the expected value increases with the expected number of spikes the neuron emits during the decision interval of length . The eligibility trace is sensitive to this deviation . In the case of our action learning model , the consequence of the above argument is that the set of significantly positive eligibility traces for fixed presynaptic neuron includes not just action neurons within the activity bump , but also those representing other directions; see Figure 8 B . Moreover , the variation of eligibility traces between neighboring neurons inside the activity bump is big , because the expected number of spikes is higher for neurons inside the activity bump . In particular , several synapses from a fixed presynaptic neuron onto neurons in the bump have eligibility traces that are significantly negative ( corresponding to the fact that some neurons in the bump fire less spikes than expected from the firing rate , see point ( ii ) above ) . This leads to the problem that eligibility traces of individual neurons do not reflect the action choice represented by the population of active neurons [80] . Simply speaking , neurons inside the bump are those that determine the action even though their eligibity trace can be negative . The parameter in our learning rule gives a systematic positive bias of the postsynaptic term for those postsynaptic neurons that have a large firing rate . Thus the eligibity trace is maximal for neurons within the bump of activity , i . e . for those representing the action that is actually chosen; see Figure 8C . Hence , if the sequence of actions leads to a reward later on , the synpatic weights between those presynaptic place cells and postsynaptic action cells that actually led to the sequence of actions are maximally strengthened . Because of the bounds on the weight dynamics , these weights will eventually converge towards a release probability of . We note that all neurons outside that activity bump have very low activity , so that has a zero average and only small fluctuations . Hence , a learning rule with is expected to work better in the case of large learning rates , and high firing rates , and a decision criterion based on a population vector calculated over a long time period . In a general spike-based learning problem where the aim is to learn a spatio-temporal spike pattern , the high variability of eligibility traces would allow to explore a large space of firing patterns . However , in our case with lateral interactions and decisions based not on detailed firing patterns , but only on population vector data integated over 200ms , the bias towards high activities identifies neurons in the bump that participate in the action choice . Indeed , a learning rule with does work in the situation where ( a ) there are no lateral interactions between the action cells or ( b ) decisions are based on less than one spike per neuron on average . In the latter case , every spike is unexpected , and basing a decision on the population vector chooses an action that is indeed caused by a fluctuation . In principle four action neurons would be sufficient to encode the direction of the next action ( e . g . , [45] , [53] ) . In this case , learning rules based on either policy gradient [45] or naive Hebb [53] work . However , it is likely that in biological brains actions are encoded by large populations of neurons . In order to achieve fast learning despite a large population of action neurons , action neurons must share information during learning – and this can be achieved by the formation of activity bumps . The results of this paper show that in the presence of activity bumps and population vector read-out based on spike counts , the spike based policy gradient rule no longer works , whereas a rule with a bias towards Hebbian correlation does . From a technical point of view , neither stochastic synapses nor voltage dependent plasticity is critical for the function of the model , however they are both desirable properties for the biophysical plausibility of the rule . In our model , the stochastic release probability of the synapses is hard-bounded in order to maintain reasonable values , for a biophysical implementation of such bounds see [46] . Also a reset it is not necessary to take place exactly every 200msec; in principle may occur at any point that the activity bump is formed . We require to reset the activity in the action neurons layer only ( or equivalently we could clamp the AC activity for say 10ms ) so that the activity profile will not become “sticky” , but in no other way the learning would be affected . Without reset , the rat will end up again learning the position of the platform , but its movements will become more curved . A negative input would be desirable after a decision is formed so that at the beginning of the learning the next action will not depend on the previous one . This negative input may arrive at any point after a decision ( activity bump ) has been formed . We chose 200ms so that this could coincide with the theta rhythms , but it could have been 150ms or 300ms , or a random interval ( as we demonstrate in simulations ) .
To derive a learning rule for a highly connected network with action cells with lateral connections receiving from input from place cells , we shall first consider a restricted scenario where the rat always starts a trial in the same initial location and is left to move around for a fixed duration . We shall denote by ( ) the spatio-temporal spike pattern generated during this time by all place ( action ) cells . The reward , administered at the end of each trial , depends on the trajectory of the rat in the water maze . Given the fixed initial location , this trajectory is determined by the firings of the action cells . So we write reward as a function , where b is the reinforcement baseline [39] , without explictly noting the dependence on the initial position of the rat . Expected reward then is [32] , [34] ( 17 ) here denote the strengths of the synapses connecting the action to the place cells , and is the probability that the network generates the total spike pattern . In our model can be decomposed as ( see also Decomposition of probability ) : ( 18 ) Here is the function giving for the action cell the single neuron probability that it generates its spike train with an input consisting of all the other spikes produced by the network . Similarly , is the single neuron probability function for the spike train produced by the place cell given its input ( determined by the other spikes in the network ) . Note that the above product form does not imply that the spike trains are statistically independent . This is obviously not the case: First , due to the lateral connections between the action cells , and , more importantly , due to the simple fact that the action cells decide on the rats trajectory and thus influence the firing of the place cells . The product form simply represents the fact that the internal stochastic processes which modulate the translation of presynaptic input to postsynaptic output are assumed to be independent between different cells . In other words , given the input spikes from all other neurons and its own previous spikes up to time , the neuron decides locally whether it fires between and or not ( i . e . , we activate an independent random process for each neuron in each time step of the simulation ) , see section Decomposition of probability . An explicit form for would be rather complicated , due to the involved calculations mapping the action cell firings to the trajectory of the rat . Luckily , we just explicitly need . Note , and this is in fact the crucial feature of the decomposition , that does not depend on all feed-forward weights , but only on the weight vector of the synapses actually projecting onto neuron . To calculate the gradient of the expected reward ( 17 ) , we first rewrite the probability as ( 19 ) and note that in view of ( 18 ) the term in square brackets in fact does not depend on ( even if this is not apparent from the notation ) . Now , for the synapse connecting place cell to action cell the gradient calculation is ( 20 ) The last line yields a batch rule for synaptic changes . We first average ( 21 ) over many trials and then use the result to update the synaptic strength . The biologically reasonable online version of this is to already update after each single trial , i . e . ( 22 ) Often we replace the reinforcement baseline with the estimate of upcoming reinforcement based on past experience [39] . In the context of on-line learning , our initial requirement of a fixed initial position is no longer necessary since we calculate the expected reward by averaging not just over trials with the same but also over trials with different initial positions . The crucial element of the learning rule is the conditional probability of creating certain outputs ( and hence taking certain actions ) given an input . In order to calculate the conditional probability that neuron fires a spike given the past , we need to introduce a neuronal model . Following the approach of Pfister et al [34] , we assume that neuronal activity can be described by the Spike Response Model ( SRM ) [20]: ( 23 ) where is the membrane potential of the neuron , is the resting potential , is the set of postsynaptic spikes , is the set of postsynaptic spikes up to time , the synaptic strength between the presynaptic neuron and the postsynaptic neuron , is the th firing time of the presynaptic neuron and the th firing time of the postsynaptic neuron . The sum is restricted to firing times before time . The kernel describes the time course on an excitatory postsynaptic potential ( EPSP ) and the spike-afterpotential . We would like to emphasize that for an exponential kernel and exponential spike- afterpotential , the SRM becomes identical to a leaky integrate-and-fire model with membrane time constant [20] as used in Eq . ( 11 ) in the results section . Given a membrane potential , action potentials are generated by a point process with stochastic intensity , where is some positive nonlinear function . To be specific , we take an exponential function ( 24 ) where the formal firing threshold , and , parameters . Thus the higher the membrane potential , the more likely is the neuron model to fire . With the above neuron model , the probability of neuron to emit a particular set of postsynaptic spikes in the period given the input and from all neurons in the network except neuron is given by: ( 25 ) with representing the postsynaptic spike train of the neuron up to time as a sum of the Dirac functions , i . e . Taking the partial derivative in respect to the synaptic weight , we have the following equation [34]: ( 26 ) where , being the set of postsynaptic spikes that occurred before , and the EPSP kernel . Note that for the exponential function , we have , so the learning rule becomes: ( 27 ) Here is the total reward received during or after a trial of total duration . In order to illustrate the mathematical structure of Eq . ( 27 ) , we consider the time point at the end of the trial and integrate backwards in time ( 28 ) where is the momentary reward at time . Here is a weighting function that allows us to give different weights to events in the past . If we take for and zero otherwise , and evaluate at time point , we retrieve exactly Eq . ( 27 ) under the assumption that the reward is given according to one of the following two schedules: ( a ) all the reward is delivered at time , i . e . , and a negative is applied at every time step; this is the scenario we have in mind with our notation that we use throughout the rest of the methods section , since it simplifies the development of the theory . Or , ( b ) no reward is given in the interval and an effective reward is applied at time , i . e . , . This is the scenario we used in the simulations in the main body of the paper . The baseline is either or . Starting from the interpretation ( a ) we can turn to an online rule in continuous time where rewards can be delivered at arbitrary moments . To arrive at a more elegant representation of the rule , we replace the step function by an exponential kernel for and zero otherwise . Then we have ( 29 ) is a learning rate and is called an eligibility trace [1] , [32] . For our specific model we have ( 30 ) Because of the exponential in the integral the eligibility trace can be rewritten as a differential equation ( 31 ) We consider stochastic binary synapses with . Synaptic transmission is stochastic with a release probability . Learning affects the release property so that increasing the weight of the synapse by the above update rule will increase the release probability . We choose proportionality factors so that the expectation of the binary synaptic transmission over time is equal to the continuous synaptic weight , i . e . . and thus , with , we have for binary synapses instead of Eq . 29 the following learning rule ( 32 ) We impose a hard bound that reflect the interpretation of as a probability of transmitter release . In order to guarantee sufficient exploration , we also impose a non-zero lower bound The factor can be absorbed by a learning rate yielding the final online-rule ( 33 ) We note the typical structure of a three-factor learning rule . The eligibility trace picks up correlations between EPSPs caused by presynaptic spike arrivals and postsynaptic firing times as in a STDP learning rule [34] which is then combined with the reward signal [33]–[35] . We extended our rule by introducing ad hoc a variant with a parameter : ( 34 ) In the limit of this reduces to the rule derived above . Eq . ( 34 ) in discrete form becomes: ( 35 ) with being the time step , being 1 if a spike is emitted in the interval and 0 otherwise and the hat ( ) operator denoting discrete firing times . The quantity is the probability that the postsynaptic neuron emits a spike in the interval given the input spike trains ( denoted in discrete time ) and is computed as ( 36 ) which computationally advantageous for large timesteps , see also [20] . In Figure 1 we plot the factor ( 37 ) The voltage trace is obtained by integrating Eq . ( 11 ) for constant input , i . e . presynaptic spike arrival is replaced by a positive constant . Interestingly the rule developed by [34] as well as the variation presented here can be mapped to Associative Reward Inaction ( ARI ) [39] , [81] in discrete time . With Eq . ( 27 ) , and ignoring the baseline subtraction , we have ( 38 ) Let us assume a rectangular EPSP of duration of one time step and unit amplitude . Hence , the EPSP can be replace by a binary variable if a spike has arrived at the synapse j at time , and with in the absence of a spike . We then have: ( 39 ) We note that according to the above derivation is a sigmoidal function of the membrane potential . Hence , dropping the hats ( that we used to denote discrete time ) we have exactly the update rule of the ARI: ( 40 ) Similarly the learning rules of [32] , [33] also correspond to ARI or its modern forms of policy gradient . In fact the rule in [33] is derived from the framework of [40] . The rule of [32] is a special case of the rules by [33] , [34] , since it makes use of a memoryless Poisson neural model , wheres our derivation here includes refractoriness via the kernel . Here we show that the probability of the place cell spike pattern and the action cell spike pattern to occur can be decomposed into the product ( 41 ) as mentioned in the Methods of the main text , Eq . ( 18 ) . The argument is similar to the unfolding in time used by Williams [39] , except that networks of spiking neurons are not Markovian . We claim that the above decomposition holds for an arbitrary network architecture including recurrent connections . Let be a collection of discrete random variables , a location index , a time index . Denote by the whole collection up to time . In our example , the index encompasses both the place and action cells . Moreover , ( ) if the corresponding cell did ( did not ) emit a spike at time . We assume that the sequence is generated by choosing at time the value with a probability . For spiking neurons the sequence determines the internal states ( membrane potentials ) at time and this modulates the probability of firing at time given the previous spike history , . We further assume that the internal stochastic processes which trigger the spikes are independent given the membranes potentials . Hence , ( 42 ) for . Because we can always write with a factor , we can iteratively apply an analogous multiplicative decomposition for , , , and receive a product representation of . To anchor the product we assume that ( 42 ) also holds at , and take this to mean that the initial values are statistically independent with probabilities given by . While consecutively applying ( 42 ) at each step of the decomposition we arrive at ( 43 ) Setting and reordering the product terms we can write ( 43 ) asand this is just the decomposition into the product across the place and action cells expressed in ( 41 ) . Model and Figures are produced with Matlab R2008b ( Linux version ) , developed by Mathworks . The model is implemented with custom-made code . For implementation details see Figures 4 and 9 . Parameter values are summarized in Tables 1 and 2 . The Euler method is used for integration . We discretize the learning rule equation according to the method in paragraph ‘From a single rule to a family of rules’ , in order to allow for large time steps . The standard time step in our simulation is . We have checked in additional simulations with smaller time steps of that the results do not depend on the step size ( data not shown ) .
|
Humans and animals learn if they receive reward . Such reward is likely to be communicated throughout the brain by neuromodulatory signals . In this paper we present a network of model neurons , which communicate by short electrical pulses ( spikes ) . Learning is achieved by modifying the input connections depending on the signals they emit and receive , if a sequence of action is followed by reward . With such a learning rule , a simulated animal learns to find ( starting from arbitrary initial conditions ) a target location where reward has occurred in the past .
|
[
"Abstract",
"Results",
"Discussion",
"Methods"
] |
[
"neuroscience/theoretical",
"neuroscience"
] |
2009
|
Spike-Based Reinforcement Learning in Continuous State and Action Space: When Policy Gradient Methods Fail
|
The upper respiratory tract mucosa is the location for commensal Streptococcus ( S . ) pneumoniae colonization and therefore represents a major site of contact between host and bacteria . The CD4+ T cell response to pneumococcus is increasingly recognised as an important mediator of immunity that protects against invasive disease , with data suggesting a critical role for Th17 cells in mucosal clearance . By assessing CD4 T cell proliferative responses we demonstrate age-related sequestration of Th1 and Th17 CD4+ T cells reactive to pneumococcal protein antigens within mucosal lymphoid tissue . CD25hi T cell depletion and utilisation of pneumococcal specific MHCII tetramers revealed the presence of antigen specific Tregs that utilised CTLA-4 and PDL-1 surface molecules to suppress these responses . The balance between mucosal effector and regulatory CD4+ T cell immunity is likely to be critical to pneumococcal commensalism and the prevention of unwanted pathology associated with carriage . However , if dysregulated , such responses may render the host more susceptible to invasive pneumococcal infection and adversely affect the successful implementation of both polysaccharide-conjugate and novel protein-based pneumococcal vaccines .
Global estimates suggest that approximately one and a half million deaths due to pneumonia , bacteraemia and meningitis are associated with pneumococcal infection annually , around two thirds of these occur in children in resource-poor countries [1]–[3] . In addition to this high disease burden , S . pneumoniae is also a common commensal of the upper respiratory tract colonising approximately 40–50% of children from 0 to 2 years of age in the United Kingdom [4] and up to 90% of African children in this same age group [5] . It is assumed that this commensal relationship is regulated by natural immunity to the pneumococcus , which is acquired from early infancy onwards [6] . This immunity is thought to result in a gradual decline in pneumococcal carriage and infection with increasing age , even in settings where the rates of invasive pneumococcal disease are high [7] , [8] . Classically , due to the undeniable protective efficacy of pneumococcal capsular polysaccharide vaccines , anti-capsular antibodies have been thought to be largely responsible for natural immunity to S . pneumoniae [6] , [9] . As a consequence , studies assessing T cell immunity to the pneumococcus , particularly in humans , have until recently been lacking . However , re-evaluation of the epidemiology has brought into question the central role of anti-capsular antibody [6] . Studies of colonization , antibody acquisition and the relationship with otitis media suggest that naturally-induced antibodies to pneumococcal protein antigens may be protective against disease [10] . The demonstration of CD4+ T cells that respond to pneumococcal protein antigens points to the possible contribution of these cells to the development of serotype independent protection against S . pneumoniae and the age-related decline in pneumococcal disease [6] , [11]–[15] . Experiments in the mouse have shown cell-mediated immunity to be an important protagonist in host immune defence against pneumococcal colonization following immunization with protein antigens . These studies have implicated the Th17 CD4 T cell subset in the promotion of mucosal clearance through the recruitment of neutrophils and macrophages . Indeed , it has been suggested that pneumolysin ( Ply ) , a cytotoxic protein antigen and TLR4 agonist which elicits protective immune responses in rodent challenge models , is essential to the generation of Th17 responses to S . pneumoniae [11] , [14] , [16] . We have investigated the nature of CD4 T cell immunity in the upper respiratory tract with increasing age , and the relationship between immunity at this site and that seen in the circulation . In addition , we have determined whether pneumococcus-specific Treg cells arise as a result of natural exposure and whether such cells modulate the nature of protective responses .
S . pneumoniae can inhabit the upper respiratory tract , particularly during childhood [6] , [10] leading to nasopharyngeal colonisation [17] , [18] . Multiple colonization events are likely to occur throughout life commencing from early infancy when they are most frequent . Due to ongoing bacterial exposure in the upper respiratory tract , lymphoid tissues within this anatomical site are likely locations for immune induction and depots for pneumococcal reactive lymphocytes . Support for this comes from a previous study assessing tonsil and blood CD4+ T cell responses to pneumococcal proteins which hinted at higher responses by tonsillar CD4+ T cells compared to those from blood [13] . To test this , we first compared CD4+ T cell responses to S . pneumoniae in upper respiratory tract lymphoid tissue with that in blood . Tonsil mononuclear cells ( MNC ) were cultured with recombinant pneumococcal Ply mutant protein [19] or supernatants generated from the culture of type 2 D39 S . pneumoniae bacterial cells ( SPNT ) as described previously [10] , [12] . CD4+ T cell proliferation was assessed after seven to nine days stimulation via 5 , 6-Carboxyfluorescein diacetate succinimidyl ester ( CFSE ) staining and flow cytometric analysis ( Figure 1a ) . When similar data from eight adult subjects ( >20years old ) was analysed , the mean percentage of CD4+ T cells proliferating to Ply was 33 . 5% ( ±6% ) and SPNT was 40 . 3% ( ±5 . 9% ) . Proliferation to a previously established positive control , influenza [20] was 49 . 2% ( ±6 . 9% ) and to the negative control of media alone , 24 . 1% ( ±6 . 2% ) ; thus proliferative responses to all three antigens were significantly greater ( p <0 . 05 ) than the background proliferation ( Figure 1b ) . The results reveal that anti-pneumococcal CD4+ T cell responses are evident in the palatine tonsil mucosal lymphoid tissues of adults . In order to compare the anti-pneumococcal immune responses in the tonsil with those systemically , we analysed paired tonsil and peripheral blood mononuclear cells ( PBMC ) CD4 T cell proliferative responses . Comparison of CD4+ cell proliferative response to both Ply and SPNT stimulation similarly revealed stronger anti-pneumococcal responses by tonsil compared to blood CD4+ T cells ( Figure 1c ) . This stronger tonsillar response occurred despite the subtraction of a slightly higher level of background response in a number of the tonsil samples , which might have been expected to mitigate against seeing such a difference . The response to influenza was similar in both compartments . These results suggest that pneumococcal-specific CD4+ T cell may be preferentially sequestered within mucosal lymphoid tissues located in the upper respiratory tract , the anatomical site for bacterial colonization . As the decline in pneumococcal carriage and invasive disease is thought to be the consequence of age-associated acquisition of natural adaptive anti-pneumococcal immune responses [4] , [5] , we next examined the mucosal CD4 T cell- mediated response to the pneumococcus from early infancy to mid-life . Evaluation of the proliferative response to Ply and SPNT by CD4+ cells of subjects aged between 2 to 39 years old revealed a gradual increase in responses with age , up to around early to mid-twenties and then a plateau thereafter ( Figure 2a ) . For anti-Ply responses the average rate of increase in the percentage of proliferating CD4+ cells between ages 2 to 12 years was calculated to be 0 . 34% per year while for SPNT this was at 0 . 4% per year . From early teens until 30 years rates dropped to 0 . 11% proliferating CD4+ cells per year for both anti-Ply and anti-SPNT responses . Anti-influenza responses showed a similar trend but at a higher intensity at all the age groups analysed . Interestingly , when we assessed the results of a separate study analysing the age-associated rates of community-acquired pneumonia ( CAP ) which is primarily due to S . pneumoniae infection , in the UK between 2000 to 2003 [21] , we observed that the combined proliferative responses of CD4 T cells to SPNT and Ply showed an inverse relationship to the levels of CAP ( Figure 2b , generated with permission from P . R . Myles ) . This finding suggests a potential role for the cell mediated anti-pneumococcal response in protecting from disease . T regulatory cells are increasingly being recognised as important modulators of responses to bacteria and viruses [22]–[26] . Interestingly , animal studies have revealed that the presence of commensal bacteria can induce Treg levels in vivo [27] , and similarly , we have established that Treg appear during the acquisition of natural immunity to Neisseria ( N . ) meningitidis , another upper respiratory tract coloniser [20] . Tregs can have unfavourable effects on host immunity by preventing the generation of effective immune responses . For example in humans , elevated Treg numbers are associated with chronic viral infections such as hepatitis B and HIV with the depletion of these cells in vitro resulting in improved T cell responses [28] , [29] . We therefore assessed the role of Tregs on the mucosal anti-pneumococcal responses by CD4+ T cells [30] , [31] . Previous works have shown that many , although not all populations of Tregs can express the IL-2 receptor subunit CD25 at high levels [22]–[25] . We therefore investigated whether pneumococcus-specific Treg cells are present in the tonsils by depleting the Treg cell containing CD25hi cell population from cultures of tonsil mononuclear cells ( MNC ) in order to assess the impact of such cells on mucosal anti-pneumococcal CD4 T cell responses . We found a significant increase in the proliferative responses of tonsil CD4+ T cells to Ply ( Figure 3a ) and SPNT ( Figure 3b ) after CD25 depletion in the majority of subjects at the age of 17 years and above but not for those below 17 years . As was observed previously [20] , analysis of the CD4 cell anti-influenza responses revealed no significant effect of CD25hi cell depletion on the level of proliferative responses to influenza ( Figure 3c ) . In order to confirm that the enhanced proliferation observed following the depletion of CD25hi cells was as a result of this population exerting an inhibitory effect in undepleted cultures , we subsequently added back the CD25hi cell fraction to the depleted MNC population at the original proportion and at three times the original proportion . Assessment of those subjects ( four of five ) showing notable increased CD4+ T cell proliferation to pneumococcus post Treg depletion revealed that restoration of CD25hi cell numbers resulted in the reversal of the observed increased proliferative responses to Ply ( Figure 4a ) in subjects above the age of 17 years . When CD25hi cells were added back at three times the original percentage , the proliferative responses decreased even further . When subjects below the age of 17 years were assessed ( Figure 4b ) , no notable changes in the proliferative response were observed between the undepleted , CD25hi depleted , CD25hi added back cell cultures . In keeping with the lack of effect of CD25hi cell depletion on influenza specific CD4+ cell responses , no significant change in the proliferative response to influenza were observed following the removal of , or the addition back of this population to depleted MNC populations , even at three times their normal frequency ( Figure 4c ) . Collectively , these data reveal that Treg cells are present within mucosal lymphoid tissues of adults and are able to significantly suppress the proliferative responses of CD4+ T cells to S . pneumoniae . CD4 T cells produce and secrete a variety of cytokines that control and co-ordinate effector mechanisms involved in pathogen clearance . In order to determine the nature of the anti-pneumococcal immune response in the mucosa , we assessed cytokine production in supernatants taken from cultures of palatine tonsil MNC following in vitro stimulation with Ply . We observed significant production of TNF-α , IL-2 and IL-10 ( p <0 . 05 ) and a modest but not significant increase in IL-17 ( p = 0 . 06 ) compared to the levels observed in unstimulated cultures ( Figure 5a ) . In order to confirm the cellular origin of the pneumococcal antigen-mediated cytokine production , we performed similar studies using intracellular cytokine staining in combination with staining for CD4 receptor . Flow cytometric analysis after 6 to 7 days stimulation and then an overnight re-stimulation with Ply revealed significant increases in CD4 T cell IFN-γ , TNF-α and IL-17 ( Figure 5b ) . To determine whether the anti-pneumococcal cytokine response is modified by Treg-mediated suppression , we then analysed cytokine production following the removal of the Treg containing CD25hi cell population from the tonsil MNC cultures pre pneumococcal challenge . Intracellular flow cytometric analysis revealed increases in IFN-γ and IL-17 positive cell numbers in the Treg depleted cell cultures . As expected , analysis of those few subjects not displaying increases in anti-pneumococcal CD4 proliferation post Treg depletion revealed little change in their IFN-γ and IL-17 expressing CD4 T cell numbers following CD25hi cell depletion ( data not shown ) . The results indicate an inhibitory effect of Tregs on the pneumococcal-induced production of these specific cytokines by mucosal CD4 T cells . No significant effect on TNF-α and IL-10 production was observed following the depletion of CD25hi cells . Initial transwell experiments indicated the involvement of cell contact-dependent mechanism/s for the observed Treg suppression of CD4 T cell proliferation ( data not shown ) . We therefore assessed whether Tregs utilised the inhibitory cell surface molecules CTLA-4 and PDL-1 [29] , [32]–[37] during their suppression of anti-pneumococcal responses by tonsil CD4+ T cells . This was achieved by pre-blocking purified CD25hi Tregs with neutralizing antibodies prior to addition back into CD25hi depleted cell population and subsequent 8 days in vitro culture in the presence of SPNT . Flow cytometric analysis revealed that approximately 80% of CD4+ CD25hi Tregs were successfully bound by the antibodies while up to 2% of non Tregs cells in the culture became bound by the neutralizing antibodies during the entire 8 day culture period ( data not shown ) indicating blocking was specific to the Treg cell population . Blocking with CTLA-4 antibody but not an isotype matched IgG1 antibody control resulted in a significant increase in the mean percentage CD4 proliferation from the undepleted culture ( Figure 6a ) . Just as depletion of CD25hi cells resulted in a significant increase ( p <0 . 05 ) in the mean percentage of CD4 T cell proliferation , pre-blocking with PDL-1 but not the isotype matched control IgG2 antibody on CD25hi cells and their subsequent restoration back into CD25hi depleted cell populations also resulted in significant increase ( p <0 . 05 ) in mean CD4 cell proliferation ( Figure 6b ) . Collectively , these data suggests that the suppression of mucosal anti-pneumococcal T cells by Treg involves surface interactions via CTLA-4 and PDL-1 inhibitory co-receptors . Blocking either CTLA-4 or PDL-1 did not increase proliferative responses up to the levels observed in post CD25+ cell depletion indicating additional mechanisms may also be involved for Treg suppression of anti-pneumococcal CD4 T cell responses . MHCII tetramers consisting of four identical biotinylated MHCII molecules presenting an epitope of a specific antigen , ligated to one another via fluorescently labelled strepavidin molecules have been increasingly utilised to allow the flow cytometric detection and quantification of CD4 T cells with specificity to the antigen being presented . The target CD4 T cells bind to the antigenic epitope/tetramer complex via their surface T cell receptors ( TCR ) thus allowing their subsequent detection [38] . We therefore utilised this technology in order to determine whether CD25+ Tregs which we observed to inhibit CD4 T cell proliferative responses to Ply can be detected within tonsillar MNC populations . Using Ply epitope presenting MHCII tetramers we confirmed the presence of S . pneumoniae specific Treg cells within the tonsil CD25+ cell population of adults by generating phycoerythrin ( PE ) -labelled HLADR04 tetramers bound to one of three different Ply peptide epitopes , or to a non-epitope peptide from Ply as a tetramer-Ply epitope negative control , which was observed to bind poorly to Ply specific cell lines we had generated ( data not shown ) . The HLADR0401 MHCII molecule was chosen as a large fraction of the Caucasian population expressed this particular serotype , thus increasing the chances of finding suitable HLADR0401 subjects for subsequent tetramer analysis . Tetramer binding was assessed on tonsil CD25 enriched cells by flow cytometry ( Figure 7 ) . Cells were stained with streptavidin-PE alone ( Figure 7a ) , the tetramer-Ply negative control ( Figure 7b ) , or three tetramer-Ply epitope cocktail ( Figure 7c ) . Approximately 45% of CD25 enriched CD4+ cells were found to be FoxP3+/CD127 low/- Treg cells . Within this population 0 . 01% of cells bound to streptavidin PE alone and 0 . 47% bound to the control tetramer . Importantly , 1 . 96% of the Treg bound to the Ply epitope tetramers . In five individuals over 17 years old assessed , 0 . 53 to 1 . 96% ( with % of cells binding to streptavidin-PE control subtracted ) of Treg cells were Ply-specific , which was significantly greater ( p <0 . 05 ) than the 0 . 005 to 0 . 46% ( with % of cells binding to streptavidin-PE control subtracted ) of Treg cells bound by the negative control Ply tetramer . Similar assessment of blood Treg population indicated lower frequencies of Ply specific Treg cells as indicated by a lack of tetramer staining ( mean 0 . 02% ) of CD4+ FoxP3+/CD127low/- peripheral blood mononuclear cells ( Figure 8 ) . This result would therefore indicate a similar sequestration of pneumococcal specific Treg cells within the tonsils as that observed for pneumococcal specific conventional CD4 T cells , and also implies that the low CD4 T cell responses in blood is not due to a high frequency of anti-pneumococcal Tregs residing within this compartment and suppressing immunity . Thus using these novel tools we have shown the presence of tonsil FoxP3+/CD127low/- , CD25 enriched CD4+ Treg cells that are specific for pneumococcal Ply antigen , likely contributing to the observed suppression of anti-pneumococcal CD4 T cell responses .
In the current study , the magnitude of CD4 T cell response to S . pneumoniae with age was assessed in view of the reported age-related decline in pneumococcal disease and the proposed involvement of cell mediated immunity to this acquisition of protection from disease . We hypothesised that assessment of CD4 T cells residing within upper respiratory tract mucosal lymphoid tissues that are in close proximity to the site of S . pneumoniae colonization may prove to be more appropriate for the study . Indeed , comparison of anti-pneumococcal responses by CD4 T cells from blood and palatine tonsils revealed a greater level of responsiveness by the cells obtained from the latter source . Although , it must be acknowledged that tonsils are only available from individuals who have either an upper airway obstruction or have suffered from recurrent tonsillitis , the results are highly suggestive of increased numbers of pneumococcal specific CD4+ T cells within lymphoid tissues in the vicinity of the nasopharynx , which we propose to be likely due to the sequestration of these cells within these organs . The containment of these cells within the upper respiratory tract would likely prove advantageous in promoting more efficient immune responses by increasing the likelihood for antigenic encounter and consequently facilitating more rapid , localised responses . Results from a previous study had indicated higher anti-pneumococcal responses by tonsil compared to blood CD4 T cells but failed to determine whether such a difference was significant [13] . Herein , we show that this is indeed the case . Assessment of the pneumococcal-specific responses by mucosal CD4 T cells with age revealed a gradual age-related increase in the magnitude of cellular proliferation from the youngest age assessed at 2 years old until the age of approximately 20 years . Cytokine analysis revealed Th1 and Th17 type anti-pneumococcal CD4 T cell responses were most evident , suggesting the recruitment and activation of macrophage and neutrophils as a primary inductive mechanism for protection by CD4 T cells from pneumococcal carriage and disease , as observed in mouse models [11] , [14] . We hypothesise that the progressive rise in CD4 T cell responses with age may be at least partially responsible for the observed decline in pneumococcal disease rates through childhood and early adulthood . Support for this proposal comes from epidemiological investigations into CAP . Epidemiological studies of CAP cases in Finland and earlier in the US found the incidence of the disease to show greatest decline during the first 14 years of life [39] , [40] . This was also supported by a more recent investigation in the UK , which reported incidence rates dropping from 20 cases/10000 inhabitants to 6 cases/10000 inhabitants per year ( Figure 2b ) . The incidence of CAP dropped more slowly by a further 15% between the ages of 15 to 29 years [21] . Thus , an inverse relationship between CAP and CD4 T cell responses can be observed , with the graph for the rates of CAP appearing to be almost a mirror image of the graph for the rate of CD4+ T cell proliferative responses to pneumococcus with increasing age . While this is suggestive of a role for mucosal CD4+ T cell responses in reducing CAP incidence , a role for other factors cannot be dismissed . Furthermore , whether the role of the CD4+ T cell response is mediated directly or via help for antibody production is unclear . Data from studies conducted by Laine et al on Kenyan subjects revealed increases in IgG and IgA to Ply as well as pneumococcal surface protein A ( PspA ) up to the age of 3 to 5 years which then plateau at least until 20 years [41] . These findings would indicate a contributory role for the antibody response in reducing pneumococcal disease early on in life but less so after 5 years of age when such responses plateau despite CAP levels still continuing to decline until teenage years . Thus , from the age of six years it is tempting to speculate that the CD4 T cell response is perhaps of greater relevance as antibody responses are seemingly static despite disease rates continuing to drop . Additional analysis of antibody responses , and in particular detailed characterisation of the levels and functional properties of both systemic and mucosal responses would be helpful in understanding the relative roles . Murine and studies in human childhood have shown an important role of CD4 T cells , particularly Th17 cells in inhibiting pneumococcal carriage lending support to a correlation between the cell-mediated response to pneumococcus and carriage rates of the bacteria [12]–[14] . Despite decreasing disease and carriage rates , CD4 T cell responses were observed to increase with age , we propose that multiple exposure events to S . pneumoniae takes place throughout life in order to maintain and , for the time period assessed in this study progressively augment the anti-pneumococcal CD4 T cell responses with age . Mucosal T cell immune responses are commonly susceptible to the suppressive actions of Treg cells , and herein for the first time , we observed this to be also the case with adult human anti-pneumococcal CD4 T cell responses . Interestingly a significant Treg effect was observed during the ages when relatively strong anti-pneumococcal CD4 T cell responses have developed . Such regulation may be useful at this stage of life when these more robust responses may have a greater likelihood to cause bystander damage to host tissues following an immune response to pneumococci . Furthermore , strong responses to colonizing pneumococcus may lead to the damage of mucosal cell walls and permit bacterial penetration into tissues resulting in infection and disease . Indeed , recent studies in mice have observed the ability of gut commensal bacteria to induce the generation of Treg cells within the mucosal tissue in order to inhibit inflammatory responses that can cause immunopathology and lead to autoimmune disease [42] , [43] . With the use of novel pneumococcal specific tetramers bound to epitopes of Ply , we were able to confirm the presence of Treg cells specific for pneumococcal Ply in tonsillar populations . It is therefore likely that these cells , as well as Treg which may recognise and respond to other pneumococcal antigens , are involved in the observed suppression of CD4 T cell responses to pneumococcus . Within mucosal lymphoid tissue we observed that at least 1 per approximately 200 FoxP3+/CD127low/- CD4+ Treg cells in the tonsil were specific for Ply . This high incidence may reflect a level of specific induction at , or retention of anti-pneumococcal Treg cells within this site . Importantly the frequency of such cells in the peripheral blood was much lower . In future studies the novel Ply presenting tetramers that we have generated would be invaluable tools in enabling a comparison of the anti-Ply CD4 T cells in each compartment . The reason for the delay in the development of Treg responses with age is unclear . We hypothesise that this may be a consequence of the types of encounter that take place between the host and the pneumococcus . It is possible that , during childhood , the greater propensity for colonisation events to become associated with infection favours the development of responses lacking T cell regulation . However , as immune responses increase with age , infection becomes rarer and the balance shifts toward carriage without invasion , which might subsequently predispose to the development of regulation . Alternatively , this observation may be due to differing carriage rates of the bacteria during aging . Support for this comes from our previous study assessing the CD4 T cell response to N . meningitidis ( MenB ) which like that of S . pneumoniae also became subject to Treg cell suppression [20] . Unlike in S . pneumoniae however , Treg suppression of MenB responses became evident earlier on in life between 8–11 years of age . While carriage rates are highest for S . pneumoniae during approximately the first decade of life and dropping to just below 10% by mid-teens onwards , for MenB carriage does not reach peak levels until late teens with pre-teens prevalence rates remaining below 10% [44] , [45] . Thus , with regulation for the two bacteria appearing at ages when carriage rates for both are also similar ( i . e . at just below 10% ) , it is tempting to speculate for the role of carriage in the development of a regulatory mechanism . A recent study by Zhang et al suggested that the adenoids of young children who were colonised , but not those who were not colonised with S . pneumoniae , contained Treg that could suppress anti-pneumococcal T cell responses [46] . Although they were not able to confirm that the Treg that they studied were pneumococcus specific as opposed to being activated polyclonally to factors such as TLR ligands in the preparations or identify the mechanisms utilised by the Tregs for suppression , their findings do support a possible role for colonization in the development of regulation . We did not distinguish between carriers and non-carriers in our studies and accordingly did not see evidence of Treg controlling reactivity in such young children . While Zhang et al's findings suggests a direct correlation between colonisation and Treg suppression of anti-pneumococcal CD4 T cells in infants , the presence of Tregs despite the observed low levels of colonisation in adults [12] , [14] would indicate that pneumococcal-specific Treg become a stable part of the repertoire with age . Transient Treg responses in children may promote colonisation and therefore be undesirable . In adults , where colonisation is less frequent , they may have a different role altogether; to dampen potentially pathological immune responses during pneumococcal exposure and promote a beneficial profile of immunity at the mucosa . Although the case for the majority , not all subjects above the age of 16 years displayed Treg inhibition of their anti-pneumococcal CD4 T cell immune responses , with some even showing decreases in responses following depletion of CD25hi cells . Such variation is almost inevitable in complex human systems , and may possibly be due to differences in bacterial carriage and/or differences in the number of S . pneumoniae exposures throughout life , as well as other factors affecting the state of the local immune system at the time of the study . For example , in individuals with an ongoing mucosal response effector cells as well as Treg cells may be depleted using CD25 , and this could deplete the proliferating pool as well as those potentially capable of controlling proliferation . The Treg cells observed in our study mediated suppression using the inhibitory cell surface molecules CTLA-4 and PDL-1 in this process . CTLA-4 on Treg cells may inhibit pneumococcal specific responses by preventing and downregulating cell co-stimulation via B7 . 1/B7 . 2 and CD28 interactions between antigen presenting cells and effector T cells respectively [33] , [47] . Treg cells may utilise PDL-1 in their suppression by engaging programmed death 1 ( PD1 ) and/or B7 . 1 on effector T cells to attenuate T cell receptor ( TCR ) signalling , partly by upregulating the T cell inhibitory basic leucine transcription factor ( BATF ) in order to suppress T cell activation and cytokine production [48] , [49] . It is interesting to note that both CTLA-4 and PDL-1 can bind B7 . 1 in light of a previous study observing the capacity for Treg cells to induce their inhibition through engagement of B7 molecules expressed on their target T cells [50] . The role of the antibody response in host defence against pneumococcal disease has been extensively documented , and it is therefore interesting to consider whether the Treg characterised here may affect these responses directly or indirectly via their control of effector T cells . Interestingly , studies in mice and human have indicated a capacity of Tregs to affect antibody production both in vitro and in mice in vivo . In mice increases in IgG and IgA in mucosal tissues and mucosal secretions have been observed as well as decreases in splenic IgG and IgM in the presence of Tregs [51] , [52] . While in humans , Tregs were observed to inhibit IgG , IgM , IgE and IgA production by tonsillar B cells following their polyclonal activation [53] . Thus a possible additional effect of the pneumococcal specific Tregs that we have observed in this study on B cell responses to S . pneumoniae along with the CD4 T cell response merits assessment . Several mouse studies have shown the importance of the Th17 CD4 response in inducing neutrophil and monocytes/macrophage mediated clearance of colonizing bacteria . Studies by Lu et al and Mureithi et al have observed in vitro IL-17 production following challenge with pneumococcal antigens with Lu et al showing an enhancing effect of IL-17 on the phagocytic killing of pneumococci by human neutrophil cells [14] , [15] . Our assessment of cytokine production by adult mucosal CD4 T cells post-pneumococcal antigen challenge also revealed increases in IL-17 levels . Such Th17 responses may limit colonization in this age group and since colonization is a prerequisite of disease , may consequently contribute to the lower disease rates observed in human adults [4] , [45] . However , our data additionally revealed that Tregs were able to significantly decrease the observed IL-17 production by anti-pneumococcal CD4 T cells , which have not yet been shown previously . The ability of Treg cells to preferentially block not only IL-17 but also IFN-γ pro-inflammatory cytokine production is interesting as this may reveal a role for these cells in blocking unwanted inflammation induced pathology at the mucosa upon contact with colonizing S . pneumoniae . Coincidently , neutrophils and macrophages , the main protagonist cell populations involved in protecting against pneumococcal carriage and/or disease are found to be principally mediated by these two particular cytokines [14] , [15] , thus providing a possible motive for the preferential effect of Tregs on their production . Other previous studies have shown the capacity for Tregs to inhibit Th17 cell responses as we have observed herein [54] , [55] . It is possible that the Treg responses observed serve to moderate potentially pathological immunity at the mucosa in order to maintain a balance between the need to contend with a potentially harmful pathogen and to preserve physiological and barrier function . Whilst regulation may be evolutionary advantageous in the context of commensalism , the potential to suppress protective immunity during pneumococcal invasion may conversely facilitate disease . Our findings have clear implications for the development of vaccines against the pneumococcus . As we move from approaches targeted solely at stimulating antibody responses to polysaccharide antigens and toward the generation of protein based pneumococcal vaccines , it will be important to consider the extent to which new approaches mimic natural immunity in providing local protection at the mucosa . Further , it will be important to recognise that vaccines that drive potent Th1 and/or Th17 responses without inducing the balancing effects of Treg could potentially lead to enhanced pathological outcomes in disease , and even during colonisation . The challenge in vaccinating adults will be to enhance the pre-existing responses that we have described as immunity wanes in the elderly , while maintaining the fine balances that mediate protection . As may be the case in other infectious diseases , vaccines should not necessarily be produced with the sole aim of stimulating as strong an immune response as is possible , but should rather be targeted at modulating immunity to achieve the desired outcome .
Palatine tonsils and blood samples were obtained from otherwise healthy individuals ( aged 2 to 39 years ) undergoing routine tonsillectomy for recurrent tonsillitis or upper airway obstruction at Bristol Royal Hospital for Children , Southmead Hospital or Saint Michael's Hospital in Bristol , United Kingdom . Tonsils were collected into HBSS media ( Invitrogen ) supplemented with 100 U/ml penicillin , 100 µg/ml streptomycin ( Sigma ) and blood samples into citrate phosphate dextrose solution ( Sigma ) . Patients with immunodeficiency or serious infections were excluded from the study . Participants with inflamed tonsils at the time of surgery were not recruited . No participants had received pneumococcal vaccine . The study was approved by , and sample collection and research were undertaken in accordance with the guidelines set out by the South Bristol local research ethics committee ( reference number; E4388 ) . Written informed consent was obtained from all participants and/or their legal guardians . Pneumococcal cell culture supernatants ( SPNT ) from a standard encapsulated type 2 ( D39 ) S . pneumoniae strain ( National Collection of Type Cultures , NCTC #7466 ) was prepared as described previously [56] and used at a concentration of 2 µg/ml for cell stimulations; the optimal level as determined in dose response studies ( data not shown ) . Recombinant Ply , a Ply protein with a Trp433-Phe mutation that reduces its haemolytic activity without affecting antigenicity generated as previously described [19] , was added at a concentration of 0 . 1 µg/ml for cell stimulations . Pneumococcal antigens were tested for the presence of contaminating Gram-negative endotoxin using the colorimetric LAL assay ( KQCL-BioWhittaker , Lonza ) . All purified proteins had endotoxin levels at concentrations that were too low to have any notable effect on CD4 T cell responses ( <0 . 6 units per microgram of protein ) . Fluzone 2002–2003 formula inactivated split-virion influenza ( flu ) vaccine ( Sanofi-Pasteur MSD ) was used at 0 . 09 µg/ml hemagglutinin . Blood and tonsil tissue ( MNC ) were isolated by histopaque density gradient separation as described previously [20] . CD25hi cells were depleted from ∼0 . 5–1×108 MNC using anti-human CD25 coated MACS microbeads ( Miltenyi Biotec ) and magnetic cell sorting ( MACS ) on LD columns ( Miltenyi Biotec ) according to the manufacturer's instructions . Purity of depleted cells were typically >96% as assessed by flow cytometry using APC labelled anti-human CD25 antibody ( BD Pharmingen ) . Approximately 8–10% of tonsil lymphocytes were found to be CD25hi . Cells were resuspended at 0 . 8×106/ml in RPMI 1640 media containing 2% human serum ( Sigma ) , 2 mmol L-glutamine ( Sigma ) , 100 U/ml penicillin ( Sigma ) and 100 µg/ml streptomycin ( Sigma ) . Cells were plated at 1 ml/well in 48 well plates ( Corning ) and stimulated with flu , pneumococcal supernatant or Todd Hewitt Broth ( negative control ) , Ply or left unstimulated as negative control for up to 9 days , at 37°C/5%CO2 . In some stimulation experiments the cell culture supernatants were collected 5 days post stimulation and stored at −70°C for subsequent cytokine assessment . Cells were subsequently analysed for cell proliferation . For the CD25hi add back experiments , MACS purified CD25hi cells were added to 0 . 8×106 CD25hi depleted cells per ml media at the original ( 8×104 i . e . 10% ) or 3 times the original ( i . e . 2 . 4×105 ) CD25hi cell proportion in undepleted MNC . Prior to stimulation , cells were labelled with CFSE dye ( Invitrogen ) according to the manufacturer's instructions , in order to permit flow cytometric tracking of cell division . Cells were stained with anti-human CD4- phycoerythrin ( PE ) -Cy7 antibody ( BD Pharmingen ) for 30 mins at 4°C and then with the vital dye TOPRO3 ( Invitrogen ) , according to the manufacturer's instructions just prior to flow cytometric ( FACS ) analysis . Cells were analysed using FACS Canto ( Becton Dickinson ) acquiring 20000 lymphocytes ( gated according to forward and side scatters ) . FACS results were subsequently analysed with FlowJo ( TreeStar ) . The identification of cells that have undergone cellular division and the gating of FACS dot plots according to CFSE staining was performed as described previously [13] and consequently background ( i . e . media alone for Ply stimulated or Todd-Hewitt broth treated for SPNT stimulated ) proliferation were subtracted from values of stimulated cells . Dead cells ( i . e . cells stained with TOPRO3 ) were excluded from the flow cytometric analysis . Our studies have shown that the data obtained with this technique , which allows identification of the dividing cells , directly correlates with that obtained from 3H-thymidine incorporation experiments from matched samples ( data not shown ) . IL- 2 , 5 , 10 , 17 , TNF-α and IFN-γ levels in the cell culture supernatants were quantified 7 days post stimulation with Ply by Luminex xMAP technology for cytokine quantification and Luminex 200 exponent system ( Luminex ) according to the manufacturer's instructions . Additionally IL-10 , -17 , TNF-α and IFN-γ production by CD4+ T cells was measured by intracellular cytokine flow cytometric analysis . Briefly , at day 6 to 7 of culture , cells were restimulated for 12 hours with Ply antigen and BD Golgi Stop protein transport inhibitor ( BD Bioscience ) was added for the last 8 hours . Cells were stained with Live/Dead near IR Stain kit ( Invitrogen ) according to the manufacturer's instructions , in order to exclude dead cells from flow cytometric analysis and then with anti-human CD4-V450 ( BD Pharmingen ) . Cells were fixed and permeabilized using BD cytofix/cytoperm kit ( BD Biosciences ) according to manufacturer's instructions and were subsequently stained with antibodies to TNFα-alexa flour 700 ( BD Pharmingen ) and IL-10-APC ( BD Pharmingen ) or IL-17- alexa flour 700 ( BD Pharmingen ) and IFNγ-APC ( BD Pharmingen ) at 4°C for 30 mins and analysed with the BD LSR flow cytometer ( Becton Dickinson ) . The percentage of cells positively staining for cytokines in unstimulated samples were subtracted from the percentage of positively cytokine stained cells in the stimulated samples . For pre-blocking experiment , 5×106 CD25hi CFSE stained cells obtained during CD25hi depletion by magnetic cell sorting on LD columns as mentioned above in 1 ml RPMI 1640 media containing 1% human serum , 2 mmol L-glutamine , 100 U/ml penicillin and 100 µg/ml streptomycin were treated with 5 µg/ml anti-human CTLA-4 ( eBoscience ) or 15 µg/ml anti-human PDL-1 ( eBoscience ) or appropriate isotype-matched antibody controls ( eBoscience ) and incubated for 3 hours at 37°C/5%CO2 . Cells were extensively washed and added back to CFSE stained CD25hi depleted cell fraction at a ratio of 1∶10 . Flow cytometric analysis revealed >80% of the CD25hi cells to be successfully blocked . Cells were stimulated with antigens and left for up to 8 days in culture at 37°C/5%CO2 and then analysed for cellular proliferation by CFSE staining . Biotinylated HLADR0401 MHCII tetramers bound to streptavidin-PE were generated at the Benaroya Research Institute ( Seattle , USA ) . Four different Ply 13mer peptide epitopes ( P1; SerAspIleSerValThrAlaThrAsnAspSerArgLeu , P2; ArgProLeuValTyrIleSerSerValAlaTyrGlyArg , P3;ValTyrLeuLysLeuGluThrThrSerLysSerAspGlu , P4; ThrSerPheLeuArgAspAsnValValAlaThrPheGln ) were identified from the complete Ply protein amino acid sequence by using TEPITOPE ( Vaccinome ) [57] . Each Ply epitope was tested for its capacity to induce proliferation of Ply-specific T cell lines in order to determine the potential applicability of the epitopes for detecting CD4 T cells . Tetramers were generated by loading the Ply peptides onto empty biotinylated DR0401 molecules and subsequent cross-linking with streptavidin-PE [58] . To test whether each tetramer were able to bind Ply specific tonsil CD4 T cells , anti-Ply T cell lines were generated from blood CD4+ T cells of HLA DRB1*04 expressing individuals . Briefly , fresh PBMC were depleted of CD8+ cells by magnetic cell sorting using anti-CD8 magnetic beads ( Miltenyi ) according to the manufacturer's guide and stimulated with 0 . 2 µg/ml Ply peptide in RPMI 1640 medium containing 10% human serum . The plates were incubated at 37°C in 5% CO2 , and after 7 days , the medium was replaced and 25 U/ml IL-2 ( Peprotech ) added . Cells were restimulated with peptide , IL-2 and irradiated autologous PBMC every 12–14 days . Cell lines in 1% human serum RPMI were stained with a tetramer presenting one of the four Ply epitopes at 10 µg/ml for 1 hour at 37°C , 5% CO2 . Subsequent FACS analysis revealed tetramers presenting P1 , P3 and P4 Ply epitopes were bound by at least 6% of cells while P2 were bound by <1% and used as a negative control . CD4+CD25 enriched tonsillar MNC or PBMC from HLADR0401 expressing subjects above 20 years old were obtained by magnetic sorting using the CD4 CD25 enrichment kit ( Miltenyi Biotec ) according to the manufacturers instruction ( typically >80% of cells were CD4+ CD25+ ) . Enriched cells were stained with either streptavidin-PE ( Becton Dickenson ) as a negative control , P1 , P3 and P4 tetramers at 10 µg/ml each or 30 µg/ml of P2 tetramer as an additional negative control as previously . Cells were stained with anti-human CD4-PECy7 ( Becton Dickenson ) , anti-human CD127-FITC ( eBioscience ) according to manufacturer's instruction followed by intracellular staining with anti-human FoxP3-APC ( eBioscience ) using FoxP3 staining buffer set ( eBioscience ) according to the manufacturer instructions and subsequently analysed by FACS and FlowJo . Distribution of the data was determined using the Kolmogorov-Smirnov test . When data was found to be normally distributed , differences between two groups were tested using paired or unpaired students t- test accordingly . The Wilcoxon signed rank test was used to test differences between undepleted MNC and CD25 depleted groups following antigenic stimulation by SPSS statistical analysis software ( IBM ) . Two way ANOVA was used for comparing results of untreated , flu , Ply and SPNT treated groups and for the effect of Treg study and for comparing undepleted MNC , CD25-depleted , IgG blocked and CTLA-4 or PDL-1 blocked groups in the CTLA-4 or PDL-1 blocking study by SPSS statistical analysis software . Amino acid sequence for pneumolysin was obtained from GenBank; Accession:ADF28490 ( DBSOURCE - GU968411 . 1 ) . GI: 294652455 .
|
The S . pneumoniae bacteria is a major cause of disease ( e . g . pneumonia and meningitis ) particularly affecting infants . In most cases bacteria can colonise the nose without causing harm , however colonisation is thought to be a prerequisite of disease . With increasing age colonization and disease , rates gradually decrease which is likely due to the development of immunity to the pneumococcus with age . The CD4 T cells of the immune system may contribute to the defence against bacterial colonisation by producing factors that promote pneumococcal killing . Herein , we show that CD4 T cells reactive to pneumococci are found in greater numbers at the site of colonisation and gradually increase in their levels from infancy . However , at the peak of CD4 T cell responses from late teens , we detected the presence of regulatory T cells ( Tregs ) which suppressed anti-pneumococci CD4 T cell activity greatly . Our finding shows that pneumococcal reactive CD4 T cells selectively populate colonisation sites and increase with age as a result of ongoing bacterial exposure throughout life , inversely correlating with colonisation and disease rates . As factors that utilise CD4 T cells become increasingly advocated as potential preventative strategies against pneumococcal carriage and disease , the observed effect of Tregs must be considered .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"adaptive",
"immunity",
"immune",
"cells",
"aging",
"and",
"immunity",
"streptococci",
"immunity",
"t",
"cells",
"immunology",
"biology",
"microbiology",
"immunoregulation",
"bacterial",
"pathogens",
"immune",
"response"
] |
2011
|
Acquisition of Pneumococci Specific Effector and Regulatory Cd4+ T Cells Localising within Human Upper Respiratory-Tract Mucosal Lymphoid Tissue
|
De novo creation of protein coding genes involves the formation of short ORFs from noncoding regions; some of these ORFs might then become fixed in the population . These orphan proteins need to , at the bare minimum , not cause serious harm to the organism , meaning that they should for instance not aggregate . Therefore , although the creation of short ORFs could be truly random , the fixation should be subjected to some selective pressure . The selective forces acting on orphan proteins have been elusive , and contradictory results have been reported . In Drosophila young proteins are more disordered than ancient ones , while the opposite trend is present in yeast . To the best of our knowledge no valid explanation for this difference has been proposed . To solve this riddle we studied structural properties and age of proteins in 187 eukaryotic organisms . We find that , with the exception of length , there are only small differences in the properties between proteins of different ages . However , when we take the GC content into account we noted that it could explain the opposite trends observed for orphans in yeast ( low GC ) and Drosophila ( high GC ) . GC content is correlated with codons coding for disorder promoting amino acids . This leads us to propose that intrinsic disorder is not a strong determining factor for fixation of orphan proteins . Instead these proteins largely resemble random proteins given a particular GC level . During evolution the properties of a protein change faster than the GC level causing the relationship between disorder and GC to gradually weaken .
Proteins without any detectable homology are often referred to as orphans . The presence of orphans can be attributed to several causes; rapid sequence divergence beyond the point of homology recognition [1 , 2] , lateral transfer of genetic material [3] , and de novo gene creation [4] . The latter is of particular interest , as it is a source of completely novel coding material . Studies of the properties of these proteins might provide unique insights into the fundamental processes in the formation of all proteins , since , in the strict sense , all proteins were once created by a de novo mechanism . Before the genomic era , the scientific consensus held that de novo creation of new genes was rare—instead it was believed that the vast majority of all genes were generated in an ancient “big bang” . However , when the first complete genomic sequences were initially published , this hypothesis was not supported [5] . In fact , to this day , when analyzing complete genomes from closely related species , a surprisingly high number of orphan proteins is still found [6–8] . It has later been shown that some of these proteins are not de novo created but rather assigned as orphans as a result of limited phylogenetic coverage in earlier studies [9] . Today supported by the vast amount of complete genome sequences available and improved search methods [10] , many of the initially identified orphans have been shown to have distant homologs in other genomes . Still , at least in yeast , a large set of genes appears to have been created through recent de novo formation [11 , 12] . These studies indicate that in yeast there is a large set of proto-genes: ORFs that remain on the verge of becoming fixed as protein-coding genes in the population [11] . This provides a possible model of how novel proteins can be generated from noncoding genetic material . In other species than yeast the genomic coverage is more limited and therefore studies have been less detailed . The availability of many , complete , evenly spaced genomes allows classifying proteins at different evolutionary age [7 , 9 , 11] , using methods such as ProteinHistorian [13] . Here , a protein can be unique to a specific species , or even to a strain; alternatively it can be present pervasively across a taxonomic group [14 , 15] . After de novo creation , a gene needs to become fixed in the population . The selective forces governing this process have been studied by examining the properties of orphan proteins . Intrinsic disorder , low complexity , subtelomeric location , high β-sheet preference as well as other features have been associated with orphan proteins [16 , 17] . It has also been proposed that with age proteins ( i ) accumulate interactions , ( ii ) become more often essential and ( iii ) obtain lower β-strand content and higher stability [18] . Some aspects of these observations , such as the fact that orphans on average are short , are likely connected to a de novo creation mechanism . However , other features , including intrinsic disorder , are not obviously related to the gene genesis and could instead be the result of the selective pressure acting during fixation . In yeast , we have earlier reported that the youngest proteins , i . e . the ones unique to S . cerevisiae , are less disordered than older proteins [7] , while in Drosophila the opposite can be seen: the youngest proteins are more disordered than the ancient ones [19] . To the best of our knowledge the origin of this difference has not been explained . Could the selective forces be that disparate between two different eukaryotes ? Or is it an artifact caused by limited genomic coverage ? One difference between the two organisms is the content of Guanine and Cytosine ( GC ) nucleotides in the coding regions: Saccharomyces cerevisiae genomes are roughly 40% GC , while in Drosophila melanogaster the GC content is 53% . To obtain a better understanding of the structural properties of orphan proteins , we determined the age of proteins in 187 eukaryotic genomes and studied a number of intrinsic properties , such as GC content , disorder , hydrophobicity , low complexity , and secondary structure . As expected we find that the most striking difference between young and old proteins is their difference in length . Further , intrinsic disorder and low complexity appear to be higher in orphans , albeit with a much smaller difference than for length , and these differences are not present in all species . The structural features in young proteins differ significantly depending on the GC content: low-GC orphans are much less disordered than high-GC orphans . In older proteins this relationship is much weaker , supporting a model where genes are created de novo starting from random DNA sequences , then their features gradually conform to those of ancient genes through adaptation .
Protein data for 400 eukaryotic species were obtained from OrthoDB , release 8 [20] , divided into 173 Metazoans and 227 Fungi , for a total of 4 , 562 , 743 protein sequences . This initial dataset was then filtered to a final size of 187 species , see below . For each species , a complete proteome was also downloaded from UniProt Knowledge Base [21] . The ProteinHistorian software pipeline [13] is aimed at annotating proteins with phylogenetic ages . It requires a phylogenetic tree relating a group of species , and a protein family file representing the orthology relationships between proteins . The pipeline assigns each protein to an age group , depending on the species tree . Here , we used ProteinHistorian with default parameters , the NCBI phylogenetic tree [22] , and protein orthology data obtained from OrthoDB . The OrthoDB method is based on all-against-all protein sequence comparisons using the Smith-Waterman algorithm and requiring a sequence alignment overlap of at least 30 amino acids across all members of an orthologous group . Therefore , the age group can be thought of as the level in the species tree on which a shared sequence of at least 30 amino acids first appeared , i . e . it assigns multi-domain proteins to the age of its oldest domains . Proteins present in OrthoDB are only those with orthologs in at least one other species , i . e . proteins without orthologs ( singletons ) are not present in OrthoDB . Therefore , to obtain a set of candidate orphan proteins , the complete proteomes of all species were downloaded from Uniprot . Thereafter , BLAST was used to extract proteins not present in the OrthoDB dataset , obtaining 356 , 884 candidate orphan proteins . However , a large fraction of these proteins are not orphans but are missing from OrthoDB for other reasons , including that they were not present when the database was created or that they have undergone large domain rearrangements . We would assume that truly de novo created orphans do not contain domains found in other proteins . Therefore to ensure that we have a unique set of orphan proteins we filtered out proteins with hits in the Pfam-A database , by using hmmscan [23] . We believe that , due to the stringent criteria used here , the majority of this remaining set is constituted of de novo created proteins , and we refer to them as orphans throughout the rest of this paper . These proteins are specific to the species taxonomic level , i . e . we expect not to find them in other species in the dataset , even in the same genus . For Saccharomyces cerevisiae , that has 16 strains in the dataset , we also included the strain specific proteins into the orphan group . Among the OrthoDB proteins , we defined genus orphans those that were assigned age = 1 ( 2 in the case of S . cerevisiae ) , while proteins having the maximum age according to ProteinHistorian were defined as ancient: these proteins are thought to be present in the common ancestor of all Fungi ( taxon id = 4751 ) or all Metazoa ( taxon id = 33208 ) . Finally , proteins whose estimated age is between genus orphans and ancient were defined as intermediate . Taxonomic genera represented by a single species in the dataset have by definition no genus orphans; for this reason , we selected for our final dataset only the 187 species that have at least one other species within the same genus . The final dataset amounts to 1 , 782 , 675 proteins distributed across 187 species; 0 . 8% of them are defined orphans and 0 . 6% as genus orphans , 15% are intermediate and the remaining 84% are ancient . One problem that exists using the NCBI phylogenetic tree is the presence of many polytomic branches , especially at the genus level , because ProteinHistorian cannot distinguish between proteins being specific to that species and proteins shared among the entire group . To solve this , we forced no polytomy on the terminal branches: multifurcating nodes were converted to a randomly bifurcated topology , transforming the NCBI tree to a fully binary structure . While a binary tree is needed for ProteinHistorian , its algorithm assumes that a protein gain is much more rare than a loss; this means that the most recent common ancestor of a protein will be at the topmost intersection of a group of species . Thus , randomly converting multifurcations to bifurcation might likely underestimate the number of genus-specific orphans , but have no effect on species-specific orphans . Clades affected by the conversion from multi- to bi-furcating branches include Caenorhabditis ( 5 species ) , Drosophila ( 5 species ) , Anopheles ( 5 species ) , Candida ( 5 species ) , Saccharomyces ( 14 strains ) , Aspergillus ( 5 species ) and Trichopython ( 5 species ) . The taxonomic tree comprising the final set of 187 species is presented in S1 Fig . We could map 1 , 357 , 518 out of 1 , 782 , 675 proteins ( ∼76% of the dataset ) to their ENA identifiers . This mapping was used to download the Coding Sequence ( CDS ) data from ENA ( https://www . ebi . ac . uk/ena/ ) ; the GC content was then calculated for each mapped gene individually . Evidence for functionality of the proteins was estimated using annotated Gene Ontology ( GO ) terms . Using the Uniprot KnowledgeBase mapping data ( ftp://ftp . uniprot . org/pub/databases/uniprot/current_release/knowledgebase/idmapping/idmapping . dat . gz ) we assigned UniprotKB identifiers to 894 , 831 out of 1 , 752 , 675 proteins ( 51% ) . These were then annotated with three terms , one for each main GO category: Molecular Function , Biological Process and Cellular Component . All GO terms are associated with evidence codes; a subset of these codes ( ‘EXP’ , ‘IDA’ , ‘IPI’ , ‘IMP’ , ‘IGI’ or ‘IEP’ ) represents experimentally validated functional annotations . If any of these codes is present we mark the corresponding protein as experimentally characterized . Intrinsic disorder content was predicted for all the proteins by using several disorder predictors; short and long disorder predictions by IUPred [24] , three type of predictions ( REM-465 , Hotloops and Coils ) by DisEMBL [25] and GlobPlot [26] . In the main figures we only report the prediction by IUPred long; the others are found in the supplementary material ( S2 to S7 Figs ) . It is worth mentioning that these predictors operate with different definitions of disorder , so a consensus should not be expected . We used SCAMPI [27] to predict the fraction of transmembrane residues in a protein . The fraction of low-complexity residues is predicted using SEG [28] . PSIPRED [29] was used to predict the secondary structure of all the proteins in the dataset , using only a single sequence and not a profile . This reduces the accuracy but the overall frequencies should not be changed significantly . We annotated each protein with the fraction of residues predicted to be in each type of secondary structure ( α-helix , β-strand , coil ) . TOP-IDP [30] is a measure of the disorder-promoting propensity of a single amino acid . For each protein , the average propensity was calculated by averaging the TOP-IDP values of all its residues . Similarly the hydrophobicity of each protein was expressed as the average hydrophobicity using the biological hydrophobicity scale [31] . Finally , we computed the propensity of each amino acid to be in a secondary structure ( helix , sheet , coil , turn ) in the same manner by using secondary structure propensity scales [32] . In order to test the statistical significance of the results , a number of tests were performed . Rank-sum tests between all possible pairs of age groups were performed for the entire dataset and for each studied property . Due to the large number of samples the p-values from these tests are always smaller than 10−141 even when the absolute difference in numbers is minuscule . To study the difference between young and old proteins on a global level , we performed a rank-sum test for orphan versus ancient proteins within each species . To exclude small variations we only considered the species where the p-value of this test was <0 . 01 . To determine the relationship between a property and GC we studied the slopes for proteins of different age . If the p-value of a linear regression test is <0 . 01 , the corresponding property is considered significantly correlated with GC . To test whether the studied intrinsic properties , as well as the frequency of any given amino acid , were solely dependent on GC content , we used a set of 21 , 000 random ORFs , generated as follows: at each GC content ranging from 20 to 90% , in steps of 1% , a set of 400 ORFs ( equally divided into 300 , 900 , 1 , 500 and 2 , 100 bp long ) was generated so that its content of GC was fixed . The ORFs were generated by randomly selecting codons among the 61 non-stop codons . The probability to select one codon given a GC content of GCfreq is set accordingly: P r o b a b i l i t y = ∏ i = 1 3 δ ( N i | G C ) * G C f r e q + δ ( N i | A T ) * ( 1 - G C f r e q ) ( 1 ) where Ni is the nucleotide of the codon in position i and δ ( N|GC is equal to 1 if the nucleotide N is guanine or cytosine and zero otherwise , etc . Finally , start and stop codons were added . These ORFs were then translated to polypeptides , and all their intrinsic properties , as well as the frequencies of their amino acid were computed , as described above .
Next we set out to estimate the functional evidence for our set of proteins; for this we explore their Gene Ontology ( GO ) annotation . For each main GO category ( process , function and component ) , we computed the fraction of proteins being annotated with at least one GO term in UniProt . In addition we calculated the fraction of proteins having at least one experimentally verified GO annotation , Table 1 . The fraction of annotated proteins increases steadily with age , from ∼3-9% in orphans to ∼25% in ancient , Table 1 . This is expected , as older proteins have more regulatory , protein-protein , and genetic interactions [18] . However , the fraction of proteins with experimental functional evidence is small ( <1% of protein ) irrespectively of age . This shows that there exists at least a fraction of proteins of any age that is functionally characterized , but it is difficult to exactly determine how substantial it is . The average length of the proteins increases by age , see Fig 1b . The average length is 100 amino acids in orphans , 150 in genus orphans , 300 for intermediate and 500 for ancient proteins . It can be noted that in the vast majority of the genomes the difference in length is significant between orphan and ancient proteins , Table 2 . This highlights the well-established fact that eukaryotic proteins expand during evolution: the expansion can occur by several mechanisms , including domain-fusions [34] , additional secondary structure elements [35] and expansion within intrinsically disordered regions [16] . GC content on the other hand does not appear to change by age , see Fig 1c . There is approximately the same number of genomes where a statistically significant increase or decrease exist , Table 2 . Next , we compared predicted structural properties of all proteins , see Figs 1d–1i , S8 and Table 2 . The amount of predicted disorder residues ranges between 20% and 40% , depending on the prediction method . For most disorder predictors the fraction of disordered residues is higher in orphans than in ancient proteins . However , there exists about a handful of genomes where the opposite trend is observed: supporting earlier observations , orphans are significantly more ordered in Candida albicans according to 5 out of 6 methods , in Saccharomyces cerevisiae s288c for 4 methods and in Fusarium pseudograminearum for 3 . An interesting case is that of Drosophila pseudoobscura , that appears to have more ordered orphans according to IUPred long , contrary to all others Drosphila species . The fraction of transmembrane residues is on average ∼2% in orphan proteins , with an increasing trend towards ancient ( 4% ) . Similarly the amount of helical residues increase slightly with age , while the fraction of low complexity residues decrease by age . For all these structural predictions the changes are quite small and there are genomes with significant increases and decreases for all measures . Above , we noted that on average orphan proteins are more disordered . However , we also noted that in a handful of genomes a statistically significant opposite trend could be observed . To investigate this further we studied the amount of predicted disorder in each genome separately . When studying intrinsic disorder , orphans and genus orphans of S . cerevisiae appear remarkably ordered ( ∼3% of the amino acids ) as shown before [7] see Fig 2a and 2b . The closely related species Candida albicans shows a similar trend; see Fig 2c . Results from additional disorder predictors are presented in S4–S8 Figs and agree well with these observations . In contrast , but also consistent with earlier studies [36] , orphans and genus orphans in most Drosophila genomes are more disordered than ancient proteins , see Fig 2d and 2e . In the worm C . elegans ( Fig 2f ) orphan proteins appear to be consistently more disordered than progressively older ones; this is true across all the considered Caenorhabditis species . These results are consistent in other predictors , see S4–S8 Figs d , e and f . In general , it is apparent that in most organisms orphans are more disordered than ancient proteins , while in yeast the opposite appears to be the case . What could possibly explain this difference ? One possibility is that the more complex regulations in animals require more disordered residues in comparison with yeast . But the average disorder content is similar in all eukaryotic species , contradicting this idea . We noted that yeasts are among the genomes with lowest GC content ( ∼40% in S . cerevisiae , 35% in C . glabrata ) . Therefore , we decided to examine the properties of proteins from different age groups in respect to their GC content . To identify the origin of the different properties of orphan and ancient proteins in different organisms we studied the distribution of different structural properties , including low complexity , fraction of transmembrane residues , secondary structure frequency and intrinsic disorder ) for all genomes against GC of the genomes , see Fig 3 . With the exception of β-sheet frequency , the difference between orphans and ancient proteins for all the considered properties is statistically significant: the p-value of a rank-sum test ( a non-parametric equivalent of the t-test ) is always < 10−11 . For proteins of all ages , low complexity ( SEG ) and predicted coil frequency increase with GC , while transmembrane , helix and sheet frequency decrease . Notable is that intrinsic disorder shows a clear , directly proportional dependency on GC: higher GC corresponds to more disorder . At the extreme ( over 60% GC ) , more than 50% of the residues in orphan proteins are predicted to be disordered , while for ancient proteins the disorder fraction is about 30% . At low GC ( below 40% ) the fraction of disordered residues is lower and similar in ancient and orphan proteins ( 15-20% ) . Further , the dependency of GC is clearly stronger for younger proteins , indicating that it is related to the creation of the protein and then gradually lost during evolution . To assess the significance of this dependency , we performed a linear regression test for each age group . The p-values of such test is presented for orphans and ancient in the boxes of Fig 3 . All the properties , with the exception of low complexity , show a p-value <0 . 01 , indicating that they are significantly correlated with GC in both orphan and ancient proteins . The GC is not constant over a genome . In complex eukaryotic organisms , the global GC content is heavily determined by the GC composition of isochores: these regions of uniform GC form a mosaic in the genomes of many complex eukaryotes , and their maintenance is likely the result of natural selection [37] . In general coding regions have higher GC than noncoding regions [38 , 39] . Further , there are also variation in GC between different regions of a genome , so when a noncoding region is turned into a gene the local GC will decide the amino acid content of the protein . Therefore , it might be more relevant to study the GC of each protein individually . In Fig 4 we show the dependency of structural properties on GC content for individual proteins . In addition , structural properties of a set of proteins generated randomly at all GC levels are shown . Orphans and genus orphans , as well as random proteins , show a definite dependency on GC . In contrast , ancient and intermediate proteins are only loosely dependent on GC . In general there is a resemblance between orphans and randomly generated proteins . However , when studying Fig 4 in more detail a few notable differences between them can be observed: orphans are more disordered and contain more low complexity regions but fewer sheets , independently of the GC level . It should be recalled that what we describe above is based on predicted structural features that are an indirect reflection of the protein sequence . If a certain group of proteins is predicted to be more disordered , or contain more sheets , it is quite likely a consequence of changes in amino acid frequencies . Next , we studied the proteins using six different amino acid propensity scales . The difference between the scales and predicted features is that scales describe general properties , and are directly calculated from amino acid frequencies , while predicted properties can also include other features . For disorder we use the TOP-IDP scale [30] , for hydrophobicity we use the biological hydrophobicity scale [31] , while sheet , turn , coil and helix propensities are analyzed using the structure-based conformational preferences scales [32] . In agreement with the predicted values; the average properties in the four age groups of proteins are overlapping , see S8 Fig . However , when taking the GC content into account the properties of the younger proteins show a strong correlation with GC , see Fig 5 . To a very large degree the properties of orphan proteins follow what would be expected for random proteins ( black line ) . However , regardless of GC , orphan proteins are more disordered and hydrophobic , have slightly higher turn and helical propensities , and also lower sheet propensities than random . Interestingly , the propensities of the two groups of older proteins also change by GC; however , this dependency is less pronounced than for orphan or random proteins . The difference seen between orphan and ancient proteins indicates that , given evolutionary time , the selective pressure to change the GC level is weaker than the selective pressure to change amino acid frequencies .
How do changes in GC content affect proteins ? In a random DNA sequence , the frequency of different codons changes depending on GC , and this , in turn , affects the resulting amino acid frequencies . To study the effect of GC content on amino acid frequency we examined the frequency of all 20 amino acids in proteins of different age . In Fig 6 , the expected and observed amino acid frequencies at different GC contents are explored . For most amino acids the observed frequencies are surprisingly well correlated with what is expected from GC alone . However , a few notable exceptions exist: In Fig 7 and S2 Table the GC content of the codons of each amino acid is compared with the propensity of that amino acid to be in a certain structural region . Three amino acids , Ala , Gly and Pro are “high GC” amino acids , i . e . they have more than 80% GC in their codons , while five amino acids , Lys , Phe , Asn , Tyr and Ile , are “low GC codons” , with less than 20% GC in their codons . All three “high GC” amino acids are disorder promoting ( high TOP-IDP ) , and four out of five “low GC” amino acids are order-promoting ( low TOP-IDP ) residues . Therefore at high GC content , codons coding for hydrophilic , disorder-promoting amino acid are prevalent . Genes low in GC tend to contain codons for hydrophobic amino acids , associated with order . All scales correlate with the GC frequencies with coefficients ranging from -0 . 42 to 0 . 39 . The strongest correlations are found with β-sheet propensity ( -0 . 42 ) and TOP-IDP ( 0 . 39 ) and the weakest with hydrophobicity ( 0 . 16 ) . We have studied the properties of proteins and their age in a large set of eukaryotic genomes . As shown before , orphan proteins are shorter than ancient proteins , but , surprisingly , we do find that on average for other structural features the young and old proteins are rather similar . However , we also observe that the properties of youngest proteins vary significantly with the GC content . At high GC the youngest proteins become more disordered and contain less secondary structure elements , while at low GC the reverse is observed . We show that these properties can be explained by changes in amino acid frequencies caused by the different amount of GC in different codons . The influence of this can be seen in the frequency of the amino acids that have a high or low fraction of GC in their codons , such as Pro . In a random sequence , Pro only represents less than 5% of the amino acids at 40% GC , but 10% at 60% GC . This actually agrees well with what is observed in orphan proteins: 5% at 40% GC vs . 9% at 60% GC , see Fig 6 . Similar changes in frequencies can be observed for several amino acids . On average , young proteins are more disordered than ancient proteins , but this property is strongly related to the GC content . In a low-GC genome the disorder content of an orphan protein is ∼30% while in a high-GC genome it is over 50% , see Fig 3 . Here we show that GC content of a genome strongly affects the amino acid distribution in de novo created proteins . It appears as if de novo created proteins that become fixed in the population are very similar to random proteins given a certain GC content . Codons coding for disorder-promoting residues are on average richer in GC , explaining the earlier contrasting observations between the low disorder among orphans in yeast ( a low GC organism ) and the high disorder among orphans in Drosophila ( a high GC organism ) . Finally , it can be observed that older proteins show a lower dependency of their structural properties on GC , but have a GC content similar to the one of orphans . This can lead to the speculation that selective pressure acts less on GC levels and more on structural features of proteins .
|
We show that the GC content of a genome is of great importance for the properties of an orphan protein . GC content affects the frequency of the codons and this affects the probability for each amino acid to be included in a de novo created protein . The codons encoding for Ala , Pro and Gly contain 80% GC , while codons for Lys , Phe , Asn , Tyr and Ile contain 20% or less . The three high GC amino acids are all disorder promoting , while Phe , Tyr and Ile are order promoting . Therefore , random protein sequences at a high GC will be more disordered than the ones created at a low GC . The structural properties of the youngest proteins match to a large degree the properties of random proteins when the GC content is taken into account . In contrast , structural properties of ancient proteins only show a weak correlation with GC content . This suggests that even after fixation in the population , proteins largely resemble random proteins given a certain GC content . Thereafter , during evolution the correlation between structural properties and GC weakens .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"microbiology",
"invertebrate",
"genomics",
"animals",
"animal",
"models",
"age",
"groups",
"fungi",
"drosophila",
"melanogaster",
"model",
"organisms",
"experimental",
"organism",
"systems",
"genome",
"analysis",
"paleontology",
"protein",
"structure",
"paleogenetics",
"fungal",
"pathogens",
"drosophila",
"saccharomyces",
"research",
"and",
"analysis",
"methods",
"mycology",
"proteins",
"medical",
"microbiology",
"microbial",
"pathogens",
"gene",
"ontologies",
"animal",
"genomics",
"insects",
"molecular",
"biology",
"yeast",
"arthropoda",
"people",
"and",
"places",
"candida",
"albicans",
"biochemistry",
"candida",
"earth",
"sciences",
"genetics",
"biology",
"and",
"life",
"sciences",
"yeast",
"and",
"fungal",
"models",
"saccharomyces",
"cerevisiae",
"genomics",
"population",
"groupings",
"computational",
"biology",
"organisms",
"macromolecular",
"structure",
"analysis"
] |
2017
|
High GC content causes orphan proteins to be intrinsically disordered
|
Plasmodium parasites , along with their Piroplasm relatives , have caused malaria-like illnesses in terrestrial mammals for millions of years . Several Plasmodium-protective alleles have recently evolved in human populations , but little is known about host adaptation to blood parasites over deeper evolutionary timescales . In this work , we analyze mammalian adaptation in ~500 Plasmodium- or Piroplasm- interacting proteins ( PPIPs ) manually curated from the scientific literature . We show that ( i ) PPIPs are enriched for both immune functions and pleiotropy with other pathogens , and ( ii ) the rate of adaptation across mammals is significantly elevated in PPIPs , compared to carefully matched control proteins . PPIPs with high pathogen pleiotropy show the strongest signatures of adaptation , but this pattern is fully explained by their immune enrichment . Several pieces of evidence suggest that blood parasites specifically have imposed selection on PPIPs . First , even non-immune PPIPs that lack interactions with other pathogens have adapted at twice the rate of matched controls . Second , PPIP adaptation is linked to high expression in the liver , a critical organ in the parasite life cycle . Finally , our detailed investigation of alpha-spectrin , a major red blood cell membrane protein , shows that domains with particularly high rates of adaptation are those known to interact specifically with P . falciparum . Overall , we show that host proteins that interact with Plasmodium and Piroplasm parasites have experienced elevated rates of adaptation across mammals , and provide evidence that some of this adaptation has likely been driven by blood parasites .
Malaria is one of the world's most notorious infectious diseases , responsible for billions of illnesses and millions of deaths in the last fifty years alone [1] . Human malaria is caused by five species in the genus Plasmodium , which are evolutionarily related to Babesia , Theileria , and other parasites in the order Piroplasmida . Approximately fifty Plasmodium species cause malaria in primates , rodents , and bats [2] , while Piroplasms infect a wider range of mammals ( Fig 1 ) . Although some wild animals appear to host malaria parasites without ill effects ( e . g . [3] ) , others are known to suffer serious symptoms and death , especially when exposed to novel parasites [4 , 5] . The severity of infection may then depend on the population history of exposure and individual acquired immunity , as it does in humans [6] . Parasites and other pathogens are important drivers of adaptive evolution in their hosts [7] . In the specific case of humans and Plasmodium , genetic variation in about 35 red blood cell or immune proteins has been associated with protection from severe complications of malaria , if not outright resistance ( reviewed in [8–10] ) . Some of these protective genes , including HBB , DARC , and GYPA , have been supported by population genetic evidence of selection in African or Southeast Asian populations within the last 75 , 000 years ( e . g . [11–13] ) . Malaria has consequently been labeled "one of the strongest selective forces on the human genome" [9 , 10] , though this statement has never been quantified . Human adaptation to malaria is likely occurring within the broader context of mammalian adaptation to widespread blood parasites . The common ancestor of modern humans existed perhaps 200 , 000 years ago , while the common ancestor of placental mammals dates back ~105 million years [14 , 15] . For comparison , the parasite genus Plasmodium experienced a major radiation ~129 million years ago [16] . Plasmodium and Piroplasms have likely infected mammals for as long as mammals have existed , but the evolutionary consequences of this long-standing relationship have never been investigated . Despite their age and diversity , Plasmodium and Piroplasms cause disease through similar mechanisms , including transmission to and from mammalian blood by the bite of a mosquito or tick . Plasmodium cells migrate first to the liver , multiply within hepatocytes , and emerge several days later to invade red blood cells ( RBCs ) [17] . Babesia parasites invade RBCs directly , while Theileria parasites infect both red and white blood cells [18 , 19] . Although Piroplasms like Babesia and Theileria are thought to lack a liver stage [20] , their infections cause substantial liver damage through increased coagulation and other mechanisms [21–23] . Parasitized cells also adhere to capillaries lining the liver , lung , brain , and other tissues , which can impair circulation and lead to life-threatening organ dysfunction ( e . g . [18 , 24–26] ) . Finally , each parasitic infection triggers a complex immune response from the host , including the removal of infected RBCs from circulation by the spleen ( e . g . [27] ) . The complexity of these host-parasite interactions makes it difficult to precisely measure their evolutionary impact . One important reason is that our knowledge of host responses is biased toward convenient samples , like blood cells , from specific groups , like humans and Plasmodium . We particularly lack information across the extant diversity of parasite and host species [2] . A second key reason is that host genes relevant to malaria are likely to be pleiotropically involved with other selected phenotypes , including responses to other pathogens [28] . In particular , some Plasmodium-associated genes in humans are also associated with viruses or bacteria , making it difficult to attribute their evolution specifically to pressure from Plasmodium [29 , 30] . Parsing the contribution of various pathogens to host evolution thus requires a broader understanding of many host genes , many tissues , and many pathogens . In this work , we examine patterns of adaptation and functional pleiotropy in a set of ~500 Plasmodium- or Piroplasm-interacting proteins ( PPIPs ) manually curated from the literature . These PPIPs represent about 5% of the mammalian proteome , as defined by the set of 9 , 338 proteins conserved across 24 well-sequenced mammal species . Previously , evolutionary analysis of an externally defined gene set has proven useful for detecting polygenic adaptation [7 , 31–33] . Here , because PPIPs represent a relatively small fraction of all conserved mammalian genes , we use permutation tests to compare PPIPs to a background of non-PPIP controls . That is , we compare PPIPs to many sets of other mammalian proteins , which we match to PPIPs by a number of important metrics . This approach has recently been used by [31] to identify viruses as a dominant driver of adaptation in mammals . Overall , we demonstrate that PPIPs have experienced ~3 times more adaptive substitutions than expected throughout mammalian evolution . The strongest adaptive signals are present in PPIPs with immune functions , which are highly pleiotropic with respect to other pathogens . However , we detect a significant excess of adaptation even in non-immune PPIPs that are not known to interact with pathogens beside Plasmodium . Additional evidence suggests that the red blood cell protein alpha-spectrin , as well as PPIPs highly expressed in the liver , may have played key roles in adaptation to blood parasites . Overall , our work supports the hypothesis that Plasmodium and Piroplasm parasites—not unlike other classes of pathogens—have been important and long-standing drivers of evolutionary change in mammals .
Malaria-like illnesses generate substantial health and economic burdens in humans , livestock , and pets [1 , 34] . These costs have motivated a large body of research into host-parasite interactions and host responses to infection . We queried the PubMed database for scientific papers whose abstracts mentioned the name of a host gene along with the terms malaria , Plasmodium , Babesia , Theileria , Rangelia , or Cytauxzoon , the latter four being the best-studied Piroplasmid genera ( Methods , PPIP Identification ) . To focus on mammalian evolution , we limited our search to 9 , 338 protein-coding genes that are conserved in 24 mammalian species with high-quality reference genomes ( Fig 1; Methods , Mammalian Orthologs; [31] ) . Most of these mammalian species belong to one of four orders—primates , rodents , artiodactyls , or carnivores—and represent a range of susceptibilities to our focal parasites ( Fig 1 ) . This search returned ~35 , 000 papers associated with ~5 , 000 mammalian genes . However , the vast majority of these results were false positives . Many short acronyms that identify genes have multiple meanings , and many papers containing these acronyms do not concern genes or proteins . We manually curated paper titles and abstracts to identify just 786 papers linking 490 proteins to Plasmodium or Piroplasms via four types of phenotypic evidence: ( 1 ) biochemical interaction between a mammalian and parasite protein; ( 2 ) statistical association between genetic variation and disease susceptibility; ( 3 ) knockout or overexpression studies impacting susceptibility; and ( 4 ) low-throughput studies showing a change in gene expression during infection ( Fig 2A; S2 Table ) . Nearly half of the 490 PPIPs ( 45% ) were supported by multiple studies , and 35% by multiple sources of evidence . Expression changes were the most common form of evidence ( 85% of PPIPs , or 60% of total evidence ) , with about 43% of PPIPs identified only via expression changes . Expression-based PPIPs likely represent both direct and indirect interactions with parasites , given the size and interconnectedness of gene expression networks ( e . g . [35] ) . We have attempted to limit indirect interactions by excluding high-throughput expression experiments , and we later show that all four evidence types , including expression changes , identify sets of genes with elevated rates of adaptation ( S4 Fig ) . Consequently , we chose to analyze all PPIPs together without making distinctions based on evidence type . The majority of PPIPs were linked to Plasmodium in studies of humans and mice ( Fig 2 ) . A fifth of PPIPs ( 21% ) were linked to Piroplasms in studies of cows , dogs , and other mammal species ( Fig 2 ) . Plasmodium- and Piroplasm-interacting proteins overlap substantially , ~14 times more than expected by chance ( p<1x10-5 , Fig 2B ) . This overlap is consistent with the similar life cycles of Plasmodium and Piroplasms , as well as the conservation of host responses to these parasites across mammals . To ask whether PPIPs generally perform functions relevant to malaria , we also tested 17 , 696 GO functional categories for PPIP enrichment ( Methods , Protein Metrics ) . After correcting for multiple testing , over 1 , 200 categories contained significantly more PPIPs than expected ( S3 Table ) . The most enriched categories were dominated by immune functions , with 51% of PPIPs falling under immune system process ( p = 6 . 16 x 10−94 ) and 83% under response to stimulus ( p = 7 . 89 x 10−69 ) ( Table 1 ) . Other highly enriched categories indicate functions more specific to malaria pathology , including cell adhesion ( p = 9 . 89 x 10−32 ) , hemostasis ( p = 1 . 68 x 10−35 ) , and hemopoiesis ( p = 2 . 20 x 10−37 ) ( Table 1 ) . Together , these results reflect the expected functions of host genes ascertained through studies of malaria-relevant processes . Many immune genes , even outside the adaptive immune system , are activated by signals from multiple pathogens ( e . g . [36 , 37] ) . This 'pathogen pleiotropy' poses an important complication when testing the link between blood parasites and host adaptation , even in genes phenotypically linked to these parasites . To quantify the extent of pathogen pleiotropy in mammals , we compared PPIPs to host proteins known to interact with viruses and bacteria ( Methods , VIPs and BIPs ) . In both cases , we focused only on genes conserved across our focal 24 mammal species . For viruses , we obtained a high-quality list of 1 , 256 manually curated virus-interacting proteins from [31] ( S4 Table ) . For bacteria , we queried the EBI IntAct database [38] for all deposited interactions between humans and bacteria , which returned 1 , 250 host proteins ( S4 Table ) . Overall , we find that 36% of all PPIPs also interact with viruses , 23% with bacteria , and 48% with viruses and/or bacteria—many more than expected by chance ( Fig 2C; all p<10-4 ) . Unsurprisingly , this overlap is strongest for immune PPIPs ( here defined as falling under the GO category immune system process ) , of which 56% interact with multiple pathogens ( p<10-4 ) . However , nearly 40% of non-immune PIPs also interact with multiple pathogens ( p<10-4 ) . Some of these "non-immune" proteins may have uncharacterized immune functions , but most are known for their involvement in general cellular processes , including metabolism and signal transduction . This suggests that a diverse array of prokaryotic , eukaryotic , and viral pathogens may interact with a surprisingly small number of host proteins , or alternatively , that these proteins represent a non-specific host response to infection . Plasmodium and Piroplasms influence several mammalian tissues as they progress through their complex life cycle . To begin investigating the specificity of PPIPs to malaria-like infections , we examined gene expression in a condensed set of 34 human tissues collected by the GTEx Consortium from uninfected individuals [39] ( Methods , Protein Metrics ) . We first found that PPIPs have an average of 8 . 2% higher total expression than randomly selected sets of non-PPIPs ( p<0 . 001; S1A Fig ) . To fairly evaluate PPIP expression enrichment in each tissue , we designed a matched permutation test that compares PPIPs to many , similarly-sized sets of control genes with similar total expression ( Methods , Permutation Tests ) . Throughout this work , we use matched permutation tests to compare PPIPs to many sets of other genes that are controlled for confounding factors . This approach allows us to isolate the evolutionary effects of interactions with Plasmodium or Piroplasms from potentially correlated factors , such as high total expression . Compared to matched control proteins , we found that PPIPs were significantly differentially expressed in 20 of 34 tissues , despite the noisiness of this measurement ( Fig 3 ) . PPIP expression was underrepresented in 16 tissues , particularly those involved in reproduction , and overrepresented in four tissues: blood , liver , lung , and spleen . All four of these overrepresented tissues are key sites of parasite replication , containment , and/or tissue damage in Plasmodium or Piroplasm infections in mammals . This result may be due in part to an ascertainment bias in sampling certain tissues , especially blood , but is also consistent with the biology of host interactions with Plasmodium and Piroplasm parasites . We have already shown that PPIPs have three unusual properties—immune enrichment ( Table 1 ) , excess interactions with other pathogens ( Fig 2C ) , and high mRNA expression ( S1A Fig ) —that may influence their rate of evolution . In order to evaluate PPIP adaptation against an appropriate background , we assessed several additional metrics for PPIPs and other proteins in order to control for them in our permutation tests ( S5 Table ) . First , we examined three additional broad measures of gene function in humans: the density of DNAseI hypersensitive elements; protein expression , as measured by mass spectrometry; and the number of protein-protein interactions ( Methods , Protein Metrics ) . For each of these metrics , PPIPs have significantly higher mean values than sets of random controls , indicating that PPIPs are more broadly functional in humans ( S1B–S1D Fig; all p<0 . 001 ) . We next tested four measures of genomic context that have been linked to the rate of sequence evolution: GC content; aligned protein length; the regional density of protein-coding bases; and the density of highly conserved , vertebrate elements [40–43] ( Methods ) . Most of these metrics do not differ between PPIPs and other genes ( S1F–S1H Fig ) , with the exception of conserved element density , which is slightly but significantly lower in PPIPs ( mean = 8 . 0% vs . 8 . 8%; p = 0 . 002; S1E Fig ) . Based on these results , we expanded our permutation test to match all five of these significantly varying measures of gene function and genomic context while generating sets of non-PPIP control genes . Each non-PPIP was considered an acceptable match for a given PPIP if its values for all five metrics fell within specific ranges of the PPIP values ( Methods , Permutation Tests ) . About 10% of PPIPs were too dissimilar from other proteins to be matched and were excluded from subsequent analysis , but these proteins show similar rates of adaptation to other PPIPs ( S6 Table ) . On average , each retained PPIP could be matched to 34 control genes , allowing the generation of many different sets of ~440 matched controls . This permutation procedure effectively equalized distributions between PPIPs and control genes for all tested functional and evolutionary metrics ( compare S1 Fig to S2 Fig ) . Finally , one of the largest differences between PPIPs and other proteins is the frequency with which they are discussed in the scientific literature ( S1I Fig ) . The average PPIP has 6 . 9 times more PubMed citations than the average mammalian protein ( Methods , Protein Metrics ) . This difference was too large to match in the permutation test without excluding the majority of PPIPs . However , we show that the citation frequency of non-PPIPs has no relationship with protein adaptation ( p ≥ 0 . 17; S3 Fig ) . This indicates that a high rate of citation for PPIPs is not statistically associated with their rate of adaptation . After controlling for each metric of function and genomic context ( S2 Fig ) , we asked whether PPIPs exhibit unusual patterns of amino acid substitution and polymorphism in mammals . Importantly , PPIPs have a typical ratio of non-synonymous to synonymous polymorphism in a combined sampling of great ape species ( Fig 4A; mean pN/ ( pS+1 ) = 0 . 21 in PPIPs vs . an average of 0 . 20 in matched controls; p = 0 . 10; Methods , Protein Metrics ) . That is , PPIPs do not appear more or less evolutionarily constrained than other similar proteins , bolstering the null expectation that they should evolve at average rates . In contrast , we find that PPIPs have a significantly elevated ratio of non-synonymous to synonymous substitutions across 24 mammal species ( dN/dS = 0 . 186 in PPIPs vs . an average of 0 . 128 in matched controls , p<10−4 ) . If we make a very conservative ( and in many ways unreasonable ) assumption that matched controls have experienced no adaptation in the history of mammals , then this 31% excess in dN/dS in PPIPs , despite unremarkable pN/ ( pS+1 ) , implies that at least 31% of all amino acid substitutions in PPIPs were adaptive . However , given that matched controls have likely experienced at least some adaptation , the proportion of adaptive substitutions in PPIPs is likely to be even larger . We investigated adaptation in PPIPs in more detail using the BS-REL and BUSTED tests available in the HYPHY software package [44–46] ( Methods , Estimating Adaptation ) . Both tests use maximum likelihood models to estimate the proportion of codons in a protein with dN/dS > 1 , consistent with adaptation in some proportion of the protein . BS-REL estimates the exact proportion in each branch of the tree , whereas BUSTED estimates whether it is greater than zero in at least one branch ( i . e . , whether a gene has experienced positive selection at some point in the history of mammalian evolution ) . Both models find additional evidence of excess adaptation in PPIPs . Nearly 36% of PPIPs have BUSTED evidence ( at p≤0 . 05 ) of adaptation in some part of the mammalian phylogeny , versus 24% of matched controls ( p<10−4; Fig 4C ) . PPIPs also have BS-REL evidence for adaptation on more branches of the mammalian tree ( p = 1 . 87× 10−4; Fig 4D ) , as well as for more codons per protein ( p<10−4; Fig 4E ) . This excess is robust to the BUSTED p-value threshold used to define adaptation , and increases as the threshold becomes more stringent ( Fig 4F , p = 7x10-5 ) . PPIPs identified via different kinds of evidence all have more adaptation than expected by chance , although PPIPs that physically interact with parasite proteins have a greater excess of adaptation than PPIPs that change expression during infection ( p = 0 . 032 , S4 Fig ) . These matched permutation tests show that PPIPs have experienced an elevated rate of adaptive substitutions in mammals . Although the hundreds of published experiments that define PPIPs ( Fig 2A ) support the idea that Plasmodium and Piroplasms may have driven this adaptation , it remains critical to address PPIP pleiotropy with other pathogens ( Fig 2C ) . Based on the available information for many host-pathogen interactions ( Methods , VIPs and BIPs ) , we divided PPIPs into three categories: "Plasmodium-only , " "Plasmodium + Piroplasms , " and "multi-pathogen , " which includes PPIPs that also interact with viruses and/or bacteria ( Fig 2C ) . For each category , we again matched PPIPs to controls and found significantly more adaptive substitutions than expected ( Fig 5A ) . Among categories , the excess adaptation for PPIPs versus controls is greater when more diverse pathogen interactions are included ( p<0 . 05 ) . That is , Plasmodium-only PPIPs have 1 . 9X more adaptation than expected; PPIPs that interact with Plasmodium and/or Piroplasms but are not known to interact with viruses or bacteria have 2 . 5X more adaptation than expected; and PPIPs that also interact with viruses or bacteria have 3 . 7X more adaptation than expected ( Fig 5A ) . This result suggests that an increased number and diversity of pathogen interactions drives a cumulative increase in host adaptation . Importantly , PPIPs that interact with more pathogens are also more likely to have immune functions ( Fig 5A , Fig 2C ) . Only 39% of Plasmodium-only PPIPs , versus 59% of multi-pathogen PPIPs , have a GO annotation for immune system process ( Fig 5A ) . Immune genes are well known to evolve at rapid rates [47–52] . Here , we also find that non-immune PPIPs have adapted at slower rates than PPIPs as a whole ( Fig 5B; p = 0 . 018; see Methods , Permutation Tests for why immune PPIPs are not analyzed directly ) . These correlations among immune function , adaptation , and multi-pathogen interactions complicate the link between malaria-like parasites and host adaptation . Fortunately , these correlations can be disentangled by considering the 239 PPIPs that do not have an annotated immune function . In these non-immune PPIPs , there is a breakdown of the link between adaptation and multi-pathogen interactions ( Fig 5C ) . That is , non-immune PPIPs known to interact only with Plasmodium have ~2X more adaptation than expected , and this excess does not significantly increase when PPIPs that interact with additional pathogens are included ( all p > 0 . 17; Fig 5C ) . We note that this lack of a significant difference is not simply due to reduced power from a reduced sample size , given that subsampling of PPIPs in Fig 5A to the sample sizes in Fig 5C retains complete power to detect differences . Overall , we show that even non-immune PPIPs not known to interact with any pathogens except for Plasmodium still show sharply elevated rates of adaptation . Although we lack complete knowledge of host-parasite interactions , to explain this result independently of Plasmodium as a selective pressure would require the existence of some other pressure or pathogen , whose interactions with mammalian genes overlap remarkably well with those of Plasmodium . Host adaptation to malaria could potentially be concentrated in any malaria-relevant tissue enriched for PPIP expression , specifically blood , liver , lung , and spleen ( Fig 3 ) . We used a threshold analysis to test whether expression in these tissues was linked to elevated adaptation . That is , we compared expression patterns for the bulk of genes to patterns in the 5% of genes with the most adaptive codons for both PPIPs and controls matched for total expression ( Fig 6 ) . In sets of matched control non-PPIPs , the most highly adaptive genes are expressed at significantly lower levels than the bulk of control non-PPIPs in all four malaria-relevant tissues ( Fig 6 ) . In contrast , for PPIPs , highly adaptive genes are not expressed at significantly lower levels in any of the tissues , despite the same overall sample size and level of total expression . In fact , in the case of the liver , the highly adaptive PPIPs are expressed at significantly higher levels than other PPIPs , the opposite direction of the pattern observed in controls ( p = 9 . 8 x 10−4; Fig 6 ) . The fact that high expression in the liver is associated with elevated adaptation in PPIPs , but not controls , suggests that the liver may have been a site of particularly strong selective pressures acting specifically on PPIPs . Plasmodium and Piroplasm infections have been reported from a wide variety of mammalian species ( Fig 1 ) . We tested whether PPIP adaptation is similarly widespread across mammals by applying BUSTED and BS-REL models to subsets of the sequence data within individual mammalian orders ( Methods , Order-specific Analyses ) . When all PPIPs are considered , we find a highly significant excess of PPIP adaptation in rodents and primates ( both p<0 . 001; Fig 7 ) . Before correcting for multiple testing , the signal is marginally significant in carnivores ( p = 0 . 052 ) and positive , but not significant , in artiodactyls ( p = 0 . 29 ) . However , it is difficult to compare significance among clades for two reasons . First , we cannot account for differences in evolutionary rate in different groups due to , e . g . , generation time . Second , our mammalian tree has not sampled equal numbers of species from each the four clades ( Fig 1 ) . We further note that we completely lack statistical power to perform clade-specific analysis on subsets of PPIPs , such as Plasmodium- or Piroplasm-only PPIPs ( Methods , Permutation Tests ) . Despite these caveats , clade-specific analyses indicate at least a trend toward high adaptation in PPIPs in all major clades of the mammalian tree . Direct , biochemical interactions between mammalian and parasite proteins may be particularly important drivers of host adaptation ( S4 Fig ) , although such interactions remain uncharacterized for the majority of PPIPs ( Fig 2A ) . We chose one well-studied PPIP to test for a direct relationship , at the amino-acid level , between host adaptation and biochemical host-parasite interactions . Of the top ten PPIPs with the strongest BUSTED evidence of adaptation , alpha-spectrin ( SPTA1 ) is the only candidate that has been extensively characterized for molecular interactions with Plasmodium proteins . Alpha-spectrin is a textbook example of a major structural component of the red blood cell ( RBC ) membrane . In humans , dozens of polymorphisms in this gene are known to cause deformations of the RBC , which may either be asymptomatic or cause deleterious anemia ( reviewed in [53] ) . These deformations are more common in individuals of African descent , leading to the hypothesis that SPTA1 is involved in malaria resistance in humans . The SPTA1 protein has a well-defined domain structure , and specific interactions with Plasmodium proteins are known for three domains ( Fig 8 ) . Repeat 4 is the binding site for KAHRP , the major P . falciparum component of the adhesive 'knobs' that form on the surface of infected RBCs [54] . Another 65-residue fragment containing EF-hand 2 has been shown to bind to PfEMP3 , an interaction that destabilizes the RBC skeleton and may allow mature merozoites to egress from the cell [55] . A central SH3 domain can also be cleaved by a promiscuous Plasmodium protease called plasmepsin II [56] , which mainly functions in hemoglobin digestion [57] . Furthermore , naturally occurring mutations in the first three SPTA1 domains have been shown to impair the growth of P . falciparum in human RBCs [58–60] . We wished to test whether sites of mammalian adaptation in SPTA1 mapped to any of these Plasmodium-relevant domains . To identify adaptive codons with higher precision and power , we aligned SPTA1 coding sequences from 61 additional mammal species ( S7 Table , S10 Table ) for analysis in MEME [61] ( Methods , Alpha-spectrin ) . Of the 2 , 419 codons in this large mammalian alignment , we found that 63 show strong evidence of adaptation ( p<0 . 01 ) , and that these are distributed non-randomly throughout the protein . Remarkably , three domains—Repeat 1 , Repeat 4 , and EF-hand 2—are significantly enriched for adaptive codons , after controlling for domain length and conservation ( Fig 8; Methods ) . That is , all three SPTA1 domains with strong evidence of adaptation in mammals are known to either interact specifically with P . falciparum proteins , or harbor human mutations that provide resistance to P . falciparum . This overlap is unlikely to occur by chance ( p = 0 . 015 ) and is robust to the p-value thresholds chosen ( S8 Table ) . Thus , evidence from SPTA1 suggests a specific connection , at least in this well-studied example , between the mechanics of Plasmodium infection and adaptation in the host red blood cell . Notably , we do not claim that all adaptation in SPTA1 is due to pressure from malaria . Adaptation has occurred in at least one codon of SPTA1 on every branch of the 85-species mammalian tree , with the top branches including the base of the Camelidae , Loxodonta africana ( elephant ) , Trichechus manatus ( manatee ) , and the base of all Eutheria ( S11 Table ) . The especially high density of adaptive changes in camels may be related to the unusual shape of their red blood cells , which has been shown to extend RBC lifespan during chronic dehydration [62 , 63] . Nonetheless , when we focus only on the three domains of SPTA1 that are enriched for adaptive substitutions ( Fig 8 ) , we find much stronger evidence of adaptation on primate branches than when the entire SPTA1 protein is considered ( S11 Table ) . This observation is consistent with known molecular interactions between P . falciparum and these specific SPTA1 domains in humans . Together , branch- and domain-specific patterns of adaptation in SPTA1 support malaria as an important , but by no means unique , influence on the evolution of mammalian red blood cells .
In this work , we have identified 490 conserved mammalian proteins that interact with Plasmodium or Piroplasm parasites . This large set of PPIPs is a substantial expansion of the list of host proteins traditionally considered to be associated with Plasmodium and Piroplasms ( e . g . [10] ) , enabling us to investigate both their broad functional properties and long-term evolutionary patterns . We find that PPIPs are strongly enriched for immune annotations , although about half are involved in diverse non-immune processes ( Table 1 ) . PPIPs are also widely expressed , but particularly overrepresented in the blood , liver , lung , and spleen—tissues highly relevant to the Plasmodium and Piroplasm life cycles . We find that PPIPs tend to interact not only with multiple blood parasites ( Fig 1B ) , but also with unrelated bacterial and viral pathogens ( Fig 1C ) . As expected , this multi-pathogen overlap is strongest for immune PPIPs . Somewhat surprisingly , this overlap also extends to non-immune PPIPs , suggesting either that unrelated parasites tend to interact with the same host proteins or that these proteins correspond to some non-specific host response . Our key result is that PPIPs have been evolving unusually quickly compared to carefully matched non-PPIPs ( Fig 4 ) . If we conservatively assume that none of the amino acid substitutions in non-PPIPs have been adaptive , then we may estimate that 31% of amino acid substitutions in mammalian PPIPs have been driven by positive selection . However , because non-PPIPs have also experienced appreciable positive selection ( Fig 4C ) , the true proportion of adaptive substitutions in PPIPs is certainly higher . Regardless of the precise number , it is clear that host proteins that interact with Plasmodium or Piroplasm parasites have evolved at an elevated rate , with a substantial proportion of amino acid changes driven by positive selection ( Fig 4 ) . Across mammals , the rate of PPIP evolution is comparable to that of other proteins previously identified as targets of strong positive selection . For example , the antiviral protein PARP14 and the sperm-expressed protein TEX15—which are not PPIPs—represent two classes of proteins that have diversified rapidly in some mammals [64 , 65] . When we consider the proportion of codons under positive selection in all our mammalian orthologs , we find that PARP14 ranks in first place ( 7 . 2% ) , TEX15 in third place ( 5 . 8% ) , and the Piroplasm-associated ENTPD1 in second place ( 6 . 3% ) ( S5 Table ) . As another example , in colobine monkeys , a shift to foregut fermentation is thought to have driven adaptive substitutions in approximately half of the residues of the antibacterial enzyme lysozyme [66] . Across our broader sampling of mammals , lysozyme ranks in 228th place ( 1 . 8% adaptive codons ) , behind 41 PPIPs ( S5 Table ) . It can be difficult to make precise comparisons across studies that include different mammal species , and with the exception of a recent study of viruses [31] , few have systematically assessed the importance of particular selective pressures across many mammals . Nonetheless , PPIPs appear to have experienced similar rates of adaptation as some of the better-known mammalian examples . Given the extreme pleiotropy of PPIPs in regards to other pathogens , a natural question is whether Plasmodium and Piroplasms are truly drivers of PPIP adaptation . Our set of 490 PPIPs is large enough to begin parsing the specific effects of blood parasites on mammalian evolution , independent of the effects of other pathogens . When we consider PPIPs that interact only with Plasmodium or Piroplasms , but not viruses or bacteria , we find a 2 . 5X enrichment of adaptation compared to matched controls ( Fig 5A ) . For non-immune PPIPs in particular , additional viral or bacterial interactions do not elevate the excess of adaptation , which remains significantly higher than in matched controls ( Fig 5C ) . This provides some evidence that blood parasites have played a specific role in influencing mammalian protein evolution , both in immune and non-immune genes . Two additional pieces of evidence are consistent with this idea . First , PPIPs with the highest levels of adaptation are also particularly highly expressed in the liver , opposite to the pattern seen in matched controls ( Fig 6 ) . This suggests the possibility that adaptation is related not simply to liver expression , but to parasite interactions that take place in the liver . In Plasmodium infections , parasites initially migrate to the liver , invade hepatocytes , and replicate many times before emerging to infect red blood cells [17] . In Piroplasm infections , liver damage is also common and associated with fatality ( e . g . [21 , 22 , 67] ) . Although some Piroplasms are thought to lack a liver stage [20 , 68] , a number of studies have reported the presence of Babesia , Rangelia , or Cytauxzoon parasites within the endothelial cells of the liver , among other tissues [23 , 69–71] . The liver may thus represent a critical opportunity for PPIPs to ameliorate the effects of Plasmodium and Piroplasm infection on the host . Second , in the well-studied case of alpha-spectrin ( SPTA1 ) , we were able to directly investigate the correspondence between sites of host protein adaptation and sites of molecular interactions with Plasmodium . We indeed found strong evidence of adaptation in three domains of SPTA1 that are known to participate in molecular interactions with Plasmodium parasites ( Fig 8 ) , consistent with host-parasite interactions specifically driving mammalian adaptation . Notably , our evidence of adaptation in SPTA1 was derived from long-term evolutionary patterns in dozens of mammal species , whereas molecular interactions with P . falciparum were identified only in humans . This is analogous to the interspecies evolutionary patterns we observe in the PPIPs identified from intraspecies association studies ( S4 Fig ) . These results suggest that the host cellular machinery underlying extant parasite interactions has been largely conserved over deep evolutionary time , potentially allowing the same proteins to participate in adaptation across multiple time scales . In the end , despite these three lines of evidence pointing towards Plasmodium and Piroplasms specifically driving PPIP adaptation , this conclusion must remain tentative because of our incomplete knowledge of host-pathogen interactions . At one level , many host genes that interact with pathogens likely remain unidentified [31] . At another level , the taxonomic distribution of pathogens on hosts remains quite poorly understood [2] . This is especially problematic for testing whether adaptation in certain mammal lineages corresponds to the densities of specific parasites . For example , based on veterinary records , we may have expected artiodactyls to adapt specifically to Piroplasms but not to Plasmodium ( Fig 1 ) . However , white-tailed deer in North America were recently discovered to carry a Plasmodium species at high frequency [72] . This finding demonstrates that absence of proof is not proof of absence when it comes to the phylogenetic distribution of pathogens , nor to interactions between parasites and host genes . Emerging high-throughput studies of host-pathogen interactions , combined with broader sampling of natural infections , will allow more precise tests of how hosts evolve in response to specific pathogens . In our case , the fact that non-immune , "Plasmodium-only" PPIPs show a clear excess of adaptation ( Fig 5 ) may reflect either specific interactions with Plasmodium or incomplete knowledge of interactions with other pathogens . Likewise , high rates of adaptation in PPIPs highly expressed in the liver may reflect adaptation to liver-antagonizing blood parasites , or to other viral and bacterial pathogens that also damage the liver . Our analysis of SPTA1 provides the most compelling evidence of a specific association between adaptation in a PPIP and interactions with Plasmodium , but because this is only a single example , we cannot claim that such associations would be found more broadly if other PPIPs were to be studied in similar detail . However , these results are hopeful for our future ability to identify specific selective pressures associated with specific pathogens . Other future work could also examine PPIPs for evidence of balancing selection , especially as more non-human polymorphism data become available . Several examples of PPIP evolution in humans indicate an important role for the maintenance of polymorphism [11 , 73] , and it is possible that sampling of PPIP polymorphism in other species has contributed to the elevated divergence shown here . Balancing selection within species and directional selection across species may even be two sides of the same coin , as evidenced by immune and other genes that appear to have experienced both [11 , 74] ( S5 Table; S4 Fig ) . Indeed , balanced states have been shown to be a natural consequence of directional selection in fast-changing environments [75] . In human populations with abundant polymorphism data , PPIPs could be used as an important resource for understanding the relationship between these two selective modes . In conclusion , we show that proteins that interact with Plasmodium and Piroplasms comprise a substantial portion of the mammalian proteome; that they exhibit high rates of adaptation across mammals; and that this adaptation may be partially driven by these blood parasites . We hope that the collection of 490 mammalian PPIPs will continue to prove a powerful and continually growing resource for exploring host-parasite interactions and adaptation .
We queried PubMed for scientific papers containing both a gene name and the term malaria , Plasmodium , Babesia , Theileria , Rangelia , or Cytauxzoon in the title or abstract , as of Feb . 20 , 2017 . Human gene names were drawn from the HUGO Gene Nomenclature Committee [76] ( http://www . genenames . org/ ) for 9 , 338 mammalian orthologs ( see Methods , Mammalian Orthologs ) . For each of the genes that returned at least one hit , we manually evaluated the titles of up to 20 associated papers to assess the link between the gene and a malaria phenotype . Many acronyms used to represent genes are also used as abbreviations for techniques , locations , drugs , or other phrases . Consequently , most genes could be eliminated based on their nominal connection with papers addressing non-genetic aspects of malaria . For papers discussing genes , we examined the abstracts for the presence and type of evidence connecting genes to malaria phenotypes . In cases where the abstract was ambiguous , we examined the full text of the paper . To limit the number of false positives , we did not classify PPIPs using evidence from RNAseq or other high-throughput experiments . Gene expression is typically regulated via large , interconnected networks ( e . g . [35] ) , such that high-throughput experiments can identify hundreds or thousands of genes whose expression is perturbed by infection . Many of these differentially expressed genes may have very small or indirect impacts on the progression of malaria , making them unlikely to be important targets of malaria-related selection . In contrast , low-throughput expression experiments are typically based on a priori knowledge or hypotheses of the more direct roles of a few host genes in malaria . Focusing on candidate genes may inflate the rate of false positives in genetic association studies [77] . Here , we make substantial efforts to ensure that any bias potentially related to PPIP identification , such as the popularity of a gene , does not impact our results ( S1–S3 Figs ) . While we cannot guarantee that every PPIP is a true positive , in part because replication has often not been attempted , PPIPs as a whole do appear to represent a meaningful class of genes . In general , misclassification of either PPIPs or non-PPIPs for any reason ( false negatives or false positives ) would reduce any true difference between the two categories , weakening our results . We used BLAT to identify homologs of 22 , 074 human coding sequences in 24 high-depth mammal genomes ( Fig 1 ) . We retained orthologs which ( 1 ) had best reciprocal hits in all 24 mammal species , ( 2 ) lacked any in-frame stop codons , ( 3 ) were at least 30% of the length of the human sequence , and ( 4 ) had clearly conserved synteny in at least 18 non-human species . Coding sequences for the resulting 9 , 338 proteins were aligned with PRANK ( S1 File ) , and any codon present in fewer than eight species was excluded from analysis . Additional details are available in [31] . GO annotations were downloaded in October , 2015 from the Gene Ontology website [78] ( http://geneontology . org/ ) . Tests of enrichment were performed using Fisher's Exact Test . Expression data for 53 human tissues were downloaded from the GTEx portal ( http://www . gtexportal . org/home/ ) on March 6 , 2017 . These 53 tissues were condensed into a set of 34 tissues by averaging the RPKM across multiple tissue samples from adipose , aorta , artery , brain , cervix , colon , esophagus , heart , and skin . Each data point was then transformed as log2 ( RPKM+1 ) . Total expression for each gene was calculated as the sum of these transformed values across all 34 tissues . Regions of DNaseI hypersensitivity , combined from 95 cell types , were obtained from the databases of the ENCODE Project Consortium [79] ( https://www . encodeproject . org/ ) . The density of DNaseI hypersensitivity regions was calculated in 50 Kb windows centered on each ortholog . Protein expression levels were obtained from the Human Proteome Map [80] ( http://www . humanproteomemap . org/ ) , which used high resolution and high accuracy Fourier transform mass spectrometry experiments . We summed spectral values over 30 tissues and cell types and took the log of these total values . The log number of interacting partners for each human protein was obtained from the Biogrid Database [81] ( http://thebiogrid . org/ ) , curated by [82] . Genomic elements conserved in 46 vertebrate species , derived from PhastCons [43] , were downloaded from the UCSC genome browser ( http://hgdownload . cse . ucsc . edu/goldenPath/hg19/phastCons46way/ ) . Conserved element density was calculated within 50 kb windows centered on each gene in the human reference . Coding density was calculated from coding nucleotides in the same 50 Kb windows . The length and GC content of each protein was derived from the mammalian alignment . The citation frequency of each gene was determined by the number of citations linked to its PubMed Gene page ( http://www . ncbi . nlm . nih . gov/gene ) as of May 11 , 2017 . Polymorphism data for great apes—chimpanzee , gorilla , and orangutan—was obtained from the Great Apes Genome Project [83] . For each individual species , the counts of polymorphic sites are low , making the pN/pS ratio a noisy measure . This problem was alleviated by combining all the great ape data , which provided an overall control for the level of purifying selection across multiple species . Virus-interacting proteins ( VIPs ) were manually curated in [31] in the same manner as PPIPs . To our knowledge , no similar collection of high-quality interactions is available for other pathogens . Therefore , we queried the EBI IntAct database ( http://www . ebi . ac . uk/intact/ ) for protein interactions between Kingdom Bacteria ( taxid:2 ) or Phylum Apicomplexa ( taxid:5794 ) and humans ( taxid:9606 ) . This approach , while much faster than manual curation , is less ideal for two reasons: ( 1 ) many interactions are not included in the database ( e . g . , only 17 human-Plasmodium interactions are included in IntAct ) , and ( 2 ) many of the included interactions are based on high-throughput assays , including yeast two-hybrid experiments , which suffer from both false negatives and false positives [84] . Consequently , we do not perform rigorous analysis specifically for bacterial-interacting proteins , as has been done for PPIPs and VIPs [31] . Rather , we use IntAct data on bacterial interactions only to classify PPIPs as 'multi-pathogen' or not ( Fig 5 ) . PPIPs were compared to other sets of genes using a permutation test of the mean . That is , the mean value for PPIPs was compared to the mean value of many sets ( 1 , 000–10 , 000 ) of control genes . P-values were defined as the fraction of permutations where the control mean was more extreme ( usually , higher than ) than the PPIP mean . PPIPs differ from other genes in a number of ways ( S1 Fig ) . In order to evaluate PPIPs against a fair background , sets of control genes were selected that were matched to PPIPs by important functional and evolutionary metrics ( S2 Fig ) . This matched permutation approach allows the evolutionary effects of interacting with Plasmodium or Piroplasms to be isolated from correlated factors that may also influence evolutionary rate . For the analyses shown in Figs 4 , 5 and 7 , each PPIP was matched to a set of control proteins based on similarity in five metrics: mRNA expression , protein expression , protein-protein interactions , DNaseI density , and conserved element density . A control protein was considered a PPIP 'match' if each of its five values fell within a given range , based on the PPIP values ( S9 Table ) . For example , margins of min = 0 . 1 and max = 0 . 2 for mRNA expression would mean that , for a control protein to be matched to a PPIP , the mRNA expression of the control must fall between 90–120% of the mRNA expression of the PPIP . The goal was to maximize the number of matched controls per PPIP while creating control sets that were statistically indistinguishable from PPIPs for all five metrics ( S2 Fig ) . To achieve this balance , maximum margins were iteratively chosen that yielded average p-values for all metrics of at least 0 . 1 over 100 permutations . Once appropriate margins were found , matched control sets of equal size to the PPIP set were obtained by randomly sampling one matched control protein for each PPIP . Margins for the main permutation test ( Fig 4 ) are given in S9 Table . For subsets of PPIPs ( e . g . Fig 5A ) , the margins were altered to generate well-matched controls in every case . About 90% of PPIPs were typically matched and the rest excluded . Sets of matched controls were chosen based on the distribution of PPIP values included in each test , so whether any given PPIP was matched depended on the other PPIPs in the test ( i . e . , one extreme PPIP may or may not be balanced out by another ) . Therefore , the sum of matched PPIPs across categories differs slightly from the total . Notably , the pool of immune controls is relatively small ( 966 genes ) compared to the pool of non-immune controls ( 7 , 548 genes ) ( S3 Table ) . This made it difficult to match immune PPIPs to immune controls without discarding many immune PPIPs . Consequently , to test hypotheses of faster immune adaptation , we compared all PPIPs to all controls and non-immune PPIPs to non-immune controls ( Fig 5 ) . The codeml model m8 from the PAML package [85] was used to estimate dN/dS for each gene across 24 mammal species ( Fig 4B ) . However , branch-site tests in PAML rely on assumptions that may be violated in the case of recurrent adaptation to a pervasive selective pressure ( see [31] ) . Consequently , we also implemented maximum-likelihood branch-site tests in the better-performing HYPHY package [44] . The BUSTED algorithm [45] was used to detect overall evidence of positive selection at any branch in the mammalian tree , and BS-REL was used to estimate the proportion of positively selected codons in each gene on each branch . Both of these algorithms rely on the same underlying codon model; details of the model are described in [44 , 45] and reviewed in [31] . Unless otherwise specified ( i . e . , Fig 4F ) , codons identified by BS-REL were 'counted' as adaptive if the BUSTED p-value for that gene was ≤0 . 05 . We note that we did not employ the classical McDonald-Kreitman ( MK ) test to test for adaptation across multiple branches of the mammalian tree . The MK test estimates the proportion of adaptive substitutions in a protein for a single lineage , based on polymorphism within that lineage and comparison to an outgroup [86] . Here , our questions concern the evolution of proteins in multiple lineages , many of which lack polymorphism data . The methods in HYPHY are designed to simultaneously test for adaptation in multiple lineages , explicitly within the context of their phylogenetic relationships , based on a single sequence from each species . Therefore , BS-REL , BUSTED , and MEME are more powerful and appropriate for our data and questions than the MK test . We split the mammal-wide alignments for each gene into four non-overlapping alignments corresponding to the following clades: primates ( human , chimpanzee , gorilla , orangutan , gibbon , macaque , baboon , marmoset , bush baby ) , rodents ( mouse , rat , guinea pig , squirrel , rabbit ) , carnivores ( panda , ferret , dog , cat ) , and artiodactyls ( sheep , cow , pig ) . We excluded microbat , elephant , and horse , as these species are not closely related to any of the four major groups [15] ( Fig 1 ) . However , we included rabbit with rodents because they are more closely related . We ran BUSTED on each alignment to yield a p-value of clade-specific adaptation for each gene . PPIPs were matched to controls as described above ( Methods , Permutation Tests ) . However , rather than counting BS-REL adaptive codons in all branches if the tree-wide BUSTED p≤0 . 05 , we ( 1 ) kept each clade codon count separate , ( 2 ) counted codons only on branches within a clade , and ( 3 ) counted codons only if the clade-specific BUSTED p≤0 . 05 . An attempt was made to examine clade-specific adaptation in Plasmodium-only and Piroplasm PPIPs separately . However , down-sampling PPIPs to the numbers actually present in each subset resulted in dramatically increased variance in the estimates of adaptation , which eliminated statistical power to distinguish between PPIPs and controls . Alpha-spectrin homologs were initially identified in 88 mammal species using NCBI Gene ( http://www . ncbi . nlm . nih . gov/gene/ ? Term=ortholog_gene_6708 ) . The sequence of the longest mRNA transcript for each species was downloaded using E-Utilities , and each transcript was trimmed to the longest ORF using TransDecoder [87] ( http://transdecoder . github . io/ ) . Coding sequences with <50% of the human CDS length were removed . The remaining 85 coding sequences were aligned with PRANK [88] using default settings ( S7 Table ) . The alignment was manually inspected and corrected using JalView [89] . A phylogenetic tree for the 85 species was also obtained from phyloT ( http://phylot . biobyte . de/ ) using NCBI Taxonomy ( S10 Table ) . To analyze positive selection in specific domains of alpha-spectrin , we employed the HyPhy test MEME [61] rather than BS-REL . For a given gene , BS-REL estimates a proportion of codons under positive selection on each branch of a phylogeny , but does not specifically identify the adaptive codons . In contrast , MEME tests each individual codon for positive selection across all branches . MEME also estimates a probability of adaptation for that codon on each branch . Unlike BS-REL , then , MEME is capable of identifying specific codons that have evolved adaptively . We used the domain designations from SMART [90] ( http://smart . embl-heidelberg . de/ ) to assign 92 . 2% of SPTA1 codons to one of 25 domains ( S8 Table ) . Then , for each domain , we calculated an 'adaptation score' as: a/v where a measures adaptation ( the proportion of codons within the domain with MEME p≤0 . 01* ) and v measures variability ( the proportion of codons within the domain that vary among species , i . e . , are not 100% conserved ) . This score controls for domain length as well as the presence of invariable sites , as both components represent proportions of codons within the domain . To calculate the significance of each domain's adaptation score ( i . e . , to ask , is it higher than expected ? ) , we randomly permuted codons among domains 10 , 000 times . *We also tested MEME p-value cutoffs of 0 . 1 , 0 . 5 , 0 . 005 , and 0 . 001 for defining a; these results are available in S8 Table . The results for p≤0 . 01 , which are reported in the main text , are representative across these cutoffs .
|
Malaria caused by the parasite Plasmodium falciparum remains the third-most deadly infectious disease of humans . Over the last 75 , 000 years , partial genetic resistance to malaria has evolved several times , earning malaria the title of "one of the strongest selective forces on the human genome . " Yet , these human adaptations are only the most recent maneuvers in an ancient evolutionary war between host and parasite . Relatives of Plasmodium infect a variety of mammalian species today , and these large groups of hosts and parasites have likely coevolved for over 100 million years . Here , we identify 490 host genes that have been experimentally linked to the outcome of parasite infection in mammals , representing approximately 5% of all conserved mammalian proteins . In many cases , these proteins have also been linked to viral or bacterial infections . We show that parasite-interacting proteins have experienced ~3 times more adaptive substitutions than expected over mammalian evolution , and that blood parasites have left their own significant mark on our ancestors' genomes . We also identify one target of long-term adaptation to Plasmodium—the red blood cell protein alpha-spectrin—that may be involved in human susceptibility to malaria , demonstrating the value of considering modern disease in its long-term evolutionary context .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"parasite",
"groups",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"plasmodium",
"parasite",
"evolution",
"tropical",
"diseases",
"vertebrates",
"parasitic",
"diseases",
"animals",
"mammals",
"parasitology",
"parasitic",
"protozoans",
"apicomplexa",
"protozoans",
"evolutionary",
"adaptation",
"malarial",
"parasites",
"pathogenesis",
"eukaryota",
"host-pathogen",
"interactions",
"biology",
"and",
"life",
"sciences",
"malaria",
"evolutionary",
"biology",
"amniotes",
"evolutionary",
"processes",
"organisms"
] |
2017
|
High rate of adaptation of mammalian proteins that interact with Plasmodium and related parasites
|
The molecular triggers for axon degeneration remain unknown . We identify endogenous Nmnat2 as a labile axon survival factor whose constant replenishment by anterograde axonal transport is a limiting factor for axon survival . Specific depletion of Nmnat2 is sufficient to induce Wallerian-like degeneration of uninjured axons which endogenous Nmnat1 and Nmnat3 cannot prevent . Nmnat2 is by far the most labile Nmnat isoform and is depleted in distal stumps of injured neurites before Wallerian degeneration begins . Nmnat2 turnover is equally rapid in injured WldS neurites , despite delayed neurite degeneration , showing it is not a consequence of degeneration and also that WldS does not stabilize Nmnat2 . Depletion of Nmnat2 below a threshold level is necessary for axon degeneration since exogenous Nmnat2 can protect injured neurites when expressed at high enough levels to overcome its short half-life . Furthermore , proteasome inhibition slows both Nmnat2 turnover and neurite degeneration . We conclude that endogenous Nmnat2 prevents spontaneous degeneration of healthy axons and propose that , when present , the more long-lived , functionally related WldS protein substitutes for Nmnat2 loss after axon injury . Endogenous Nmnat2 represents an exciting new therapeutic target for axonal disorders .
The endogenous molecular trigger for Wallerian degeneration remains unknown . Recent progress towards understanding how the slow Wallerian degeneration fusion protein ( WldS ) delays degeneration of injured and sick axons has not addressed this wider question [1]–[8] , and this aberrant protein is only expressed in a few strains of mouse , rat , and fly . Knowledge of the normal regulation of axon survival in wild-type animals should not only lead to greater mechanistic insight but could also have important therapeutic implications for axon protection since pharmacological manipulation of endogenous processes is likely to be more achievable than overexpression of exogenous proteins . Many stresses that induce Wallerian or Wallerian-like degeneration involve a partial or complete block of axonal transport . Since transport is bi-directional , degeneration could be triggered by failed anterograde delivery of essential survival factors or by failed removal of harmful substances by retrograde transport . Defective anterograde transport seems more directly associated with axon loss than dysfunctional retrograde transport [9]–[13] . Therefore , extending an old model [14] , we propose a “survival factor delivery hypothesis” of axon degeneration . We suggest that axon integrity requires continuous anterograde delivery of one or more labile , cell body–synthesized survival factors . Other axonal components should be dispensable , more long-lived , or synthesized locally . Once supply is disrupted , following injury or other insult , levels of the limiting survival factor ( s ) will drop below a critical threshold due to natural turnover , activating an intrinsic axon degeneration program . This model has several attractions . First , the initial latent phase of Wallerian degeneration [14] , [15] would reflect the rate of survival factor turnover before the critical threshold is reached . Second , altered turnover would explain how low temperature and proteasome inhibition extend this latent phase [16]–[18] . Finally , redistribution of the remaining survival factor ( s ) in the distal stump by axonal transport could underlie the progressive nature of Wallerian degeneration [14] , [19] , [20] . No such endogenous survival factor has been identified , but the WldS protective mechanism offers important clues . WldS contains the N-terminal 70 amino acids of multiubiquitination factor Ube4b fused , in frame , to NAD+ synthesizing enzyme Nmnat1 [21] . Both regions are required for full WldS function in vivo [4] , [5] . The N-terminal VCP binding region probably targets the essential Nmnat activity to a specific subcellular location [2] , [5] . Despite being predominantly nuclear [6] , recent studies indicate a cytoplasmic and potentially axonal site of action for WldS [2] , [3] , [7] , [8] , rekindling interest in its relationship to the earlier model of a putative endogenous survival factor ( s ) [14] . WldS , an aberrant protein , cannot be one of these factors , but Nmnat1 [22] and the other mammalian Nmnat isoforms ( Nmnat2 and Nmnat3 [23] , [24] ) are candidates since they all possess the same critical enzyme activity and can all delay axon degeneration in primary neuronal culture when expressed exogenously at high levels [1] , [25] , [26] . Only Nmnat3 has so far been shown to confer robust protection to injured axons in vivo when wild-type proteins ( except for a tag used for detection ) are overexpressed [2] , [4] , [8] . In support of the “survival factor delivery hypothesis” we show that briefly suppressing protein synthesis in cell bodies of uninjured primary neuronal cultures induces Wallerian-like degeneration . The ability of a single protein ( WldS ) to block this suggests that only one or a few critical factors are directly involved . We hypothesized that WldS substitutes for one or more mammalian Nmnat isoforms , so we compared their properties against those predicted for a critical axon survival factor . We reasoned that depletion should trigger Wallerian-like degeneration without injury , its natural half-life should be consistent with the latent phase of Wallerian degeneration ( WldS should be much more stable to extend this period ) , the survival factor should be degraded by the proteasome to explain why proteasome inhibition extends axonal survival , it should be present in axons , and it should significantly prolong injured axon survival when highly overexpressed ( to outweigh its short half-life ) . Nmnat2 uniquely fits this profile , indicating that its depletion after injury is a trigger for Wallerian degeneration and that “dying-back” pathology is likely to reflect defects in Nmnat2 axonal transport or synthesis .
Our main hypothesis predicts that blocking synthesis of one or more putative axon survival factors should trigger Wallerian-like degeneration without injury , similar to that induced by blocking axonal transport [27] , [28] . To test this we initially inhibited all protein translation in mouse superior cervical ganglia ( SCG ) explant cultures , using two unrelated inhibitors , cycloheximide ( CHX ) and emetine , to rule out nonspecific effects . One µg/ml CHX , which suppresses global protein synthesis by more than 95% [29] , [30] , not only stopped neurite outgrowth as expected [30] , [31] but also induced widespread blebbing of distal neurites ( Figure 1A and 1C ) . Ten µg/ml CHX or 10 µM emetine caused more rapid and extensive blebbing of neurites , presumably due to more complete suppression of protein synthesis , followed by fragmentation and detachment shortly afterwards ( Figure 1A and 1C ) , similar to the degeneration of transected neurites . To test whether the degeneration is Wallerian-like , we used cultures from slow Wallerian degeneration ( WldS ) mice and found a delay of over 48 h ( Figure 1B and 1C ) . Similar results with rat SCG cultures and mouse dorsal root ganglion ( DRG ) cultures indicate that these events are not restricted to one species or neuron type ( Figure S1 ) . Delayed degeneration in WldS cultures after inhibition of translation also shows that local translation of mRNAs in neurites is unlikely to underlie WldS-mediated axon protection as hypothesized previously [32] . Similarly , localized translation is not required in injured neurites for WldS-mediated protection , and it is also not needed for Wallerian degeneration itself ( Figure S2 ) . Rapid cleavage of neurofilament heavy chain ( NF-H ) is an early molecular change that occurs as injury-induced Wallerian degeneration is initiated after the latent phase both in vitro and in vivo [6] , [18] . We found that this also occurs after protein synthesis suppression in wild-type cultures but not in WldS cultures ( Figure 1D ) . Thus , molecular assays also indicate this degeneration is Wallerian-like . Importantly , degeneration induced by protein synthesis suppression is not due to loss of neuronal viability but is a much earlier event independent of cell death . Even 7 d after treatment with 1 µg/ml CHX , long after complete degeneration of neurites , many SCG cell bodies retain the ability to re-grow neurites when this reversible inhibitor is removed ( Figure 1E ) . Most cell bodies in 7-d CHX-treated dissociated cultures also excluded Trypan Blue , further indicating neuron viability ( unpublished data ) . To test directly whether a critical axon survival factor ( s ) has to be synthesized and delivered from cell bodies , we used compartmented cultures where distal neurites can be treated separately from cell bodies and proximal neurites ( Figure 2 ) . Neurites degenerated only when inhibitors were applied to the compartment containing neuronal cell bodies and proximal neurites . Consistent with a previous report [33] , translation inhibitors applied only to distal neurites caused no significant degeneration within this timeframe . Indeed , neurites continued to grow ( unpublished data ) . Thus , suppression of protein synthesis in the cell body triggers Wallerian-like neurite degeneration , providing strong support for the survival factor delivery hypothesis and suggesting the survival factor ( s ) is proteinaceous . We then investigated the molecular basis of these findings . Because Nmnat1 contributes essential Nmnat enzyme activity to the WldS fusion protein , we reasoned that WldS might protect axons by substituting for injury-induced loss of an endogenous Nmnat activity . Transcripts of all three mammalian Nmnat isoforms are expressed in mouse SCG neurons ( Figure S3 and [26] ) , suggesting each is a reasonable candidate . Moreover , although their predominant localizations are nuclear ( Nmnat1 ) , Golgi-associated ( Nmnat2 ) and mitochondrial ( Nmnat3 ) [34] , the recent finding that WldS acts at a non-nuclear site despite its nuclear abundance [3] reminds us that low levels of protein can act elsewhere , especially if enzyme activity amplifies the effect . We therefore decided to test whether any of the Nmnat isoforms possess the predicted properties of an endogenous axon survival factor in our model . The first key prediction is that survival factor depletion should induce Wallerian-like neurite degeneration without injury as levels drop below a critical threshold . We used pools of siRNAs ( siNmnat1 , 2 , or 3 ) to knock down expression of the murine Nmnat isoforms and confirmed specificity for the appropriate isoform by assessing their ability to prevent expression of N-terminal FLAG-tagged Nmnat ( FLAG-Nmnat ) proteins in transfected HEK 293T cells and SCG neurons ( Figure 3 ) . To assess the effect of Nmnat isoform knock-down in SCG neurons , we used a microinjection-based strategy ( see Figure S4 ) , enabling us to consistently introduce similar amounts of siRNA . Neurons in wild-type dissociated cultures were first injected with each siRNA pool , with DsRed2 expression allowing visualization of injected neurons and their neurites . Of the three Nmnat siRNA pools , only injection of siNmnat2 caused a significant reduction in the percentage of healthy neurites compared to the non-targeting siRNA pool ( siControl ) ( Figure 4A and 4B ) . Some of the neurites of the siNmnat2-injected neurons already appeared abnormal 24 h after injection , when the entire lengths of the DsRed2-labeled neurites could first be clearly visualized , and almost all showed abnormal morphology or had completely degenerated 72 h after injection . In contrast , injection of siControl , siNmnat1 , and siNmnat3 all caused relatively little degeneration ( Figure 4A and 4B ) , and neurites continued to grow ( unpublished data ) . Combined injection of all three Nmnat siRNA pools did not significantly accelerate neurite degeneration relative to siNmnat2 alone ( Figure 4C ) . Thus , Nmnat2 knock-down is sufficient to induce neurite degeneration , whereas knock-down of the other Nmnat isoforms has no clear effect on neurite survival . To confirm that the siNmnat2-induced neurite degeneration is Wallerian-like , we microinjected WldS neurons with siNmnat2 and found degeneration was completely blocked for at least 72 h ( Figure 4A and 4D ) . To rule out a contribution from any off-target effect of the four individual siRNAs within the siNmnat2 pool , we tested whether they could cause neurite degeneration when injected individually or in non-overlapping sub-pools ( Figure S5 ) . One siRNA alone ( J-059190-11 ) and two others in combination ( J-059190-10 and J-059190-12 ) triggered significant neurite degeneration that was similar to that induced by the complete pool . A clear combinatorial effect was also seen as J-059190-11 injected at the concentration it contributes to the siNmnat2 pool caused significantly less neurite degeneration than the pool itself . Together , these observations show that siNmnat2-induced neurite degeneration is due to knock-down of Nmnat2 . The siNmnat2-induced neurite degeneration is distinctive , characterized by the appearance of multiple neuritic DsRed2-containing swellings and a distal-to-proximal “dying-back” progression that appears to be independent of neuronal viability ( Figure 4E and 4F ) . In contrast , the small amount of background neurite degeneration seen with all the siRNA pools ( including siControl ) coincides with cell death and is faster and morphologically distinct ( Figure 4G ) . Some loss of neuronal viability occurred in these experiments , irrespective of the siRNA injected , but a small , additional decrease in neuronal viability following siNmnat2 knock-down was also apparent ( Figure S6 ) . Even though this reduction in neuronal viability , relative to siControl , was proportionately much smaller than the reduction in neurite survival ( Figure S6F ) , we sought to completely exclude the possibility that cell death might be responsible for the siNmnat2-associated neurite degeneration . We were able to almost completely eliminate neuronal cell death in the siNmnat2 injection experiments in two ways ( Figure 5 ) . First , we reduced expression of the fluorescent marker after finding that toxicity was causing the ( caspase-independent ) background cell death . Second , we found that the small siNmnat2-associated decrease in neuronal viability could be prevented by the pan-caspase inhibitor z-VAD-fmk ( Figure 5A ) , indicating that this death is caspase-dependent . Importantly , the amount of siNmnat2-induced neurite degeneration was unchanged when cell death was reduced in these ways ( compare Figure 5C and 5D to Figures 4A , 4B , and S6 ) . This clearly shows that neurite degeneration precedes any associated loss of neuron viability in these experiments . It is also consistent with WldS-mediated protection of neurites ( Figure 4D ) being able to reduce siNmnat2-associated neuronal loss to control levels ( Figure S6C ) , despite the fact that WldS cannot directly prevent neuronal cell death in SCG cultures [35] . In addition , failure of z-VAD-fmk to prevent siNmnat2-induced neurite degeneration provides further evidence that it is Wallerian-like as Wallerian degeneration has been shown to be unaffected by a range of anti-apoptotic interventions [36]–[38] . Thus , constitutive expression of endogenous Nmnat2 in SCG neurons is required to prevent spontaneous “dying-back” Wallerian-like neurite degeneration . Importantly , these data also indicate that endogenous Nmnat1 and Nmnat3 cannot compensate for loss of Nmnat2 , despite the ability of these proteins to protect injured neurites when sufficiently overexpressed [1] , [25] . In our model , axon degeneration is initiated when survival factor levels drop below a critical threshold after synthesis or delivery is blocked . If Nmnat2 depletion acts as a trigger for Wallerian degeneration , Nmnat2 half-life should be compatible with the short latent phase of 4–6 h before transected SCG neurites degenerate . WldS , on the other hand , should be more stable to directly substitute for loss of endogenous Nmnat2 . A direct comparison of the relative turnover rates of the FLAG-tagged murine Nmnat isoforms and WldS in co-transfected HEK 293T cells ( Figure 6A ) showed that FLAG-tagged Nmnat2 is turned over rapidly when protein synthesis is blocked with an in vitro half-life of less than 4 h . In contrast , there was minimal turnover of FLAG-tagged WldS , Nmnat1 , and Nmnat3 up to 72 h . Similar results were also obtained with C-terminal FLAG-tagged proteins ( unpublished data ) . We also found that proteasome inhibition with MG-132 largely prevented turnover of FLAG-tagged Nmnat2 in these cells for at least 24 h ( Figure 6B ) . Importantly , turnover of endogenous Nmnat2 in SCG explants following protein synthesis inhibition was found to be similarly rapid ( Figure 6C ) . The half-life of Nmnat2 is also consistent with the time when wild-type SCG neurites become committed to degenerate after inhibition of translation ( Figure S7A ) . Neurites exposed to CHX for just 4 h remain healthy and continue to grow for over 5 d , but they become irreversibly committed to degenerate when exposed to CHX for just 8 h , despite only minimal evidence of degeneration when CHX is removed . Intermediate treatment for 6 h gave a mixed outcome . This suggests that degeneration of these neurites can be prevented by reestablishing synthesis of the labile survival factor ( s ) providing levels have not dropped below a critical threshold . The precise threshold can only be determined when the duration of downstream events leading to activation and execution of degeneration are better understood . Importantly , WldS expression not only delays the onset of neurite degeneration following protein synthesis suppression , it also delays their commitment to degenerate at least 3-fold ( Figure S7B ) . Therefore , the half-life of Nmnat2 , but not Nmnat1 and Nmnat3 , is compatible with its turnover being a trigger for Wallerian degeneration . Furthermore , the longer half-life of WldS is consistent with it substituting for Nmnat2 loss for a prolonged period . According to our model , the putative axon survival factor should also be present in neurites under normal conditions , and its level in transected neurites should drop significantly prior to initiation of degeneration at 4–6 h . Therefore , we assessed Nmnat2 levels in neurite-only extracts from SCG explant cultures at the time of transection and 4 h afterwards when the gross morphology of the transected neurites still appears relatively normal ( Figure 7A ) . Neurite extracts contained significant amounts of Nmnat2 at the time of transection and this fell to ∼30% of steady-state levels within 4 h . Furthermore , loss of endogenous Nmnat2 occurs before cleavage of NF-H , which accompanies physical break-down of SCG neurites after injury [18] or protein synthesis suppression ( Figure 1D ) , and before β-Tubulin degradation . An increase in Nmnat2 levels in the corresponding cell body/proximal neurite extracts 4 h after separation of their transected distal neurites is also seen . This probably represents accumulation of Nmnat2 in a greatly reduced cellular volume ( see Discussion ) . Proteasome inhibition modestly extends the latent phase of Wallerian degeneration in SCG explant cultures [18] , so we tested whether this correlates with reduced turnover of endogenous Nmnat2 given that FLAG-tagged Nmnat2 is degraded via the proteasome in HEK cells ( Figure 6B ) . Neurites treated with the proteasome inhibitor MG-132 appear relatively normal 8 h after transection , with no associated NF-H cleavage , whereas untreated neurites show extensive physical and molecular signs of degeneration ( Figure 7B ) . We found that loss of Nmnat2 was also significantly reduced by MG-132 at this time ( Figure 7B ) , consistent with depletion of endogenous Nmnat2 being a critical trigger for axon degeneration . The fact that Nmnat2 turnover was not completely prevented might explain why the duration of neurite protection by MG-132 is fairly limited [18] , although prolonged proteasome inhibition is also toxic to axons [39] . Nmnat2 loss within 4 h in transected wild-type neurites seems unlikely to be a consequence of axon degeneration , as cytoskeletal proteins and neurite morphology are little altered at this time point ( Figure 7A ) . However , to rule this out conclusively , we assessed Nmnat2 turnover in transected WldS neurites ( Figure 7C ) , which do not degenerate for several days . Nmnat2 levels in WldS neurites fell with a remarkably similar time course to those in wild-type neurites . In contrast , cleavage of NF-H was prevented , showing that proteins that degrade as a consequence of degeneration are stabilized in WldS neurites . As predicted , WldS levels in neurites also remained relatively constant . Indeed , levels of WldS protein are only moderately reduced in neurites 48 h after transection ( Figure S8 ) . Thus , Nmnat2 is rapidly depleted in distal stumps of injured neurites , as a result of natural turnover rather than a consequence of degeneration . This is consistent with Nmnat2 loss triggering Wallerian degeneration . The continued presence of WldS in transected WldS neurites long after Nmnat2 is lost shows that WldS does not act by stabilizing Nmnat2 but instead supports a model in which WldS substitutes for the functionally related Nmnat2 . We also found that an Nmnat2–enhanced green fluorescent protein ( eGFP ) fusion protein localizes to SCG neurites in highly defined particles shortly after being expressed ( Video S1 and Figure 7D ) . In contrast , eGFP alone showed uniform distribution in neurites ( unpublished data ) . Particles containing Nmnat2-eGFP travel bi-directionally , but the majority move in an anterograde direction ( 72 . 2%±3 . 8% based on particle movements in 18 neurites ) . The average and maximal velocities of particles moving anterogradely ( 0 . 58±0 . 09 and 1 . 52±0 . 12 µm/sec ) and retrogradely ( 0 . 29±0 . 06 and 1 . 18±0 . 10 µm/sec ) are consistent with fast axonal transport . This indicates that Nmnat2 undergoes rapid net anterograde delivery from the cell body to neurites . This is another important prediction of our model , as rapid delivery is needed to replenish constant turnover of Nmnat2 in distal neurites ( above ) . Finally , if Nmnat2 is an endogenous axon survival factor , overexpression should protect transected neurites by preloading them with increased amounts of the protein . However , due to its relatively short half-life , protection should be highly dose-dependent and prolonged protection might only be achieved with very high levels of Nmnat2 . In contrast , relatively long-lived WldS should also confer protection at much lower levels . We tested the ability of exogenous expression of tagged Nmnat2 and WldS to protect transected neurites in a microinjection-based assay ( Figure S9 ) . Dilution of the injected construct allowed controlled amounts to be reproducibly introduced into neurons . At low vector concentration ( 1 ng/ml ) , WldS conferred robust protection to neurites for 24 h after cutting , whereas Nmnat2 provided almost no protection ( Figure 8A and 8B ) . In contrast , at 50-fold higher construct concentrations , both Nmnat2 and WldS conferred protection to almost all cut neurites at 24 h ( Figure 8A and 8B ) . Although we used identical expression cassettes to give the best chance of equal expression of the two proteins in this assay , the shorter half-life of FLAG-Nmnat2 probably manifests as a lower steady-state level at the time of cut relative to FLAG-WldS . Indeed , in transfected HEK 293T cells , we found that 2 . 5 times more FLAG-Nmnat2 construct was required to give steady-state protein levels approximately equal to FLAG-WldS ( and the other Nmnat isoforms ) . Importantly , whilst we found that injection of the FLAG-Nmnat2 construct at 2 . 5 ng/ml gave slightly increased protection 24 h after cut relative to 1 ng/ml , this was still greatly reduced protection compared to the FLAG-WldS construct at the lower concentration ( Figure 8A ) . Thus exogenous Nmnat2 only confers significant protection of cut neurites when expressed at high levels , consistent with its short half-life , whilst more stable WldS protects even at low levels .
Our results provide direct support for the hypothesis that constant delivery of a labile , cell body–synthesized survival factor is required to stop healthy mammalian axons undergoing Wallerian degeneration . Defects that prevent its delivery , including axon injury [6] , [40] , axonal transport impairment [27] , [41] , cell death [35] , or disruption of protein synthesis in the cell body ( Figures 1 and 2 ) , all trigger WldS-sensitive axon degeneration . We identify Nmnat2 as one such critical axon survival factor , required to maintain normal axon integrity and sufficient to preserve injured ones at high doses . Nmnat2 half-life , uniquely among the three Nmnat isoforms , is consistent with the timing of the latent phase of Wallerian degeneration and commitment to degenerate in primary culture . We also show for the first time that endogenous Nmnat2 is present in neurites , where levels drop rapidly after injury . Importantly , this is not a consequence of neurite degeneration but represents natural turnover prior to activation of degeneration . These findings have significant implications for our molecular understanding of Wallerian degeneration and “dying-back” axonopathies , and for the mechanism by which WldS and other Nmnat isoforms delay axon degeneration . The most compelling evidence that Nmnat2 is required for maintenance of healthy axons is our observation that siRNA-mediated knock-down of Nmnat2 alone induces neurite degeneration in the absence of injury and that this precedes any effect on neuronal viability . The initiation and progression of this degeneration is clearly slower than that caused by protein synthesis suppression , but this is consistent with the mechanisms involved . The critical rate-limiting factor following translation inhibition is protein half-life , but for siRNA-mediated knock-down additional time is needed for mRNA degradation . Pharmacological inhibition of translation could also result in more efficient and homogenous knock-down . It is also possible that depletion of other axon survival factors after global suppression of protein synthesis may contribute to this difference in timing . Nmnat1 and Nmnat3 seem unlikely to be among them in this experimental system because of their long half-lives , the absence of any clear effect of their siRNAs , and the fact that endogenous levels of both proteins cannot compensate for loss of Nmnat2 . Nmnat2 is a labile protein in HEK cells , in whole SCG explants , and in transected neurites . The rate of Nmnat2 turnover is consistent with the trigger for axon degeneration being depletion below a critical threshold . Nmnat2 falls to barely detectable levels in transected wild-type SCG neurites prior to any significant physical signs of degeneration , which suggests that the critical threshold level of Nmnat2 is quite low . However , the precise threshold level is difficult to determine because the duration of downstream steps needed to bring about degeneration is unknown . Steady-state levels of Nmnat2 in SCG neurites also seem quite low and this could account for the short latent phase between neurite transection and degeneration in these cultures . Of the three mammalian Nmnat isoforms , Nmnat2 did not initially appear the most obvious candidate for an endogenous axon survival factor , despite being the most abundantly expressed isoform in the nervous system at the mRNA level [23] , [34] . First , its predominant Golgi localization seemed inconsistent with an axonal location . However , a recent report shows axons in primary neuronal cultures contain Golgi components [42] and , as with WldS [3] , predominant localization may not reflect the site of its axon protective role . We have now clearly detected endogenous Nmnat2 in SCG neurites by immunoblotting and have shown that an Nmnat2-eGFP fusion localizes to distinct , rapidly transported particles in these neurites ( Figure 7 and Video S1 ) . It will be interesting to determine the precise nature of these particles . Second , the inability of Nmnat2 to protect 5-d lesioned axons in Drosophila , unlike the other Nmnat isoforms and WldS , initially suggested it was either ineffective or by far the least potent isoform [2] . However , it has more recently been shown that exogenous expression of Nmnat2 can protect injured mammalian axons [26] . We propose that the short half-life of Nmnat2 could provide an explanation for this discrepancy , with the degree of protection being related to the levels of Nmnat2 expression achieved in the different systems . It is also possible that some protection of lesioned Drosophila axons might be evident at a less stringent time point ( wild-type fly axons begin to degenerate just 1 d after injury ) . Thus , a short half-life , one of the most critical inherent properties of the endogenous survival factor in our model , might mask the capacity of exogenous Nmnat2 to protect in some situations . Conversely , greater stability probably underlies the ability of exogenous Nmnat1 and Nmnat3 to protect injured axons/neurites more robustly in this and other in vivo and/or in vitro situations [1] , [2] , [8] , [25] . Our data suggest that WldS may protect axons by directly substituting for loss of endogenous Nmnat2 after injury or other stresses . This is based on three principal observations . First , WldS is inherently more stable than Nmnat2 , decaying less in 48 h than Nmnat2 does in 4 h after neurite transection ( Figures 6 , 7 , and S8 ) . Importantly , continued degradation of Nmnat2 in WldS neurites rules out an alternative hypothesis , that WldS could delay axon degeneration by stabilizing Nmnat2 . Second , WldS and Nmnat2 share the same enzyme activity , which is required for their ability to protect axons [5] , [26] , [43] . Third , both are present in axons ( Figure 7 , Video S1 , and [3] ) , and the presence of WldS in microsome fractions [3] , [8] is consistent with a possible shared localization with Nmnat2 in Golgi , or Golgi-derived structures in axons [42] . The ability of exogenous nuclear Nmnat1 and mitochondrial Nmnat3 to confer axon protection in a number of situations outwardly seems to contradict the claim that Nmnat localization is actually important , but there is increasing evidence to support it . First , endogenous Nmnat1 and Nmnat3 ( which do appear to be expressed in SCG neurons; Figure S3 ) cannot compensate for loss of Nmnat2 ( Figure 4 ) , probably as a result of strict compartmentalization . Alternatively , this could simply reflect the relative contributions of each isoform to total basal Nmnat activity in these axons . Second , redistribution of predominantly nuclear WldS and Nmnat1 into the cell body and axon enhances their ability to delay Wallerian degeneration [3] , [7] , [8] . Finally , Nmnat1 and Nmnat3 only confer protection when overexpressed . This appears to be accompanied by significant mis-localization ( Figure 3B , unpublished observations , and [8] ) , which may cause a serendipitous increase in effective Nmnat levels in the relevant axonal location . The ability of barely detectable extra-nuclear WldS to protect injured WldS mouse axons suggests that minimal mis-localization of relatively stable Nmnat1 and Nmnat3 may be sufficient to confer strong protection . The absence of significant axon protection in transgenic mice expressing Nmnat1 in neurons at similar levels to WldS in WldS neurons [4] , [8] suggests either that Nmnat1 localization is more rigorously controlled in vivo or that Nmnat1 overexpression is greater in vitro . Importantly , Nmnat1 can only protect mammalian axons in vivo when specifically mutated to cause mis-localization [7] . The main known function of the mammalian Nmnat isoforms is NAD+ biosynthesis , and the ability of Nmnat1 , Nmnat2 , and WldS to delay Wallerian degeneration requires Nmnat enzyme activity [1] , [5] , [26] , [43] . NAD+ production may therefore underlie the ability of endogenous Nmnat2 to prevent spontaneous axon degeneration . However , there is much disagreement over the ability of NAD+ to protect axons directly [1] , [4] , [8] , [44] , [45] , or even its involvement at all [43] . Indeed , siRNA-mediated knock-down of Nampt , the rate-limiting enzyme upstream of Nmnat in the NAD+ salvage pathway , does not itself trigger axon degeneration despite a substantial 70%–90% reduction in NAD+ levels , leading to the suggestion that an alternative Nmnat metabolite may be involved [43] . Regarding downstream events , the rapid initiation and progression of Wallerian degeneration is more consistent with an active degeneration program than passive degeneration resulting simply from loss of an essential metabolic activity . Recently , dual leucine kinase ( DLK ) and JNK signalling have been implicated in regulating Wallerian degeneration of DRG neurites [46] . We have found the same JNK inhibitor ( SP600125 ) used in that study also significantly delays Wallerian degeneration of SCG neurites ( unpublished data ) . The localization of Nmnat2 in defined particles in axons and the role it plays in them could now be key to identifying additional downstream events . Whilst neurite degeneration in primary neuronal cultures is a useful model of in vivo axon degeneration , high levels of protein overexpression can give misleading outcomes ( discussed above ) and other differences need to be considered . For example , there is a much longer latent phase before fragmentation of axons in transected sciatic nerves in vivo ( 36–40 h [15] ) than for transected SCG neurites in culture ( ∼8 h ) . This could reflect differences in the half-life of Nmnat2 in vivo and in culture ( for which there is some precedent [47] ) , steady-state levels of Nmnat2 , or the involvement of additional factors that are more critical for axon survival in vivo . However , it would be somewhat surprising if Nmnat2 did not play a critical role in vivo based on its rapid turnover and it being required for neurite survival in vitro . Other Nmnat isoforms remain candidates in vivo , particularly Nmnat3 as its mitochondrial localization makes its presence in axons likely . Indeed , a contribution from other molecules could help to explain the longer latent phase in vivo . It will also be interesting to see whether endogenous Nmnat proteins are involved in axon survival in non-mammalian organisms such as Drosophila . Whilst loss of the single Drosophila Nmnat homolog causes degeneration of photoreceptors , this appears to be a more general effect on neuronal viability , rather than axon health , and does not require its NAD+-synthesizing activity [48] . This contrasts with the protection against axon degeneration by mammalian Nmnat isoforms and WldS , which does require enzyme activity [1] , [5] , [26] , [43] . Neuronal viability could therefore be dependent on a reported Nmnat-associated chaperone activity [49] , with axons having a more specific dependency on enzyme activity . Thus , it is possible that the small decrease in neuron survival associated with Nmnat2 knock-down in SCG neurons ( Figure 5A ) could be due to loss of chaperone activity , although it is not yet known whether Nmnat2 possesses this activity like the other mammalian isoforms [49] . Loss of Nmnat2 could also underlie “dying-back” axon degeneration in disease . Due to its rapid turnover , Nmnat2 might fail to reach distal axons in sufficient quantities when axonal transport is either pathologically compromised [12] , [27] , [50] or slows during normal ageing [51] . Impairment of protein synthesis would be predicted to have a similar outcome , which could explain axon degeneration accompanying viral infections as the cellular protein synthesis machinery is overwhelmed [52] . More subtle effects on protein synthesis resulting from cell body defects , such as vacuolization , could underlie Wallerian-like “dying-back” axon degeneration and/or neuromuscular junction loss in slowly developing , chronic diseases like ALS [53] , [54] , even in the absence of neuronal loss . Our model would additionally explain why the longest axons are often most susceptible in disease . The ability of some larger mammals to support very long axons ( up to several meters in some cases ) raises the intriguing possibilities that Nmnat2 is inherently more stable in larger species or that chaperones stabilize it during transport . We propose that increasing Nmnat2 stability or its delivery to axons could have important therapeutic implications for these and other disorders characterized by Wallerian-like degeneration . Both treatments should delay the point at which axons become committed to degenerate ( like WldS ) . Such therapies might be particularly effective when axonopathy results from a short-term impairment ( e . g . , of cell body metabolism , axonal transport , glial support , etc . ) lasting just a few hours to a few weeks . Examples include Taxol-induced neuropathy , relapsing-remitting multiple sclerosis , some viral disorders , and stroke . Axons could be saved permanently if the degeneration commitment point is delayed long enough for the causative defect to be removed or to abate naturally . Although WldS mice have already been shown to be resistant to Taxol-induced neuropathy [41] , developing therapies based on the WldS neuroprotective mechanism has been limited by the technical challenge of introducing exogenous WldS ( or other stable Nmnat isoforms ) . In contrast , pharmacological manipulation of endogenous Nmnat2 should be more feasible . Finally , the increase in Nmnat2 levels in SCG cell bodies/proximal neurite stumps that we observed shortly after transection of their neurites is also intriguing ( Figure 7A ) . The simplest explanation is that this represents accumulation of Nmnat2 in a reduced cellular volume following neurite removal while synthesis continues at pre-injury levels . However , the possibility that this could represent a stress response cannot be completely excluded at this time , especially in light of the recent report that the Drosophila Nmnat isoform can act as a chaperone [49] . Irrespective of the mechanism involved , this increase in Nmnat2 levels might nevertheless facilitate subsequent neurite regeneration . In summary , we propose a model in which sustained expression and anterograde delivery of Nmnat2 is required to prevent activation of an intrinsic axon degeneration program . Degeneration is triggered when synthesis and/or delivery of Nmnat2 is disrupted and rapid turnover causes its level to drop below a critical threshold . We additionally propose that the relatively stable WldS fusion protein delays axon degeneration by directly substituting for loss of Nmnat2 and that localization may be an important factor . Endogenous Nmnat2 represents an exciting target for therapeutic manipulation .
Expression vectors encoding FLAG-tagged murine Nmnat isoforms and WldS were generated by amplification of the full coding region of each gene by Reverse Transcriptase PCR ( RT-PCR ) ( see below ) from 1 µg total RNA from wild-type and WldS mouse brain . Products were cloned into pCMV Tag-2B ( Stratagene ) to generate FLAG-Nmnat/WldS expression vectors or pEGFP-N1 ( BD Biosciences Clontech ) to generate a Nmnat2-eGFP expression vector . Sequencing ( Cogenics ) was performed to confirm the absence of PCR errors . Other plasmids used were pDsRed2-N1 for expression of variant Discosoma red fluorescent protein ( DsRed2 ) and pEGFP-C1 for expression of eGFP ( both BD Biosciences Clontech ) . Dharmacon ON-TARGETplus SMART pools of siRNA ( Thermo Scientific ) specifically targeted against mouse Nmnat1 ( L-051136-01 ) , Nmnat2 ( L-059190-01 ) , or Nmnat3 ( L-051688-01 ) were used in this study . Dharmacon ON-TARGETplus siControl non-targeting siRNA pool ( D-001810-10 ) was used as a control in experiments . Each pool consists of 4 individual siRNAs . The siRNAs making up the ON-TARGETplus Nmnat2 SMART pool ( J-059190-09 , -10 , -11 , and -12 ) were also tested individually or in subpools . Total brain RNA was extracted using TRIzol reagent ( Invitrogen ) , and RNA from dissociated SCG neuronal cultures was isolated using RNeasy columns ( Qiagen ) . One µg of brain RNA and 30% of that recovered from SCG cultures was reverse transcribed into cDNA using Superscript II ( both Invitrogen ) . Control samples without reverse transcriptase were processed simultaneously to rule out DNA contamination of samples . Standard PCR amplification was performed using REDTaq DNA polymerase ( Sigma ) . Primers used for detection of Nmnat isoform transcripts in SCG neuron RNA were as follows: Nmnat1 5′-ttcaaggcctgacaacatcgc-3′ and 5′-gagcaccttcacagtctccacc-3′ , Nmnat2 5′-cagtgcgagagacctcatccc-3′ and 5′-acacatgatgagacggtgccg-3′ , Nmnat3 5′-ggtgtggaggtgtgtgacagc-3′ and 5′-gccatggccactcggtgatgg-3′ . Products were sequenced to confirm correct amplification . 1 , 000× aqueous stock solutions of emetine ( dihydrochloride hydrate ) and CHX in DMSO ( both Sigma ) were diluted 1∶1000 in culture media to give final concentrations indicated ( 1 µg/ml CHX = 3 . 5 µM ) . InSolution MG-132 ( Calbiochem ) was diluted to 20 µM . MG-132 was added to SCG explant cultures 3 h prior to neurite transection . This pre-treatment is required to see neurite protection in these cultures [18] . Media was changed once with addition of fresh inhibitors when cultures were treated for more than 5 d . CHX-containing media was completely removed and replaced with media containing no CHX in experiments involving temporary suppression of protein synthesis . Microinjection was performed on a Zeiss Axiovert 200 microscope with an Eppendorf 5171 transjector and 5246 micromanipulator system and Eppendorf Femtotips . Plasmids and siRNAs were diluted in 0 . 5× PBS and passed through a Spin-X filter ( Costar ) . The mix was injected directly into the nuclei of SCG neurons in dissociated cultures . ON-TARGETplus siRNA pools were injected at a concentration of 100 ng/µl and individual siRNAs or sub-pools as indicated in the text , pDsRed2-N1 at 50 ng/µl , pEGFP-C1 at 10 ng/µl , the Nmnat2-eGFP expression construct at 50 ng/µl , and FLAG-Nmnat/WldS expression constructs or FLAG-empty control ( pCMV Tag-2B ) at 10 ng/µl for siRNA-mediated knock-down analysis by immunostaining ( Figure 3B ) and at 1 , 2 . 5 or 50 ng/µl in neurite transection experiments ( Figure 8 and Figure S9 ) . Seventy to 150 neurons were injected per dish . Injection of relatively few neurons per dish facilitated visualization of individual labelled neurites as neurites tend to cluster together in bundles . For detection of FLAG-tagged protein expression by immunostaining , neurons were fixed with 4% paraformaldehyde ( 20 min ) , permeabilized with 1% Triton X-100 in PBS ( 10 min ) , blocked in 50% goat serum in PBS containing 1% BSA ( 30 min ) , and stained using monoclonal M2 anti-FLAG ( Sigma ) ( 1∶400 in PBS , 1% BSA for 1 h ) and an Alexa568-conjugated secondary antibody ( 1∶200 in PBS , 1% BSA for 1 h ) . Cells were mounted in Vectashield containing DAPI ( Vector Laboratories ) for counterstaining of nuclei . For comparing the quantification of neuronal viability based on gross morphology with other indicators of health ( Figure 5B ) , cultures were incubated with 1 µg/ml propidium iodide ( Invitrogen ) for 15 min and were then fixed with 4% paraformaldehyde ( 20 min ) before being mounted in Vectashield containing DAPI . Neurites were cut with a disposable scalpel roughly half-way between their cell bodies and their most distal ends . Where applicable , inhibitors of translation or vehicle ( DMSO ) were added to the media less than 10 min before transection . Uncut neurites treated with DMSO continue to grow normally ( unpublished data ) . Microinjection of a row of cell bodies in dissociated SCG cultures enabled neurites to be cut so that all injected cell bodies and their proximal neurites were located on the opposite side of the cut site to the distal stumps ( Figure S9 ) . Bright-field and fluorescence images were captured on an Olympus IX81 inverted fluorescence microscope using a Soft Imaging Systems F-View camera linked to a PC running the appropriate imaging software . Wherever possible , images of the same field of neurites or neuronal cell bodies were captured at the indicated time points after initial manipulation . Images were processed using Adobe Photoshop Elements 4 . 0 . The intensity of FLAG immunostaining relative to eGFP fluorescence in individual injected neurons ( Figure 3 ) was quantified using ImageJ software . Images were captured for analysis using identical microscope settings between samples for each channel . Time-lapse images of Nmnat2-eGFP transport were acquired 6 h after injection of the expression vector using an Olympus CellR imaging system comprising IX81 microscope linked to a Hamamatsu ORCA ER camera and a 100×1 . 45 NA apochromat objective . Cultures were maintained at 37°C in a Solent Scientific environment chamber . Wide-field epifluorescence images were captured at 2 Hz for 1 min . ImageJ software plug-ins were used for analysis of the stacks ( kymograph generation and analysis of particle velocities ) and conversion of an image stack into an annotated movie ( Video S1 ) .
|
In a normally functioning neuron , the cell body supplies the axon with materials needed to keep it healthy . This complex logistical activity breaks down completely after injury and often becomes compromised in neurodegenerative diseases , leading to degeneration of the isolated axon . Whilst there are probably many important cargoes delivered from the cell body that isolated axons cannot exist without indefinitely , proteins that are short-lived will be depleted first , so loss of these proteins is likely to act as a trigger for degeneration . Using clues from a mutant mouse whose axons are protected from such degeneration , we have identified delivery of Nmnat2 , a protein with an important enzyme activity , as a limiting factor in axon survival . Importantly , Nmnat2 is very labile and its levels decline rapidly in injured axons before they start to degenerate . Even uninjured axons degenerate in a similar way without it . These properties are consistent with loss of Nmnat2 being a natural stimulus for axon degeneration , and it might therefore be a suitable target for therapeutic intervention .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"neurological",
"disorders/peripheral",
"neuropathies",
"neuroscience/neuronal",
"signaling",
"mechanisms",
"neuroscience/neurobiology",
"of",
"disease",
"and",
"regeneration",
"neurological",
"disorders/spinal",
"disorders"
] |
2010
|
Endogenous Nmnat2 Is an Essential Survival Factor for Maintenance of Healthy Axons
|
Arthropod vectors have multiple physical and immunological barriers that impede the development and transmission of parasites to new vertebrate hosts . These barriers include the peritrophic matrix ( PM ) , a chitinous barrier that separates the blood bolus from the midgut epithelia and modulates vector-pathogens interactions . In tsetse flies , a sleeve-like PM is continuously produced by the cardia organ located at the fore- and midgut junction . African trypanosomes , Trypanosoma brucei , must bypass the PM twice; first to colonize the midgut and secondly to reach the salivary glands ( SG ) , to complete their transmission cycle in tsetse . However , not all flies with midgut infections develop mammalian transmissible SG infections—the reasons for which are unclear . Here , we used transcriptomics , microscopy and functional genomics analyses to understand the factors that regulate parasite migration from midgut to SG . In flies with midgut infections only , parasites fail to cross the PM as they are eliminated from the cardia by reactive oxygen intermediates ( ROIs ) —albeit at the expense of collateral cytotoxic damage to the cardia . In flies with midgut and SG infections , expression of genes encoding components of the PM is reduced in the cardia , and structural integrity of the PM barrier is compromised . Under these circumstances trypanosomes traverse through the newly secreted and compromised PM . The process of PM attrition that enables the parasites to re-enter into the midgut lumen is apparently mediated by components of the parasites residing in the cardia . Thus , a fine-tuned dialogue between tsetse and trypanosomes at the cardia determines the outcome of PM integrity and trypanosome transmission success .
Insects are essential vectors for the transmission of microbes that cause devastating diseases in humans and livestock . Many of these diseases lack effective vaccines and drugs for control in mammalian hosts . Hence , reduction of insect populations , as well as approaches that reduce the transmission efficiency of pathogens by insect vectors , are explored for disease control . Tsetse flies transmit African trypanosomes , which are the causative agents of human and animal African trypanosomiases . These diseases can be fatal if left untreated and inflict significant socio-economic hardship across a wide swath of sub-Saharan Africa [1 , 2] . The phenomenon of antigenic variation the parasite displays in its mammalian host has prevented the development of vaccines , and easily administered and affordable drugs are unavailable . However , tsetse population reduction can significantly curb disease , especially during times of endemicity [3 , 4] . In addition , strategies that reduce parasite transmission efficiency by the tsetse vector can prevent disease emergence . A more complete understanding of parasite-vector dynamics is essential for the development of such control methods . For transmission to new vertebrate hosts , vector-borne parasites have to first successfully colonize their respective vectors . This requires that parasites circumvent several physical and immune barriers as they progress through their development in the vector . One prominent barrier they face in the midgut is the peritrophic matrix ( PM ) , which is a chitinous , proteinaceous structure that separates the epithelia from the blood meal [5–7] . In Anopheline mosquitoes the presence of the PM benefits the vector by regulating the commensal gut microbiota and preventing pathogens from invading the hemocoel [8] . In tsetse and sand flies , the PM plays a crucial role as an infection barrier by blocking parasite development and colonization [9 , 10] . The presence of the PM can also be exploited by microbes to promote their survival in the gut lumen . The agent of Lyme disease , Borrelia burgdorferi , binds to the tick vector’s gut and exploits the PM for protection from the harmful effects of the blood-filled gut lumen [11] . Unlike vectors that produce a type I PM in response to blood feeding , tsetse’s sleeve-like type II PM is constitutively produced by the cardia organ , which is located at the junction of the fore- and midgut . Upon entering the gut lumen , long-slender bloodstream form ( BSF ) trypanosomes ( Trypanosoma brucei ) are lysed while short-slender BSFs differentiate to midgut-adapted procyclic forms ( PCF ) [12] . During these lysis and differentiation processes , BSF parasites shed their dense coat composed of Variant Surface Glycoproteins ( VSGs ) into the midgut environment [12] . These molecules are then internalized by cells in the cardia , where they transiently inhibit the production of a structurally robust PM . This process promotes infection establishment by enabling trypanosomes to traverse the PM barrier and invade the midgut ectoperitrophic space ( ES ) [9] . After entering the ES , trypanosomes face strong epithelial immune responses , which hinder parasite gut colonization success . Detection of PCF parasites in the ES induces the production of trypanocidal antimicrobial peptides [13 , 14] , reactive oxygen intermediates ( ROIs ) [15] , PGRP-LB [16] and tsetse-EP protein [17] . A combination of these immune effectors eliminates trypanosomes from the majority of flies , leaving only a small percentage of flies infected with PCF parasites in their midgut . The PCF parasites move to tsetse’s cardia where they differentiate into long and short epimastigote forms . These cells then cross the PM for the second time to enter back into the fly’s gut lumen and migrate through the foregut into the salivary glands ( SG ) for further differentiation into mammalian infective metacyclic forms [18 , 19] . Interestingly , the SG infection process , which is necessary for disease transmission , succeeds in only a subset of flies with midgut infections [20] . Even though midgut trypanosomes fail to colonize tsetse’s SG in a subset of flies , parasites persist in the midgut for the remainder of the fly’s adult life . The physiological barriers that prevent SG colonization in the subset of midgut-only infected flies remain unknown . In this study , we investigated the molecular and cellular mechanisms that prevent parasites from colonizing the SG in a subset of flies with successful midgut infections . Our results show a robust host oxidative stress response reduces parasite survival in the cardia . While preventing parasites from further development , this immune response is costly for tsetse’s cellular integrity and results in extensive damage to cardia tissues . In contrast , less cellular damage is observed in the cardia of flies with midgut parasites that give rise to SG infections . Our results indicate that the ability of the parasites to successfully bypass the PM barrier in the cardia is essential for the establishment of SG infections . We discuss the molecular interactions that regulate this complex and dynamic vector-parasite relationship in the cardia organ , an essential regulator of disease transmission .
Tsetse display strong resistance to infection with trypanosomes . By 3–6 days post acquisition ( dpa ) , parasites that have entered into the ES of the midgut are eliminated by induced vector immune responses from the majority of flies . When newly emerged Glossina morsitans adults ( termed teneral ) are provided with an infectious bloodmeal in their first feeding , midgut infection success is typically around 30–40% [21 , 22] . However , in mature adults that have had at least one prior normal bloodmeal , the infection rate is lower , with only 1–5% of flies housing midgut infections [23 , 24] . PCF parasites in susceptible flies replicate in the ES and move forward to the cardia where they differentiate into long and short epimastigote forms . About 6–10 dpa , parasites in the cardia re-enter the gut lumen and migrate through the foregut into the SG , where they differentiate to mammalian infective metacyclic forms within 20–30 days [18 , 19] . To generate a suitable sample size of infected flies for down-stream experiments , we provided three groups of teneral G . m . morsitans adult females independent blood meals containing BSF trypanosomes ( Trypanosoma brucei brucei RUMP 503 ) and obtained midgut infection rates of about 30% when microscopically analyzed 40 dpa ( Fig 1A ) . When we analyzed the SG infection status of these gut infected flies , we detected SG infections in about 65% of individuals ( Fig 1B ) . We chose the 40-day time point to accurately score midgut and SG infections , as in our experimental system and insectary environment SG infection status becomes accurately verifiable by microscopy at 30 dpa at the earliest . Hence , two forms of fly infections exist: non-permissive infections where parasites are restricted exclusively to the gut ( hereafter designated ‘inf+/-’ ) , and permissive infections where parasites are present in the gut and SGs ( hereafter designated ‘inf+/+’ ) ( Fig 1C ) . We next investigated whether parasites residing in the non-permissive ( inf+/- ) gut infections suffer a developmental bottleneck that result in the selection of trypanosomes that are incapable of progressing towards metacyclic infections in the SG . We challenged two groups of teneral adults per os with Trypanosoma brucei brucei obtained from midguts of either inf+/+ or inf+/- flies . We observed a similar proportion of inf+/+ and inf+/- phenotypes regardless of the parasite population ( inf+/+ or inf+/- ) provided for the initial infection ( Fig 1D ) . This indicates that trypanosomes in the inf+/- gut population are still developmentally competent , and can complete their cyclic development to SG metacyclics . Thus , we hypothesized that the cardia physiological environment may determine the developmental course of trypanosome infection dynamics . Our prior studies on the role of the PM during the initial parasite colonization event showed that release of VSG from the ingested BSF parasites as they differentiate to PCF cells interferes with gene expression in the cardia resulting in loss of PM integrity early in the infection process . We thus checked to ensure that parasite re-entry into the gut lumen 6–10 dpa was not due to the residual effect of BSF-shed VSG on PM integrity . To do so we analyzed the expression of PM-associated genes ( pro1 , pro2 and pro3 ) in cardia three and six days after supplementing flies with a bloodmeal containing purified VSG . Our results show that expression of the PM associated genes is significantly reduced at the day-three time point , but that their expression fully recovers by the day-six time point . These findings indicate that parasite re-entry into the gut lumen in the cardia is unlikely affected by loss of PM integrity that results from the initial VSG effects ( Fig 1E ) . To investigate the molecular aspects of the infection barriers preventing SG colonization that subsequently limit parasite transmission in the inf+/- group , we used the infection scheme described above and pooled infected cardia into inf+/- and inf+/+ groups ( n = 3 independent biological replicates per group , with ten cardia per replicate ) . For comparison , we similarly obtained dissected cardia from age-matched normal controls ( called non-inf; n = 3 independent biological replicates per group , with ten cardia per replicate ) . We next applied high-throughput RNA-sequencing ( RNA-seq ) to profile gene expression in the three groups of cardia . We obtained on average > 23M reads for each of the nine libraries , with 77 . 8% ( non-inf ) , 75 . 4% ( inf+/- ) and 64 . 5% ( inf+/+ ) of the total reads mapping to Glossina morsitans morsitans transcriptome ( S1 Fig ) . The trypanosome reads corresponded to about 4 . 3% in non-permissive ( inf+/- ) dataset and about 15% in the permissive ( inf+/+ ) dataset ( Fig 2A ) . To estimate relative parasite densities in the two cardia infection states , we measured the expression of the trypanosome housekeeping gene gapdh in inf+/- and inf+/+ cardia by quantitative real time PCR ( qRT-PCR ) and normalized the values using tsetse gapdh . We noted significantly higher parasite gene expression values in the inf+/+ cardia compared to the inf+/- cardia samples ( Student t-test , p = 0 . 0028; Fig 2B ) . We also confirmed that inf+/+ cardia had higher parasite density by microscopically counting trypanosome numbers in the dissected cardia organs using a hemocytometer ( S2 Fig ) . Thus , the difference in the representative parasite transcriptome reads in the two infected groups of cardia is due to an increase in the number of trypanosomes residing in the inf+/+ cardia rather than an increase in parasite transcriptional activity . Interestingly , we noted no difference in the number of trypanosomes present in inf+/- and inf+/+ midguts ( S2 Fig ) . Hence , it appears that parasite density either decreases in the cardia , or fewer parasites colonize the organ despite the fact that inf+/- and inf+/+ flies maintain similar parasite densities in their midguts . To detect the presence of parasites in the foregut , and to understand the different parasite developmental stages that could be present in the inf+/- and inf+/+ cardia , we investigated parasites by microscopy from these tissues at 40 dpa . We did not observe parasites in the foregut of inf+/- infections , confirming that they are restricted from further development in the cardia in this group of flies . In the inf+/+ state however , we noted presence of many parasites in the foregut in all examined flies . We also looked at the relative presence of the different parasite developmental forms ( short and long-epimastigote and trypomastigotes ) populating the two different cardia phenotypes by examining parasite morphology and the localization of the nucleus and kinetoplast , as previously described [18 , 19] . We observed that the majority of the parasites present in both cardia infection states on day 40 were trypomastigotes with fewer epimastigotes , and that no significant differences between the two cardia infection states was noted ( S1 Dataset ) . Tsetse’s cardia is composed of several different cell-types with potentially varying functions ( schematically shown in Figs 2C and S3 ) [25–29] . These include an invagination of cells originating from the foregut , which are enclosed within an annular pad of columnar epithelial cells originating from the midgut . The cells occupying this pad secrete vacuoles that deliver components of the PM [27 , 29] . The cardia organ is surrounded by muscles that form a sphincter around the foregut , which likely regulates blood flow during the feeding process . Additionally , large lipid-containing cells are localized under a layer of muscle below the sphincter . The function of these cells remains unclear . Microscopy analysis of infected cardia supported our previous molecular findings , as we observed fewer parasites in the cardia of inf+/- ( Fig 2D and 2E ) when compared to inf+/+ flies ( Fig 2F and 2G ) . Parasites from the inf+/- cardia were restricted to the ES , whereas parasites were observed in both the ES and the lumen of inf+/+ cardia . Hence , the parasite populations resident in inf+/+ cardia had translocated from ES to the lumen , while parasites in inf+/- cardia failed to bypass the PM barrier . These data suggest that the cardia physiological environment may influence the parasite infection phenotype and transmission potential . For succesful transmission to the next mammalian host , trypanosomes that reside in the ES of the midgut must traverse the PM barrier a second time to re-enter into the gut lumen , move forward through the foregut and mouthparts and colonize the SGs . Traversing the PM a second time is thought to occur near the cardia region [25 , 29–31] due to newly synthesized PM likely providing a less robust barrier than in the midgut region . We investigated whether the functional integrity of the PM in the two different infection states varied in the cardia organ . We mined the non-inf cardia transcriptome dataset ( S2 Dataset ) and identified 14 transcripts associated with PM structure and function [6 , 9 , 32] , which cumulatively accounted for 35 . 7% of the total number of reads based on CPM value ( Fig 3A ) . The same set of genes represented 26 . 5% and 34 . 5% of the inf+/+ and inf+/- transcriptome data sets , respectively ( Fig 3A ) . Thus , PM-associated transcripts are less abundant in the inf+/+ cardia relative to inf+/- and control cardia . We next evaluated the expression profile of PM-associated transcripts and identified those that are differentially expressed ( DE ) with a fold-change of ≥1 . 5 in at least one infection state compared to the control ( non-inf ) ( Fig 3B ) . We observed a significant reduction in cardia transcripts encoding the major PM-associated proventriculin genes ( pro2 , pro3 ) in the cardia inf+/+ , but not the cardia inf+/- dataset . Both Pro2 and Pro3 are proteinaceous components of the PM [6] . Interestingly , the expression of chitinase was induced in both inf+/- and inf+/+ datasets . Because Chitinase activity can degrade the chitin backbone of the PM , increased levels of its expression would enhance the ability of the parasites to bypass this barrier . Overall , the inf+/+ cardia expression profile we observed here is similar to the profile noted in the cardia 72 hours post BSF parasite acquisition early in the infection process [9] . Results from that study demonstrated that reduced expression of genes that encode prominent PM associated proteins compromised PM integrity , thus increasing the midgut parasite infection prevalence [9] . Loss of PM integrity in the inf+/+ state could similarly enhance the ability of parasites to traverse the PM to re-enter the gut lumen and invade the SGs . We hypothesized that PM integrity is a prominent factor in the ability of trypanosomes to traverse the barrier in the cardia and continue their migration to the SGs . We addressed this hypothesis by experimentally compromising the structural integrity of the PM in flies that harbored established gut parasite infections . We modified a dsRNA feeding procedure that targets tsetse chitin synthase ( dsRNA-cs ) , which effectively inhibits the production of a structurally robust PM [7] . We challenged flies with BSF trypanosomes as teneral adults and then administered blood meals containing dsRNA-cs on day 6 , 8 and 10 post parasite acquisition . This is the time interval when we expect the parasites colonizing the ES of the midgut to bypass the PM barrier in the cardia to re-enter into the lumen [19 , 33 , 34] . Control groups similarly received dsRNA targeting green fluorescent protein ( dsRNA-gfp ) . Decreased expression of chitin synthase in the experimental dsRNA-cs group relative to the control dsRNA-gfp group was confirmed by qPCR analysis ( S4 Fig ) . Twenty days post dsRNA treatment , midguts and SGs were microscopically dissected and the SG infection status scored . We detected SG infections in 68% of dsRNA-cs treatment group compared to 47% in dsRNA-gfp control group ( Fig 3C ) . Thus , the PM compromised group of flies showed a significant increase in inf+/+ phenotype relative to the control group ( GLM , Wald-test , p = 0 . 0154 ) . These findings suggest that compromising the PM structure later in the infection process increases the proportion of gut infected flies that give rise to mature SG infections ( inf+/+ ) . Thus , tsetse’s PM acts as a barrier for parasite translocation from the ES to the lumen of the midgut , an essential step for successful SG colonization . We sought to determine if components of inf+/+ parasites infecting the cardia could manipulate cardia physiology to bypass the PM . For this , we used a modified version of a host microbial survival assay that was successfully used to evaluate PM structural integrity [7 , 9 , 35] . In this assay , tsetse with an intact PM fail to immunologically detect the presence of the entomopathogenic Serratia marcescens in the gut lumen . The bacteria thus proliferate uncontrolled in this environment , translocate into the hemocoel and kill the fly [7] . Conversely , when PM structure is compromised , the fly’s immune system can detect the presence of Serratia early during the infection process and express robust antimicrobial immunity that limits pathogen replication and increases host survival [7] . We provided mature adults blood meals supplemented with both entomopathogenic Serratia and heat-treated inf+/+ cardia extracts , while two age-matched control groups received either both Serratia and cardia extracts prepared from flies that had cleared their midgut infections ( designated rec-/- for "recovered" ) or only Serratia ( control ) . We found that survival of flies that received inf+/+ extracts was significantly higher than either of the two control groups ( Fig 3D ) . These findings suggest that cardia inf+/+ extracts contain molecule ( s ) that negatively influence tsetse‘s PM integrity , thereby enabling these flies to more rapidly detect Serratia and express heightened immune responses to overcome this pathogen . Our transcriptional investigation indicated that PM associated gene expression decreased in the inf+/+ state but not in the inf+/- state ( Fig 3B ) . In the survival assay we described above , we fed flies with cardia extracts from inf+/- containing heat-killed parasites at an equivalent quantity as the one used in the inf+/+ state . When supplemented with the cardia+/- extracts , the survival of flies was decreased to the same level as the two controls , suggesting that extract from cardia inf+/- did not compromise PM integrity ( Fig 3D ) . Collectively , these findings confirm that the parasites in cardia inf+/- differ in their ability to interfere with PM integrity when compared with those in the cardia inf+/+ state . This suggests that parasites in inf+/+ cardia display a different molecular dialogue with tsetse vector tissues . To understand the cardia-trypanosome interactions , we investigated the parasite populations in the inf+/+ state by transmission electron microscopy ( TEM ) analysis . We observed that trypanosomes aggregate in the annular cleft formed between the foregut and the midgut parts of the cardia where PM components are synthesized ( Figs 4A and S5 ) . Tsetse's PM is composed of three layers; a thin layer that is electron-dense when observed with TEM , a thick layer that is electron-lucent when observed with TEM and a third layer that is not distinguishable when observed with TEM [36] . The newly synthesized PM in the annular cleft is formed by secretions from the annular pad of epithelial cells [27 , 29] , hence lacking the typical electron-dense and electron-lucent layers observed in the fully formed PM in the midgut ( Fig 4B and 4C ) . From the six inf+/+ cardia analyzed by TEM , we observed trypanosomes embedded in the newly secreted PM as well as present in the lumen . In fact , we had shown above that the expression of putative PM-components decreased in cardia inf+/+ ( Fig 3B ) , and the structural integrity of the PM is compromised based on the Serratia detection assay ( Fig 3D ) . Thus , the structurally weakened PM could enable the trypanosomes to bypass this barrier in inf+/+ flies . Our EM observations ( all six inf+/+ cardia ) also showed that parasites assemble into compact masses ( similar to the previously reported "cyst-like" bodies [33] ) in between the layers of the PM ( Figs 4C and S6 ) . In three of the six infected inf+/+ cardia analyzed , we noted that the electron-dense layer of the PM restricting a cyst-like body appeared disrupted , which could enable the entrapped parasites to escape the barrier ( S7 Fig ) . The parasite aggregates we observed in the cardia near the site of PM secretion could represent a social behavior that influences cardia-trypanosome interactions and ultimately parasite transmission success . In vitro , trypanosomes are capable of displaying a similar social behavior termed ‘social motility’ ( SoMo ) [37] . In this situation early-stage PCF parasites ( similar to the forms that colonize the fly midgut ) cluster and migrate together on semi-agar plates [38] . In the tsetse vector , phases where trypanosomes group in clusters and move in synchrony have been observed during the infection process independent of the developmental stage of the parasite [34] . Furthermore , parasites co-localize in the cardia near the cells that produce the PM [34] , similar to our EM observations . By forming aggregates , trypanosomes could enhance their ability to resist adverse host immune responses and/or escape the ES by crossing through the newly secreted layers of the PM . In addition , the parasites can also actively compromise the PM integrity at this site , as suggested by the PM integrity assay ( Fig 3D ) , but the parasite components that interfere with host functions as such remain to be determined . We also observed extracellular vesicles associated with trypanosomes in TEM images , which could potentially carry molecules that interact with host cells or PM structure ( Fig 4 ) . To understand the parasite-PM interactions in the cardia inf+/- state , we similarly investigated the parasite populations residing in the cardia inf+/- samples by TEM analysis . We observed that parasites in cardia inf+/- are not present in the lumen and are thus unable to escape the ES where the newly synthesized PM is secreted ( Fig 5 ) . We noted that high densities of parasites are either lining along the PM secreting cells ( Fig 5A and 5B ) or are embedded in the PM secretions ( Fig 5C and 5D ) . Trypanosomes observed in this region also presented multiple vacuolation and nuclear condensation , which are indicative of cell death processes in these parasites . Contrary to the cardia inf+/+ transcriptome data , the expression of the majority of PM-associated genes in cardia inf+/- are not significantly decreased ( Fig 3B ) . Moreover , the Serratia assay we applied by co-feeding flies cardia inf+/- extracts indicated no compromise of PM integrity as this group of flies did not survive the bacterial infection ( Fig 3D ) . Thus , it appears that parasites in the cardia inf+/- are restricted by the PM to remain in the ES even at its point of secretion . Also , while cyst-like bodies were frequent in the cardia inf+/+ , only a few cyst-like bodies could be observed in cardia inf+/- . Finally , the presence of many physiologically unfit trypanosomes indicates that the inf+/- state represents a hostile environment for the parasite , restricting its survival and transmission ( Fig 5C and 5D ) . To understand the factors that can successfully inhibit parasite survival in the cardia inf +/- , we examined the inf+/- and inf+/+ cardia datasets relative to the control non-inf state for differential vector responses . We found that 25% ( 2093 ) of the total transcripts identifed were differentially expressed ( DE ) . Of the DE transcripts , 31% ( 646 ) were shared between the two cardia infection phenotypes , while 36% ( 756 ) and 33% ( 691 ) were unique to the inf +/- and inf+/+ infection phenotype , respectively ( S8 Fig ) . Of the shared DE transcripts , 89% ( 576 ) were similarly regulated between the inf+/+ and inf+/- states while 11% ( 70 ) were uniquely regulated in the two infection states . For putative functional significance , we selected transcripts presenting a fold-change of ≥2 between any comparison and a mean CPM value of ≥50 in at least one of the three cardia states . We identified 576 transcripts that were modified in the presence of trypanosomes independent of the cardia infection phenotype , hence representing the core response of the cardia against the parasite infection ( S8 Fig ) . Among these core responses were three antimicrobial peptide ( AMP ) encoding genes , including two cecropins with fold-changes of >200 ( GMOY011562 ) and >280 ( GMOY011563 ) , and attacin D , with a fold change of >27 . Production of AMPs by midgut epithelia is among the major trypanocidal responses , and the fact that both cardia inf+/+ and cardia inf+/- expressed these genes at the same level indicates that the ability of inf+/- flies to restrict trypanosomes in the ES is unlikely driven by an AMP-related immune response . We next investigated DE transcripts unique to the two infection phenotypes ( S8 Fig ) . Two putative immunity products , Immune responsive product FB49 and serpin 1 , were expressed 223 and 74 times higher , respectively , in inf+/+ compared to inf +/- cardia . Both of these products are induced upon microbial challenge in the tsetse [13 , 39] . Additionally , ferritin transcript abundance was >2 times higher in inf+/+ compared to inf +/- cardia . In the subset of transcripts specifically more abundant in the cardia inf+/- , we noted two transcripts encoding proteins involved in the circadian clock , Takeout and Circadian clock-controlled protein , which were 600 and 2 times more abundant relative to cardia non-inf , respectively . Also , transcripts encoding Kazal-type 1 protein , a protease inhibitor , and Lysozyme were more abundant in cardia inf+/- . Given that no single immune-related gene product could explain the cardia inf+/- ability to restrict trypanosomes in the ES , we chose to further evaluate the cardia cellular physiology under the inf+/+ and inf+/- infection states . To obtain a global snapshot of cardia functions that could influence parasite infection outcomes , the DE cardia transcripts between control ( non-inf ) and either cardia inf+/- or cardia inf+/+ datasets were subjected to Gene Ontology ( GO ) analysis ( using Blast2GO ) ( S3 Dataset ) . We noted 88 GO terms that were significantly down-regulated preferentially in the inf+/- state , while only 15 GO terms were significantly down-regulated in the inf+/+ state . The 88 GO terms detected in the inf+/- dataset included 5 , 11 and 67 terms associated with mitochondria , muscles and energy metabolism , respectively . To understand the physiological implications of the inf+/- infection phenotype in the cardia , we investigated the transcriptional response of the organ as well as the ultrastructural integrity of the mitochondria and muscle tissue . Gene expression patterns indicate that mitochondrial functions are significantly down-regulated in the inf+/- cardia relative to the inf+/+ state ( Fig 6A ) . More specifically , the putative proteins associated with energy metabolism , including the cytochrome c complex , the NADH-ubiquinone oxidoreductase and the ATP-synthase that function at the organelle’s inner membrane , were strongly reduced . Loss of mitochondrial integrity was further demonstrated by microscopic analysis of cardia muscle cells ( Figs 6B–6D and S9 ) and fat-containing cells ( Fig 6E–6G ) . In the cardia inf+/- phenotype , TEM observations showed mitochondrial degradation around myofibrils associated with muscle cells ( Fig 6C ) , while few such patterns were noted in the control cardia ( Fig 6B ) and cardia inf+/+ ( Fig 6D ) . The mitochondria within the lipid containing cells of both inf+/- and inf+/+ presented a disruption in the organization of their cristae , suggesting a disruption of the inner membrane ( Fig 6F and 6G ) , in support of transcriptomic level findings ( Fig 6A ) . In addition to putative mitochondrial proteins , we found that the expression of genes encoding structural proteins responsible for muscle contraction , such as myosin and troponin , is also significantly reduced upon infection , particularly in the cardia inf+/- state ( Fig 7A ) . Electron microscopy analysis also revealed a disorganization of the Z band of sarcomeres in muscle tissue surrounding the midgut epithelia in inf+/- cardia , but not in the control and inf+/+ cardia ( Fig 7B–7D ) . Extensive loss of muscle integrity was noted along the midgut epithelia in the inf+/- state . In addition , dilatation of the sarcoplasmic reticulum , muscle mitochondria swelling and vacuolation were observed , suggesting compromised muscle functions associated with this infection phenotype ( S9 Fig ) . The detrimental effects of trypanosome infection on cardia structure and function are more apparent in the inf+/- compared to inf+/+ state , despite the higher number of parasites present during the latter phenotype . Mitochondria produce reactive oxygen intermediates ( ROIs ) [40] , which in excess can damage the organelle and surrounding cellular structures [41 , 42] . The structural damage we observed in mitochondria , muscle tissue and fat cells of inf+/- cardia is symptomatic of oxidative stress [43] . Additionally , our TEM observations demonstrate that parasites in inf+/- cardia exhibit cell-death patterns such as vacuolation and swelling ( Fig 8A and 8B ) , while parasites in inf+/+ cardia appear structurally intact ( Figs 2F and 4 ) . Because ROIs modulate trypanosome infection outcomes in tsetse [15 , 44] , we hypothesized that ROIs may be responsible for controlling trypanosomes in inf+/- cardia and for producing an oxidative environment that concurrently results in tissue damage . We observed a significant increase of peroxide concentrations in both inf+/- ( 406nM; TukeyHSD posthoc test , p<0 . 0001 ) and inf+/+ ( 167nM; TukeyHSD posthoc test , p = 0 . 0008 ) cardia relative to the control cardia ( 19 nM ) , with peroxide levels significantly higher in the inf+/- state ( TukeyHSD posthoc test , p<0 . 0001 ) ( Fig 8C ) . When we experimentally decreased oxidative stress levels in infected flies by supplementing their blood meal with the anti-oxidant cysteine ( 10μM ) ( Fig 8D ) , 85% of midgut infected flies developed SG infections , while only 45% of midgut infected flies had SG infections in the absence of the antioxidant ( GLM , Wald-test p<0 . 001 ) . Our results indicate that the significantly higher levels of ROIs produced in the inf+/- cardia may restrict parasite infections at this crucial junction , while lower levels of ROIs present in the inf+/+ cardia may regulate the parasite density without impeding infection maintenance . Homeostasis of redox balance is one of the most critical factors affecting host survival during continuous host-microbe interaction in the gastrointestinal tract [45] . In the mosquito Anopheles gambiae , increased mortality is observed when ROIs are produced in response to Plasmodium berghei infections [46] . A similar trade-off expressed in the inf+/- cardia may restrict parasite infections while causing collateral damage to essential physiologies . Conversely , strong anti-parasite responses that compromise essential physiologies are absent in the cardia of the inf+/+ group , thus allowing the parasites to continue their journey to colonize the SG and successfully transmit to a new host . Additionally , flies with SG parasite infections also suffer from longer feeding times due to suppressed anti-coagulation activity in the SG , which may further help with parasite transmission in this group of flies [47] .
Trypanosome transmission by tsetse reflects a tug-of-war that begins with parasite colonization of the midgut and ends when parasites are transmitted to the next vertebrate via saliva . Initially during the infection process , BSF trypanosome products manipulate tsetse vector physiology to bypass the gut PM to colonize the midgut ES [9] . Our results show that to successfully colonize the SG , trypanosomes may again manipulate tsetse physiology to escape the midgut ES for access to the foregut , and subsequently to the SG . To re-enter the lumen , it is hypothesized that trypanosomes cross the PM in the cardia where newly synthesized PM is less structurally organized and hence can provide an easy bypass [25 , 29–31] . Here , we provide evidence in support of this hypothesis by showing that in flies where trypanosomes successfully colonize the SG , the parasites are accumulating in the region where the PM is newly secreted , and are observed both embedded in the PM secretions and free in the lumen ( summarized in Fig 9 ) . To facilitate their passage , components of trypanosomes in the cardia can apparently manipulate PM integrity by influencing the expression of PM-associated genes through molecular interference , the mechanisms of which remain to be studied . Alternatively , trypanosome-produced molecules may directly reduce the integrity of the PM as a barrier . The presence of trypanosomes in the cardia triggers immune responses which include the production of ROIs . In flies where midgut infections fail to reach the SG ( inf+/- ) , increased levels of peroxide produced in the cardia may restrict parasite survival and prevent them from further development in the fly . Given that the inf+/- phenotype is costly and leads to collateral damage in the cardia tissues of infected flies , it is possible that flies may be able to sustain this phenotype under laboratory conditions where resources are abundant for a minimal effort . It remains to be seen if the inf+/- phenotype could sustain in natural populations in the field . Because in field infection surveillance studies estimating the time of initial parasite infection acquisition is not possible , concluding the cardia infection status in natural populations is difficult . It may however be possible to initiate parasite infection experiments using field-caught teneral flies to partially evaluate the potential colony-bias that could arise under insectary conditions using fly lines that have been kept in captivity for many years . Trypanosome colonization of tsetse’s SG could represent a trade-off where vector tolerance to parasites leads to minimal self-inflicted collateral damage . Interestingly , different tsetse species may have evolved varying strategies to defend against parasitism . For instance , under similar laboratory conditions and using the same parasite strain for infection , Glossina pallidipes heavily defends against the initial infection , as the occurrence of the inf+/- phenotype in this species is rare despite similar resistance to SG transmission [48] . On the other hand , the closely related species G . morsitans , which we investigate here , has developed a different strategy to combat against parasite transmission [48] . Glossina morsitans presents a less efficient defense against the initial parasite infection in the midgut compared to G . pallidipes , but can similarly control the parasite transmission by restricting SG infections in midgut infected flies . Investigating the causes leading to this drift in strategies could lead to the development of new control strategies based on enhancing the immune defenses of the vector against parasites . Our work highlights the central role tsetse’s PM plays in parasite-vector interactions and infection outcome . This work opens up the possibility for exploiting this matrix as a target for vector control strategies to enhance its barrier function to block parasite transmission .
This work was carried out in strict accordance with the recommendations in the Office of Laboratory Animal Welfare at the National Institutes of Health and the Yale University Institu- tional Animal Care and Use Committee . The experimental protocol was reviewed and approved by the Yale University Institutional Animal Care and Use Committee ( Protocol 2014–07266 renewed on May 2017 ) . Glossina morsitans morsitans were maintained in Yale’s insectary at 24°C with 50–55% relative humidity . All flies received defibrinated bovine blood ( Hemostat Laboratories ) every 48 hours through an artificial membrane feeding system . Only female flies were used in this study . Bloodstream form Trypanosoma brucei brucei ( RUMP 503 ) were expanded in rats . Flies were infected by supplementing the first blood meal of newly eclosed flies ( teneral ) with 5x106 parasites/ml . Where mentioned , cysteine ( 10μM ) was added to the infective blood meal to increase the infection prevalence [15] . For survival assays , Serratia marcescens strain Db11 was grown overnight in LB medium . Prior to supplementation with Serratia , the blood was inactivated by heat treatment at 56°C for 1 hour as described in [7] . At day 40 post parasite challenge , all flies were dissected 48 hours after their last blood meal , and midgut and salivary glands ( SG ) were microscopically examined for infection status . Flies were classified as inf+/+ when infection was positive in both the midgut and the SG , as inf+/- when infection was positive in the midgut but negative in the SG . Cardia from inf+/+ and inf+/- flies were dissected and immediately placed in ice-cold TRIzol ( Invitrogen ) . For each infected group , inf+/+ and inf+/- , 10 cardia were pooled into one biological replicate and three independent biological replicates were obtained and stored at -80°C prior to RNA extraction . Similarly , three independent biological replicates containing 10 cardia from age-matched flies that had only received normal blood meals ( non-inf ) were prepared . Total RNA was extracted from the nine biological replicates using the Direct-zol RNA Minipreps kit ( Zymo Research ) following the manufacturer instructions , then subjected to DNase treatment using the Ambion TURBO DNA-free kit AM1907 ( Thermo Fisher Scientific ) . RNA quality was analyzed using the Agilent 2100 Bioanalyzer RNA Nano chip . mRNA libraries were prepared using the NEBNext Ultra RNA Library Prep Kit for Illumina ( New England BioLabs ) following the manufacturer recommendations . The nine libraries were sequenced ( single-end ) at the Yale Center for Genome Analysis ( YCGA ) using the HiSeq2500 system ( Illumina ) . Read files have been deposited in the NCBI BioProject database ( ID# PRJNA358388 ) . Using CLC Genomics Workbench 8 ( Qiagen ) , transcriptome reads were first trimmed and filtered to remove ambiguous nucleotides and low-quality sequences . The remaining reads were mapped to Glossina morsitans morsitans reference transcriptome GmorY1 . 5 ( VectorBase . org ) . Reads aligning uniquely to Glossina transcripts were used to calculate differential gene expression using EdgeR package in R software [49] . Significance was determined using EdgeR exact test for the negative binomial distribution , corrected with a False Discovery Rate ( FDR ) at P<0 . 05 . Identified genes were functionally annotated by BlastX , with an E-value cut-off of 1e-10 and bit score of 200 , and peptide data available from D . melanogaster database ( FlyBase . org ) . Blast2GO was utilized to identify specific gene ontology ( GO ) terms that were enriched between treatments based on a Fisher’s Exact Test [50] . Cardia tissues from three non-inf , five inf+/- and six inf+/+ 40 day-old flies were dissected in 4% paraformaldehyde ( PFA ) and placed in 2 . 5% gluteraldehyde and 2% PFA in 0 . 1M sodium cacodylate buffer pH7 . 4 for 1 hour . Observed infected cardia were obtained from two different groups of flies independently infected with trypanosomes ( n1 = 3 and n2 = 2 for inf+/-; n1 = 3 and n2 = 3 for inf+/+ ) . Tissues were processed at the Yale Center for Cellular and Molecular Imaging ( CCMI ) . Tissues were fixed in 1% osmium tetroxide , rinsed in 0 . 1M sodium cacodylate buffer and blocked and stained in 2% aqueous uranyl acetate for 1 hour . Subsequently , tissues were rinsed and dehydrated in a series in ethanol followed by embedment in resin infiltration Embed 812 ( Electron Microscopy Sciences ) and then stored overnight at 60°C . Hardened blocks were cut in sections at 60nm thickness using a Leica UltraCut UC7 . The resulting sections were collected on formvar/carbon coated grids and contrast-stained in 2% uranyl acetate and lead citrate . Five grids including two sections prepared from each different insects were observed using a FEI Tecnai Biotwin transmission electron microscope at 80Kv . Images were taken using a Morada CCD camera piloted with the iTEM ( Olympus ) software . Contrasts of the pictures were adjusted using Photoshop CC 2018 ( Adobe ) . At day 40 post parasite challenge , flies were dissected 72 hours after their last blood meal , and midgut and salivary glands ( SG ) were microscopically examined for infection status . Cardia were dissected , pooled by 5 in ice-cold TRIzol ( Invitrogen ) in function of their infection status ( inf+/+ or inf+/- ) , and then flash-frozen in liquid nitrogen . RNA was extracted using the Direct-zol RNA MiniPrep ( Zymo Research ) following the manufacturer instructions , then subjected to DNase treatment using the Ambion TURBO DNA-free kit AM1907 ( Thermo Fisher Scientific ) . 100ng of RNA was utilized to prepare cDNA using the iScript cDNA synthesis kit ( Bio-Rad ) following the manufacturer instructions . qPCR analysis was performed using SYBR Green supermix ( Bio-Rad ) and a Bio-Rad C1000 thermal cycler . Quantitative measurements were performed in duplicate for all samples . We used ATTCACGCTTTGGTTTGACC ( forward ) and GCATCCGCGTCATTCATAA ( reverse ) as primers to amplify trypanosome gapdh . We used CTGATTTCGTTGGTGATACT ( forward ) and CCAAATTCGTTGTCGTACCA ( reverse ) as primers to amplify tsetse gapdh . Relative density of parasite was inferred by normalizing trypanosome gapdh expression by tsetse gapdh expression . Statistical comparison of relative densities was performed on Prism 7 ( GraphPad software ) using a Student t-test . Direct counting of parasites was operated by dissecting the cardia and the whole remaining midgut from flies prepared similarly than above . Individual tissues were homogenized in PSG buffer ( 8 replicates for each tissue ) . Homogenate was then fixed in an equal volume of 4% PFA for 30 min . The solution was then centrifuged 15 min at 110g , the supernatant was discarded and the pellets containing the trypanosomes from cardia and midguts were suspended in 100μl and 2 , 500μl PSG buffer , respectively . Trypanosomes from the total solution were counted using a hemocytometer . Statistical comparison of numbers was performed on Prism 7 ( GraphPad software ) using a Mann-Whitney rank test . At day 40 post parasite challenge , flies were dissected 72 hours after their last blood meal , and midgut and salivary glands ( SG ) were microscopically examined for infection status . Around 40 inf+/+ and inf+/- were independently pooled together , and then roughly homogenized in 500μl of PSG buffer ( PBS+2% glucose ) . Each homogenate was centrifuged 10min at 30g to precipitate midgut debris , and then each supernatant containing parasites was transferred to a new tube to be centrifuged 15min at 110g to precipitate the parasites . Supernatants were then discarded and each pellet containing midgut procyclic trypanosomes either from inf+/+ or inf+/- flies was suspended in 500μl PSG . Parasites were counted using a hemocytometer . Newly emerged adult females were provided a blood diet including 10μM Cysteine and supplemented with 5×106 of procyclic trypanosomes from either inf+/+ or inf+/- flies prepared as described above . All flies were subsequently maintained on normal blood thereafter every 48 h . Four independent experiments were done for each type of trypanosomes . Midgut and salivary gland infections in each group were scored microscopically two weeks later . Precise sample sizes and count data are indicated in S1 Dataset . Statistical analysis was carried out using the R software for macOS ( version 3 . 3 . 2 ) . A generalized linear model ( GLM ) was generated using binomial distribution with a logit transformation of the data . The binary infection status ( inf+/+ or inf+/- ) was analyzed as a function of the origin of the procyclic trypanosomes ( inf+/+ or inf+/- ) and the experiment it belongs to . The best statistical model was searched using a backward stepwise procedure from full additive model ( i . e . parasite origin+experiment# ) testing the main effect of each categorical explanatory factor . Using the retained model , we performed a Wald test on the individual regression parameters to test their statistical difference . Precise statistical results are indicated in S1 Dataset . Cardia from inf+/+ and inf+/- flies were dissected 40 dpa . Ten organs were pooled and gently homogenized in 100μL PBS and parasite numbers were evaluated using a hemocytometer . As cardia inf+/- contain less trypanosomes than cardia inf+/+ , homogenates from cardia inf+/+ were diluted to the density of parasites present in cardia inf+/- . Equal numbers of parasites were then fixed in 2% Paraformaldehyde ( PFA ) PBS by adding an equal volume of 4% PFA PBS to the cardia inf+/+ and inf+/- homogenates . Parasites were then centrifuged for 10min at 500g and the resulting pellet was resuspended and washed in PBS . Samples were then centrifuged for 10min at 500g and the resulting pellet was resuspended in 200μL distilled water . 50μL of parasite-containing solution was deposited on poly-lysine coated slides and air dried . Slides were permeabilized for 10min in 0 . 1% Triton X-100 PBS , and then washed in PBS 5min and in distilled water 5min . Fluorescent DNA staining was then applied by covering the slides with a solution of DAPI in distilled water ( 1μg/mL ) for 20 min in the dark . Slides were subsequently washed in distilled water two times for 5 min before being air dried in the dark . Microscopic observations were realized using a Zeiss AxioVision microscope ( Zeiss ) . Detailed counts are indicated in S1 Dataset . Soluble VSG ( sVSG ) was prepared as described in [9] . Eight-day old adult flies received a blood meal containing purified sVSG ( 1μg/ml ) , or bovine serum albumin ( BSA ) ( 1μg/ml ) as a control . To assess the effect of sVSG on gene expression at three days , cardia organs were microscopically dissected at 72h post treatment . To assess the effect of sVSG on gene expression at six days , remaining flies that were not dissected at three days were given a second normal blood meal , and the cardia organs were microscopically dissected at 72h post second feeding . Five biological replicates for each treatment and each time point were generated . Five dissected cardia were pooled for each replicate and their RNA was extracted . 100ng RNA was used to generate cDNA . Quantitative real-time PCR ( qRT-PCR ) was used to evaluate the expression of the PM-associated genes proventriculin-1 , -2 and -3 as described in [9] . Normalization was performed to the internal control of gadph mRNA for each sample . Pairwise comparisons for each time point of the genes relative expression between sVSG and BSA treated flies was carried out with the Prism 7 software ( GraphPad software ) using a Student t-test . Precise statistical results are indicated in S1 Dataset . Green fluorescent protein ( gfp ) and chitin synthase ( cs ) gene specific dsRNAs were prepared as described in [7] . Newly emerged adult females were provided with a trypanosome supplemented blood diet that also included 10μM Cysteine . All flies were subsequently maintained on normal blood thereafter every 48 h . After 6 days ( at the 3rd blood meal ) , flies were divided into two treatment groups: first group received dsRNA-cs and the second group control dsRNA-gfp . The dsRNAs were administered to each group in 3 consecutive blood meals containing 3mg dsRNA/20μl blood ( the approximate volume a tsetse fly imbibes each time it feeds ) . Four independent experiments using the same pool of dsRNA were generated for each treatment . Midgut and salivary gland infections in each group were scored microscopically three weeks later . Precise sample sizes and count data are indicated in S1 Dataset . Statistical analysis on the infection outcomes following the antioxidant feeding was carried out using the R software for macOS ( version 3 . 3 . 2 ) . A generalized linear model ( GLM ) was generated using binomial distribution with a logit transformation of the data . The binary infection status ( inf+/+ or inf+/- ) was analyzed as a function of the dsRNA treatment ( dsRNA-gfp or dsRNA-cs ) and the experiment it belongs to . The best statistical model was searched using a backward stepwise procedure from full additive model ( i . e . dsRNA treatment+experiment# ) testing the main effect of each categorical explanatory factors . Using the retained model , we performed a Wald test on the individual regression parameters to test their statistical difference . Precise statistical results are indicated in S1 Dataset . Quantitative real-time PCR ( qRT-PCR ) was used to validate the effectiveness of our RNAi procedure as described in [7] . For each treatment of each experiment , we dissected the cardia of five randomly selected flies 72h after their third dsRNA-supplemented blood meal . The five dissected cardia were pooled together and their RNA was extracted . 100ng RNA was used to generate cDNA . RNA extractions from experiment #3 failed , but as the same dsRNA pools were used for all experiments and considering the consistency of the knockdown we observed , we decided to maintain experiment #3 in our counting results . To assess the PM integrity , we applied a host survival assay following per os treatment of each group with Serratia marcescens as described in [7 , 9] . We provided to three groups of 8 day-old flies ( in their 4th blood meal ) either cardia extracts obtained from challenged flies that cleared the trypanosomes and are subsequently recovered from initial infection ( rec-/- ) , or a cardia extract from inf+/- flies , or a cardia extract from inf+/+ flies . We included a fourth group of 8-day old flies that received an untreated blood meal . Cardia extract was obtained by dissecting , in PBS , the cardia from 40 days-old infected as described above . Approximately fifty cardia from either rec-/- , inf+/- or inf+/+ flies were pooled together , and then gently homogenized . Parasites were counted from the homogenates of inf+/- and inf+/+ using a hemocytometer . The three cardia homogenates were then heated at 100°C for 10 minutes . inf+/- and inf+/+ extracts were provided to reach a concentration of 5×105 parasites per ml of blood . As inf+/- cardia contain fewer parasites than inf+/+ cardia , the volume of the inf+/+ extract provided was adjusted by dilution in PSG buffer to be equal to inf+/- volume . Rec-/- extract was provided at an equal volume than infected extracts to ensure the presence of a similar quantity of extract molecules coming from the cardia in these groups . 48 hours after the flies received blood meal supplemented with the different extracts , all flies were provided a blood meal supplemented with 1 , 000 CFU/ml of S . marcescens strain Db11 . Thereafter , flies were maintained on normal blood every other day , while their mortality was recorded every day for 30 days . Precise counting data are indicated in S1 Dataset . Statistical analysis was carried out using the R software for macOS ( version 3 . 3 . 2 ) . We used an accelerated failure time model ( Weibull distribution ) where survival was analyzed as a function of the extract received ( survreg ( ) function of "survival" package ) . Pairwise tests were generated using Tukey contrasts on the survival model ( glht ( ) function of "multcomp" package ) . Precise statistical results are indicated in S1 Dataset . Newly emerged adult females were provided with a trypanosome-supplemented blood diet that also included 10μM Cysteine . All flies were subsequently maintained on normal blood thereafter every 48 h . After 10 days ( at the 5th blood meal ) , flies were divided into two treatment groups: first group received the anti-oxidant Cysteine ( 10μM ) and the second group was fed normally as a control . Cysteine was administered each blood meal until dissection . Four independent experiments were done for each treatment . Midgut and salivary gland infections in each group were scored microscopically three weeks later . Precise sample sizes and count data are indicated in S1 Dataset . Statistical analysis was carried out using the R software for macOS ( version 3 . 3 . 2 ) . A generalized linear model ( GLM ) was generated using binomial distribution with a logit transformation of the data . The binary infection status ( inf+/+ or inf+/- ) was analyzed as a function of the treatment ( control or cysteine ) and the experiment it belongs to . The best statistical model was searched using a backward stepwise procedure from full additive model ( i . e . antioxidant treatment+experiment# ) testing the main effect of each categorical explanatory factors . Using the retained model , we performed a Wald test on the individual regression parameters to test their statistical difference . Precise statistical results are indicated in S1 Dataset . ROS were quantified using the Amplex Red Hydrogen Peroxide/Peroxidase Assay Kit ( ThermoFisher Scientific ) , following the manufacturer recommendations . 40 days post parasite challenge , flies were dissected 72 hours after their last blood meal , and midgut and salivary glands ( SG ) were microscopically examined for infection status . For each infection phenotype ( i . e . inf+/+ or inf+/- ) , 3 replicates containing each 10 cardia tissues pooled and homogenized in 80μl of ice-cold Amplex Red Kit 1X Reaction Buffer were generated . Three replicates of age-matched non-infected cardia tissues were conceived in the same manner . 50μl of assay reaction mix was added to 50μl of the supernatant of each samples , and then incubated 60 minutes at RT . Fluorescence units were determined using a BioTek Synergy HT plate reader ( 530nm excitation; 590nm emission ) . Peroxide concentrations were determined using the BioTek Gen5 software calculation inferred from a standard curve ( precise results are indicated in S1 Dataset ) . Statistical analysis was performed on Prism 7 ( GraphPad software ) using a one-way ANOVA where ROS concentration was analyzed as a function of the infection status . Pairwise comparisons were carried out using a TukeyHSD posthoc test .
|
Insects are responsible for transmission of parasites that cause deadly diseases in humans and animals . Understanding the key factors that enhance or interfere with parasite transmission processes can result in new control strategies . Here , we report that a proportion of tsetse flies with African trypanosome infections in their midgut can prevent parasites from migrating to the salivary glands , albeit at the expense of collateral damage . In a subset of flies with gut infections , the parasites manipulate the integrity of a midgut barrier , called the peritrophic matrix , and reach the salivary glands for transmission to the next mammal . Either targeting parasite manipulative processes or enhancing peritrophic matrix integrity could reduce parasite transmission .
|
[
"Abstract",
"Introduction",
"Results",
"and",
"discussion",
"Conclusion",
"Methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"parasitic",
"diseases",
"parasitic",
"protozoans",
"protozoans",
"mitochondria",
"bioenergetics",
"infectious",
"disease",
"control",
"cellular",
"structures",
"and",
"organelles",
"cardia",
"digestive",
"system",
"infectious",
"diseases",
"gene",
"expression",
"exocrine",
"glands",
"gastrointestinal",
"tract",
"biochemistry",
"trypanosoma",
"eukaryota",
"blood",
"anatomy",
"cell",
"biology",
"physiology",
"genetics",
"salivary",
"glands",
"biology",
"and",
"life",
"sciences",
"energy-producing",
"organelles",
"organisms"
] |
2018
|
A fine-tuned vector-parasite dialogue in tsetse's cardia determines peritrophic matrix integrity and trypanosome transmission success
|
Immunotherapy using regulatory T cells ( Treg ) has been proposed , yet cellular and molecular mechanisms of human Tregs remain incompletely characterized . Here , we demonstrate that human Tregs promote the generation of myeloid dendritic cells ( DC ) with reduced capacity to stimulate effector T cell responses . In a model of xenogeneic graft-versus-host disease ( GVHD ) , allogeneic human DC conditioned with Tregs suppressed human T cell activation and completely abrogated posttransplant lethality . Tregs induced programmed death ligand-1 ( PD-L1 ) expression on Treg-conditioned DC; subsequently , Treg-conditioned DC induced PD-L1 expression in vivo on effector T cells . PD-L1 blockade reversed Treg-conditioned DC function in vitro and in vivo , thereby demonstrating that human Tregs can promote immune suppression via DC modulation through PD-L1 up-regulation . This identification of a human Treg downstream cellular effector ( DC ) and molecular mechanism ( PD-L1 ) will facilitate the rational design of clinical trials to modulate alloreactivity .
Regulatory T cells ( Tregs ) promote immune tolerance to self-antigens and alloantigens ( reviewed in [1] ) . Genetic deficiency of Tregs mediated by lack of Foxp3 transcription factor yields autoimmunity in mice [2] and humans [3] . Numerical or functional deficiency of Tregs in murine models exacerbates autoimmune disease [4] , [5] , predisposes to solid organ and hematopoietic stem cell graft rejection [6] , [7] , and associates with acute and chronic graft-versus-host disease ( GVHD ) [8]–[10] . Importantly , clinical studies have demonstrated Treg defects in humans with autoimmune disease [11] , [12] and GVHD [13]–[15] . Given this background , a rationale has been outlined to evaluate adoptive cell therapy using ex vivo–expanded Tregs as an approach to treat autoimmune [16] or alloimmune [17] conditions . Negative selection against the IL-7 receptor alpha chain ( CD127 ) enriches for human Tregs [18] and thereby may represent a useful tool for such cell therapy efforts; however , there are currently no reports pertaining to the regulatory function of cells expanded from CD127-depleted human T cells . Given this information , our experiments focused on human Tregs generated ex vivo by enrichment for CD127-depleted CD4+ T cells and by culture in conditions demonstrated to promote Treg expansion , including CD28 costimulation IL-2 , TGF-β [19] , and rapamycin [20] . A more comprehensive understanding of cellular and molecular mechanisms of adoptively transferred Treg products would facilitate the rational design of clinical trials evaluating Tregs . Such an understanding may be difficult to ascertain given the varieties of Tregs [21] and numerous molecular mechanisms operational in murine Treg cells , including: CTLA-4 [22] , TGF-β [23] , PD-L1 [24] , GITR [25] , or IL-10 [9] . The cellular mechanism of Tregs also is complex and varied depending on the particular experimental model; importantly , recent evidence indicates that murine Tregs inhibit responder T cells indirectly via modulation of dendritic cells ( DC ) [26] , [27] . Identification of cellular and molecular mechanisms of human Tregs , in particular ex vivo–generated Tregs , has been relatively elusive . For example , ex vivo–generated human Tregs suppressed an allogeneic mixed lymphocyte reaction ( allo-MLR ) by an undefined mechanism that operated independent of IL-10 or TGF-β [28] . Indeed , the role of antigen-presenting-cell ( APC ) modulation as a human Treg mechanism has been somewhat neglected in part because published studies have typically utilized APC-free suppressor assays . Nonetheless , one recent study determined that freshly isolated Tregs inhibited myeloid DC inflammatory cytokine secretion and costimulatory molecule expression; such Treg-conditioned DC had reduced capacity to stimulate alloreactivity in vitro [29] . In light of this relative paucity of information relating to the mechanism of ex vivo–generated human Tregs , our primary objective was to elucidate the cellular and molecular pathways associated with human Treg cell suppressor function . Because of our focus on allogeneic hematopoietic stem cell transplantation ( HSCT ) , the role of Tregs in GVHD protection , and the role of host APC for GVHD induction [30] , we elected to study human Tregs in vitro using an allo-MLR driven by a defined population of myeloid DC and in vivo using a xenogeneic GVHD ( x-GVHD ) model similar to that previously utilized to study human Tregs [31] .
Total CD4+ and CD4+CD127− T cells were costimulated and expanded in medium containing IL-2 , TGF-β1 , and rapamycin to generate control bulk “CD4” and “Treg” populations that were directly compared in each experiment . Expanded T cells maintained their CD127− status , were comparable in terms of expansion ( Figure S1A ) , coexpression of CD62L with CCR7 ( Figure S1B ( i ) ) and Foxp3 expression ( Figure S1B ( ii ) ) . Because Foxp3 is expressed in human Tregs and transiently expressed in human effector T cells [32] , we reasoned that bulk CD4 cell Foxp3 content may represent a marker of effector differentiation . To address this , we compared ex vivo–expanded T cells for simultaneous expression of Foxp3 and effector cytokines , including IL-2 ( Foxp3+IL-2+ events ) and IFN-γ ( Foxp3+IFN-γ+ events ) . Indeed , relative to Tregs , control CD4 cells had increased coexpression of Foxp3 with IL-2 ( Figure S1C ( i ) ) and Foxp3 with IFN-γ ( Figure S1C ( ii ) ) . Furthermore , relative to control CD4 cells , expanded Tregs mediated increased suppression of CD4+ and CD8+ T cell alloreactivity ( Figure S1D ( i ) and S1D ( ii ) ) ; suppression was observed at a Treg cell to responder T cell ratio of 1∶20 that approximates the physiologic ratio ( see dose-response curve , Figure S1E ) . Further experiments were performed to characterize the mechanism of immune modulation mediated by expanded Tregs generated from CD4+CD127− cells . Blockade of TGF-β , IL-10 , IDO , CTLA4 , or LAP did not abrogate Treg suppression in the allo-MLR ( unpublished data ) . However , experiments utilizing transwell plates indicated that Treg suppression in the allo-MLR was contact dependent ( unpublished data ) . Programmed death ( PD ) ligand 1 ( PD-L1 , or B7-H1 ) is expressed on DC [33] , human tumor cells [34] , and normal human tissue [35] and interacts with PD receptors on T cells to modulate the balance of tolerance and immunity ( reviewed in [36] ) . In murine systems , Treg cell expression of PD-L1 associates with suppressor function [24]; in addition , endothelial cell [37] or CD8α+ DC [38] expression of PD-L1 promotes murine Treg generation . In humans , intratumor Tregs directly inhibited responder T cell proliferation through PD-L1 [39] . Because Tregs in our experiments expressed increased PD-L1 ( Figure 1A; representative flow plot ( i ) and ( ii ) ; summary ( iii ) ) , we reasoned that Tregs might modulate DC via the PD-1 pathway . Indeed , allogeneic DC isolated from the Treg-containing MLR expressed increased PD-L1 relative to DC isolated from the standard MLR ( Figure 1B; representative plot ( i ) and ( ii ) ; summary ( iii ) ) ; remarkably , DC harvested from control CD4-containing MLR failed to up-regulate PD-L1 . Of note , Treg-conditioned DC did not have increased expression of PD-1 ( CD11c+PD-1+ cells , <1% ) . PD-L1 inhibits T cell function via the PD-1 receptor and B7-1 ( CD80 ) [40] . To determine PD-L1 binding pathways in our system , we first measured effector T cell expression of PD-1 and CD80 after incubation with three types of allogeneic myeloid DC ( control , Treg conditioned , or control CD4 conditioned ) . Effector CD4+ T cells ( Figure 1C; representative flow plot ( i ) and ( ii ) ; summary ( iii ) ) and CD8+ T cells ( Figure 1C representative flow plot ( iv ) and ( v ) ; summary ( vi ) ) up-regulated PD-1 expression , but not CD80 expression , upon exposure to Treg-conditioned DC , but not CD4-conditioned DC . We next utilized a PD-L1 fusion protein to characterize binding pathways . Using laser scanning cytometry ( LSC ) , we found that effector T cells up-regulated total PD-L1 binding partners in the presence of Treg-conditioned DC , but not CD4-conditioned DC ( Figure 1D , left panel ) ; importantly , effector T cell PD-L1 binding was abrogated by T cell preincubation with anti-PD1 , but not anti-CD80 ( Figure 1C , right panel ) . And finally , effector T cell PD-L1 binding was quantified by flow cytometry ( Figure 1E ( i ) – ( iii ) ) . Remarkably , PD-L1 binding was greatly increased on effector T cells exposed to Treg-conditioned DC ( % effector T cell PD-L1 binding increased from 7 . 3±0 . 4 to 92 . 6±2 . 8 , p = 0 . 001 ) ; similar to results using LSC , effector T cell PD-L1 binding was abrogated by T cell preincubation with anti-PD1 , but not anti-CD80 ( Figure 1E ( iv ) ) . Secondary transfer experiments were performed to evaluate whether Tregs mediated suppression in part through DC modulation ( experimental scheme , Figure 2A ) . Indeed , allogeneic DC conditioned with Tregs yielded reduced levels of CD4+ and CD8+ responder T cell proliferation relative to CD4-conditioned allogeneic DC ( representative results , Figure 2B; pooled results , Figure 2C ) . Importantly , blockade of DC expression of PD-L1 partially corrected the observed stimulatory deficit of Treg-conditioned DC on CD4+ and CD8+ T cell proliferation ( representative results , Figure 2B; pooled results , Figure 2C ) . Next , we utilized an in vivo xenogeneic transplantation model to further characterize the ability of Tregs or Treg-conditioned DC to modulate the PD1 pathway . As expected , recipients of Treg-conditioned DC , which expressed increased PD-L1 in vitro prior to adoptive transfer , had an increased in vivo number of dendritic cells in the spleen that expressed PD-L1 ( Figure 3A; representative flow plots ( i ) , ( ii ) , and ( iii ) ; summary data , 3b ( i ) ) ; relative to recipients of control DC , recipients of Treg-conditioned DC also had an increase in PD-L1–expressing DC in the bone marrow ( p = 0 . 006 ) . Remarkably , recipients of Treg-conditioned DC also had increased numbers of effector CD4+ and CD8+ T cells in the spleen that expressed PD-L1 in vivo ( Figures 3B ( ii ) and ( iii ) , respectively ) ; such recipients also had increased numbers of T cells that expressed PD-L1 in the bone marrow ( p = 0 . 003 ) . In marked contrast , recipients of control CD4-conditioned DC did not have increased responder T cell PD-L1 expression . Interestingly , recipients of Treg-conditioned DC also had increased numbers of effector CD8+ and CD4+ cells in the spleen that expressed PD-1 in vivo ( Figures 3B ( iv ) and ( v ) , respectively ) ; in the bone marrow , such recipients also had increased numbers of CD8+PD-1+ cells ( p = 0 . 02 ) and CD4+PD-1+ cells ( p = 0 . 009 ) . Further experiments were performed to assess the functional significance of this sequential increase in PD-L1 expression from Treg cell , to conditioned DC , and then to responder T cells in vivo . Recipients of Treg-conditioned DC that were incubated with anti–PD-L1 prior to adoptive transfer had lower numbers of PD-L1–expressing DC in vivo , although cohort comparisons did not reach statistical significance ( Figure 3C ( i ) ) ; a repeat experiment yielded similar findings ( unpublished data ) . Blockade of PD-L1 on Treg-conditioned DC yielded a reduction in the in vivo number of effector CD8 cells expressing PD-L1 ( Figure 3C ( ii ) ) . Blockade of PD-L1 on Treg-conditioned DC also reduced the number of PD-L1–expressing responder effector CD4+ cells in the spleen ( Figure 3C ( iii ) ) . Finally , PD-L1 blockade of Treg-conditioned DC reduced the in vivo number of effector CD8+ cells in the spleen that expressed PD-1 ( Figure 3C ( iv ) ) ; the number of CD4+PD1+ T cells in the spleen was not significantly altered by PD-L1 blockade ( Figure 3C ( v ) ) . In sum , these data indicate that PD-L1 expression on Treg-conditioned DC was functionally significant in vivo , particularly with respect to up-regulating downstream expression of PD1 and PD-L1 on effector CD4+ and CD8+ T cells . Next , we evaluated whether human Treg and Treg-conditioned DC might modulate xenogeneic GVHD in a PD-L1–dependent manner . Previous xenogeneic GVHD models have utilized human peripheral blood mononuclear cells ( PBMCs ) that contain unmanipulated human T cells [41] , [42] , or more recently , ex vivo costimulated T cells [43] . Our initial xenogeneic GVHD experiments utilized PBMC or purified lymphocytes as the human T cell inocula . However , despite following the protocol utilized by previous publications , we found an unacceptably low rate of lethal GVHD ( <10% lethality by day 45 postinfusion ) ; an inability to consistently generate lethality was associated with a low level of human T cell engraftment ( see Figure S2A ) . Subsequent experiments were designed to identify human inocula that yielded enhanced human T cell engraftment and a resultant increase in lethal xenogeneic GVHD incidence . An initial experiment found that engraftment of purified human T cells was enhanced by coinfusion of a human , but not murine , source of APC ( unpublished data ) . Based on these data , in a subsequent experiment , immune-deficient murine hosts received one of five distinct human T cell–containing inocula: ( 1 ) PBMC; ( 2 ) lymphocytes plus monocytes; ( 3 ) lymphocytes plus DC; ( 4 ) ex vivo–activated effector T cells plus monocytes; and ( 5 ) ex vivo–activated T cells plus DC . At day 30 postinfusion , recipients of the ex vivo–activated T cells plus DC had the highest levels of human CD4+ and CD8+ T cell engraftment ( Figure S2A ( i ) and ( ii ) , respectively ) ; furthermore , recipients of ex vivo–activated T cells plus DC had the highest capacity for secretion of human IFN-γ at day 30 postinfusion ( Figure S2B ) . Therefore , in order to evaluate the effects of Tregs , Treg-conditioned DC , and the PD-1 pathway in a more stringent model of xenogeneic GVHD , subsequent experiments utilized human inocula that contained ex vivo–activated T cells and DC . Further in vivo experiments were performed to evaluate the effect of Tregs and Treg-conditioned DC on human T cell engraftment , cytokine activation , and induction of lethal xenogeneic GVHD . Recipients of human inocula that contained either Tregs or Treg-conditioned DC had reduced absolute numbers of human T cells as measured in the spleen at day 45 posttransplant ( Figure 4A ) ; the absolute number of human T cells present in vivo was also reduced when the evaluation was performed in the bone marrow for recipients of both Tregs ( p = 0 . 03 ) and Treg-conditioned DC ( p = 0 . 01 ) . Tregs and Treg-conditioned DC transfer resulted in reduced absolute numbers of both human effector CD8+ and CD4+ cells ( Figure 4B; representative data ( i ) ; summation of data in ( ii ) and ( iii ) , respectively ) . Human CD4+ T cell numbers in the bone marrow was also reduced for recipients of Tregs ( p = 0 . 01 ) but not significantly reduced in recipients of Treg-conditioned DC ( p = 0 . 08 ) ; human CD8+ T cell numbers in the bone marrow were also reduced for recipients of Tregs ( p = 0 . 02 ) but not significantly reduced in recipients of Treg-conditioned DC ( p = 0 . 09 ) . Of note , recipients of Tregs , but not recipients of Treg-conditioned DC , had a statistically significant reduction in the absolute number of posttransplant CD8+Tc1 and CD4+Th1 cells capable of IFN-γ secretion ( Figure 4C ( i ) and ( ii ) , representative flow plots; 4C ( iii ) and ( iv ) , summation of data ) . In sum , these data indicated that both Treg cells and Treg-conditioned DC were capable of inhibiting human T cells in vivo , with Treg therapy manifesting more potent regulation both in terms of limiting T cell numbers and T cell effector function . Xenogeneic GVHD was evaluated by weight loss measurement , survival analysis , and histology evaluation of GVHD target tissues . Recipients of Tregs or Treg-conditioned DC were uniformly protected against lethal xenogeneic GVHD ( Figure 5A ( i ) ) ; importantly , recipients of control CD4-conditioned DC uniformly died of xenogeneic GVHD . Posttransplant weight loss , which is a more sensitive clinical parameter of xenogeneic GVHD , was moderated by Treg-conditioned DC therapy and virtually eliminated by Treg therapy ( Figure 5A ( ii ) ) . In a second experiment , we confirmed the ability of Treg-conditioned DC to completely abrogate the generation of lethal xenogeneic GVHD; importantly , protection against lethal xenogeneic GVHD conferred by the Treg-conditioned DC was completely abrogated by anti–PD-L1 , but not by isotype control antibody ( Figure 5B ) . Of note , both control DC and Treg-conditioned DC engrafted and persisted in vivo; importantly , such numbers were not substantially influenced by Treg therapy or anti-PDL1 antibody . That is , at day 25 posttransplant , the absolute numbers of CD11c+ DC per spleen ( each value , ×103; n = 5 per cohort ) in transplant recipients that received effector human T cells in combination with the indicated specific type of human DC were 136±11 ( control DC ) , 107±6 ( control DC and Treg therapy ) , 418±98 ( Treg-conditioned DC ) , 163±63 ( Treg-conditioned DC , anti–PDL1-treated ) , and 279±77 ( Treg-conditioned DC , isotype antibody treated ) ( each comparison , p = NS by ANOVA test ) . GVHD control mice uniformly developed a diffuse skin rash and hair loss; skin histology analysis documented cutaneous acanthosis and hyperkeratosis in GVHD controls , but not in Treg recipients ( representative histology; Figure 5C ( iii ) and ( iv ) , respectively ) . Furthermore , GVHD controls , but not Treg recipients , developed diffuse lymphocytic infiltration of the liver ( representative histology; Figure 5C ( i ) and ( ii ) , respectively ) .
The rational design of adoptive cell therapy protocols using ex vivo–expanded Tregs would be facilitated by an improved understanding of their cellular and molecular mechanism of action , which has been difficult to ascertain , particularly with respect to human Tregs . In this report , utilizing a novel method of generating human Tregs based on CD127 negative selection , we have elucidated a unique Treg mechanism of immune suppression analogous to previously described models of infectious tolerance [44] that is mediated at least in part by modulation of allogeneic dendritic cells through the PD-L1 pathway . This mechanistic understanding is particularly pertinent to efforts that will utilize Tregs for the prevention or treatment of GVHD , which is driven by allogeneic DC [30] and is amenable to suppression through PD-1 [45] . Our results are the first , to our knowledge , to describe a mechanism of human Treg action that involves potent in vitro and in vivo suppression of effector T cells through a secondary cellular messenger , myeloid dendritic cells . A similar biology has been described in murine models , whereby Tregs create a weak stimulator DC through induction of immunosuppressive indoleamine 2 , 3-dioxygenase ( IDO ) via a CTLA-4– or IFN-γ–dependent pathway [26] . Interestingly , the reverse biology has also been described in murine models , whereby murine plasmacytoid DC that produce IDO promote the generation of immunosuppressive Tregs that express PD-L1 [24] . Of note , in our experiments , inhibition of IDO by 1-MT treatment did not abrogate suppression mediated by Treg-conditioned myeloid DC ( unpublished data ) . Similar to a previous study using freshly isolated human Tregs [29] , we found that ex vivo–generated human Tregs inhibited myeloid DC secretion of the proinflammatory cytokines IL-6 and TNF-α ( unpublished data ) and induced a DC phenotype with greatly reduced capacity to induce responder T cell proliferation in vitro . Most importantly , we have significantly extended this prior work through our discovery that myeloid DC conditioned by Tregs were effective in vivo for the complete elimination of posttransplant lethal xenogeneic GVHD induced by effector T cells . In addition to this apparent DC-mediated mechanism of GVHD protection , other non-APC mechanisms are likely operative for the Tregs that we studied , because transplant cohorts that received Tregs had more robust protection against xenogeneic GVHD than recipients of Treg-conditioned DC ( lowest CD4+ and CD8+ T cell engraftment , lowest posttransplant IFN-γ secretion , and lowest degree of weight loss posttransplant ) . Furthermore , this is the first demonstration that human Tregs mediate immune suppression in vivo through modulation of the PD-1 pathway . First , we observed that ex vivo–expanded human Tregs expressed increased PD-L1 relative to control expanded CD4+ T cells . Second , Treg-conditioned DC expressed greatly increased PD-L1 relative to DC conditioned with control CD4 cells . As such , Treg PD-L1 appeared to directly induce DC PD-L1 expression; the potential existence of such a PD-L1 “positive feedback loop” adds to the known complexity of PD-1 pathway regulation [36] and to our knowledge has not been previously described for murine or human Tregs . Third , this feedback appeared to extend to the distal stage of effector T cell regulation because effector CD4+ and CD8+ cells under the influence of Treg-conditioned DC , but not control CD4-conditioned DC , had nearly universal expression of PD-L1 binding partners . Finally , we determined that such effector T cell binding to PD-L1 was preferentially mediated through PD-1 rather than the other receptor associated with this pathway , CD80 . It is interesting to note that a recent study found that PD-1 expression on Treg cells in patients with viral hepatitis played a negative regulatory role for Treg cell function via limitation of STAT-5 phosphorylation [46] . In our experiments , the ex vivo–activated Treg cells expressed a high level of PD-1 , yet were able to mediate potent suppression of effector T cells in vivo at relatively dilute Treg to effector T cell ratio; as such , it does not appear that the PD-1 pathway exerted a functionally significant down-regulatory effect on the Tregs utilized in our model . It is interesting to note that the Treg-conditioned DC did not express significant PD-1; it is thus possible that the capacity of this cell population to effectively prevent xenogeneic GVHD may reside in part on a limited susceptibility to PD-1–mediated suppression . Importantly , this biology was functional in vivo because: ( 1 ) Treg-conditioned DC maintained expression of PD-L1 after adoptive transfer; ( 2 ) effector ( Teff ) cells up-regulated both PD-L1 and PD1 in vivo in the presence of Treg-conditioned DC; and ( 3 ) a significant proportion of this immune modulation was abrogated if Treg-conditioned DC were blocked with anti–PD-L1 . Remarkably , the survival advantage conferred by Treg-conditioned DC was fully abrogated by anti–PD-L1 . In sum , these data demonstrate that modulation of the PD-1 pathway represents a significant mechanism of action of ex vivo–expanded Tregs that involves an interaction between Tregs , DC , and effector T cells in an apparent positive feedback loop . Further experiments will be required to better understand this process of intercellular PD-1 pathway modulation . Potentially , the PD-L1 suppressor phenotype might be transferred from Tregs to DC and then to effector T cells by a process of trogocytosis [47] , which results in the generalized transfer of cell membrane proteins , including costimulatory molecules [48] . However , because we observed that Tregs up-regulated DC expression of PD-L1 , but not other cell surface molecules such as PD-1 , we speculate that alternative mechanisms of intercellular regulation may be operational . These findings have several implications for ongoing efforts to utilize ex vivo–generated Tregs for adoptive cell therapy . First , we have found that CD127− selection represents a suitable alternative to CD25+ selection for attempts to enrich for Tregs prior to ex vivo expansion; further experiments will be required to directly compare these two methodologies to determine whether such methods result in differential modulation of APC function through the PD-1 pathway . It is perhaps important to emphasize that the regulatory T cell or Treg-conditioned DC modulation of xenogeneic GVHD was robust because it occurred at the relatively low regulatory cell to effector cell ratio of 1∶20 , which is considered to represent a physiologic ratio . Second , our demonstration that ex vivo–generated Tregs operate to a significant degree indirectly through allogeneic myeloid DC may help guide protocol design , particularly in the setting of allogeneic HSCT . One theoretical limitation to Treg cell therapy is the transfer of “contaminating” effector T cells or the conversion of Tregs to proinflammatory Th17 cells [49] that are known to induce GVHD [10] . The APC mechanism we have identified offers a solution to this potential limitation: that is , one could harvest host-type monocytes pretransplant and generate myeloid DC in a manner similar to the methods that we utilized , condition such DC with ex vivo–generated Tregs , and then transfer only the conditioned host DC prior to allogeneic HSCT . Such an approach would be analogous to that proposed for type II DC ( DC2 cells ) that promote Th2 cytokines and prevent murine GVHD [50] . Our results also indicate that the capacity of adoptively transferred Tregs to modulate GVHD may relate in part to the bioavailability of host-type myeloid DC . This consideration may have relevance to the choice of host conditioning for Treg protocols: predictably , non-myeloablative regimens may be favorable in this regard because such regimens would preserve host myeloid DC as a key secondary cellular mediator of the Treg therapy . It should be stated that xenogeneic models of GVHD likely do not fully reflect the biology of clinical GVHD , and as such , the potential clinical implications of the findings in our model must be interpreted with caution . Specifically , the xenogeneic transplantation model that we utilized did not incorporate a human hematopoietic stem cell component , and as such , the potential effect of the regulatory T cells or Treg-conditioned DC that we evaluated on stem cell engraftment was not assessed . However , we found that human dendritic cell and effector T cell engraftment was persistent at the relatively late time points of day 25 and day 45 posttransplant , respectively; thus , it is possible that human hematopoietic progenitor cell engraftment would also be durable under the conditions that we evaluated . Such a possibility would be consistent with data emanating from murine models of MHC-disparate transplantation , which have found Treg adoptive transfer to augment allogeneic hematopoietic stem cell engraftment [7] , [51] . Our findings relating to PD-1 pathway modulation may also hold clinical implications . Adoptive cell therapy using Tregs or Treg-conditioned DC may be conceptualized as a vehicle for PD-L1 delivery . Such a cell therapy approach may have immediate practical benefit for the treatment of the myriad of diseases that may benefit from PD-1 agonism [52] . That is , although an antibody-based method of PD-1 antagonism has already been investigated in phase I clinical trials [53] , it is unclear whether agonistic PD-1 antibodies will be available or safely administered . Finally , the mechanisms we have identified will provide a rationale for monitoring PD-1 and PD-L1 expression on posttransplant T cells and DC as a biological marker for in vivo activity of the administered Tregs or Treg-conditioned DC; in addition , surface PD-L1 expression may be utilized as a marker to facilitate a functionally defined release criteria for the experimental cell therapy products . In conclusion , ex vivo expansion of CD127-negatively selected CD4+ T cells yielded a human Treg product that inhibited alloreactivity in vitro and in vivo , in large part due to modulation of myeloid DC and a multifaceted promotion of the PD-1 pathway in Tregs , DC , and effector T cells . As such , we have identified two distinct cell therapy vehicles , Tregs and Treg-conditioned myeloid DC , each of which show promises as a novel approach to modulate human effector T cells through the PD-1 pathway .
Female RAG2−/−γc−/− mice were obtained from Taconic and utilized at 8–12 wk of age . Experiments were performed according to a protocol approved by the National Cancer Institute Animal Care and Use Committee . Mice were housed in a sterile facility and received sterile water and pellets . As in previously reported methods [31] , [41] , mice were injected with 0 . 1 ml of chlodronate-containing liposomes ( Encapsula Nanoscience ) for macrophage depletion and given low-dose irradiation ( 350 cGy ) . X-VIVO 20 media was obtained from BioWhitaker and AB serum was from Gem Cell . CD4 microbeads were from Miltenyi Biotec . Sheep anti-mouse ( SAM ) IgG dynabeads were from Dynal . Anti-CD3 , anti-CD28 coated tosyl-activated magnetic beads were manufactured as previously described [54] . Rapamycin was from Wyeth ( Rapamune ) . Recombinant human ( rh ) IL-2 and rhIL-12 were from PeproTech , and rhTGF-β1 , αTGF-β1 , -β2 , -β3 , and purified αPD-L1 were from R&D Systems . All other antibodies ( unless otherwise stated ) were provided by BD Biosciences; anti-human Foxp3 APC was from eBioscience . Luminex kits for detection of IFN-γ and TNF-α were from Bio-Rad . 5- ( and-6 ) -Carboxyfluorescein diacetate , succinimidyl ester [5 ( 6 ) -CFDA , SE; CFSE] was from Invitrogen . Normal donor peripheral blood cells were collected by apheresis on an IRB-approved protocol . Total lymphocytes were isolated by elutriation [55] . Total CD4+ T cells were then enriched by CD4 microbeads according to manufacturer instructions . To isolate CD127-depleted CD4+ T cells: ( 1 ) elutriated lymphocytes were adjusted to 100×106 cells/ml and incubated with anti-CD127 ( 10 µg/ml , 30 min , 4°C ) ; ( 2 ) cells were washed , mixed with SAM dynabeads ( bead∶cell ratio , 4∶1 ) , incubated ( 30 min , 4°C ) , separated ( hand-held magnet , Dynal ) ; and ( 3 ) CD127-depleted cells were subjected to CD4 cell isolation by microbeads . Total CD4+ and CD4+CD127− T cells were cultured in polystyrene tissue culture flasks ( Corning ) . Cells were activated by anti-CD3 , anti-CD28 costimulation ( bead∶cell ratio , 3∶1 ) , and cultured in X-VIVO 20 with 5% heat-inactivated ( HI ) AB serum containing rapamycin ( 1 µM ) , TGF-β1 ( 20 ng/ml ) , rhIL-2 ( 100 IU/ml ) . rhIL-2 alone was added at days 2 , 4 , and 6 . Cultures were started at 1 . 5×106 cells/ml , maintained at 1×106 cells/ml through day 7 , and then split daily to 0 . 5×106/ml by addition of IL-2 and rapamycin-replete medium through day 12 . T cells were washed with PBS supplemented with 0 . 1% BSA and 0 . 01% azide , and stained using anti-: CD4 PE-cy7 ( clone S3 . 5; Caltag ) , Foxp3 APC ( clone 249D; eBioscience ) , CCR7 PE ( clone 150503; R&D ) , CTLA-4 Biotin ( clone BN13 ) , CD27 FITC ( clone M-T271 ) , and CD62L APC-cy7 ( clone DREG-56; Biolegend ) . For intracellular ( IC ) flow cytometry , fixation and permeabilization buffer was utilized ( eBioscience ) ; four-color IC flow cytometry was performed with combinations of anti-: IL-2 biotin ( clone B33-2 ) , IFN-γ APC ( clone B27 ) , CD4 Pe-Cy5 ( clone RPA-T4 ) , and Foxp3 PE ( clone PCH101; eBioscience ) . DC were evaluated using anti-: CD80 Bio ( clone L307 . 4 ) , CD86 APC ( clone 2331 ) , CD14 PE ( clone M5E2 ) , CD83 FITC ( clone HB15e ) , CD40 APC ( clone 5C3 ) , and PDL1 PE-cy7 ( clone MIH1 ) . Monocytes from four healthy , randomly selected donors were obtained by apheresis and elutriation; HLA typing confirmed that the donors did not share major haplotypes . Each monocyte population was cultured in X-VIVO 20 medium with 5% HI-AB serum , rhGM-CSF ( 50 ng/ml ) , and rhIL-4 ( 20 ng/ml ) . On day 5 , each DC culture was enumerated and subjected to flow cytometry to document a DC phenotype ( CD14− , CD11c+ , CD83+ , CD80+ , CD86+; unpublished data ) . The four separate DC populations were pooled in equal proportions , and aliquots of the final product were cryopreserved and utilized for each experiment . Normal donor lymphocytes ( “responder T cells”; 2×105 cells ) were cocultured with allogeneic DC ( 5×104 cells ) in 96-well round-bottom plates ( T cell to DC ratio , 20∶1 ) . To detect proliferation , responder T cells were CFSE-labeled before coculture . From the same normal donors , Tregs were generated from CD4+CD127− cells , or as a control , from total CD4+ cells . Initial experiments determined that a Treg to responder T cell ratio of 1∶20 consistently yielded suppression of proliferation . Proliferation of CD4+ and CD8+ responder T cells was evaluated by CFSE dye dilution; percent suppression of CD4 and CD8 responder T cell suppression was calculated , with values representing the ratio of total divided peaks to both divided and nondivided peaks , normalized to the sham-treated experimental group . During the MLR , neutralizing antibodies were added , including anti-: CTLA-4 , IL-10 , TGF-β1 , TGF-β2 , TGF-β3 , LAP , and their respective isotope controls . All antibodies were used at 20 µg/ml . A combination of anti-CTLA-4 , anti-TGF-β , and anti-LAP was also tested . 1-methyl-d-tryptophan ( 1-MT , 1 mM; Sigma ) was utilized to inhibit indoleamine 2 , 3-dioxygenase ( IDO ) . Transwell plates with a 4-mm membrane ( Corning LifeSciences ) were utilized to assess Treg contact dependency . For the secondary transfer experiments , DC were incubated with Tregs for 48 h ( Treg to DC ratio , 1∶1 ) . Tregs were then removed using T cell–positive selection ( anti-CD3 microbeads and subsequent magnetic column separation; Miltenyi ) ; the resultant population was >99% pure for DC content , as determined by flow cytometry using CD11c in combination with CD80 , CD86 , and CD40 . Such Treg-conditioned DC were then used as stimulator cells , with degree of proliferation determined relative to DC conditioned with control T cells ( cells generated ex vivo from total CD4 cells ) or sham-treated DC . MLR assays using preconditioned DC were also performed with anti–PD-L1 ( 20 µg/ml ) or isotype control antibody . On day 5 , responder T cells were evaluated for expression of PD-L1 binding partners PD1 and CD80 . The responder T cells were blocked with a specific αPD1 ( 1 µg/1×106 cells ) and αCD80 ( 1 µg/1×106 cells ) antibody and then PD-L1 binding was studied by incubation with recombinant PD-L1-Fc fusion molecule ( R&D ) ; secondary incubation was performed with FITC-labeled rabbit anti-human IgG , Fc-fragment antibody ( Jackson Laboratory ) . Stained T cells were delivered to 96-well plates with a plastic #1 cover slip bottom ( 1×105 cells in 200 µl ) and analyzed ( iCys Laser Scanning Cytometer; Compucyte Corporation ) . Cells were scanned ( 488-nm laser ) and fluorescence was detected ( 530/30-nm band-pass filter ) . Scan images and fluorescence data were generated ( iGeneration and innovator software; Compucyte ) . Images were collected at 0 . 5-µm scan resolution . Human effector CD4+Th1/CD8+Tc1 ( Teff ) cells were generated by T cell culture for 6 d by costimulation and expansion of T cells in rhIL2 ( 20 IU ) , αIL-4 ( 100 ng/ml ) , rhIL-12 ( 20 ng/ml ) , and rapamycin ( 1 µM ) . On day 6 of culture , Teff cells were harvested and injected ( i . v . by retro-orbital method , as previously described [43] ) into Rag2−/−γc−/− mice conditioned with chlodronate and radiation; Teff cell dose was either 1 or 3×107 cells/recipient ( higher dose used for evaluation of posttransplant lethality ) . Specific cohorts additionally received ex vivo–generated Tregs ( generated from CD4+CD127− cells ) or control T cells ( generated from total CD4 cells ) at a dose of 0 . 5 or 1 . 5×106 cells/recipient such that the in vivo ratio of effector Teff cells to Tregs always matched that utilized in the allogeneic MLR ( 20∶1 ) . As indicated , cohorts additionally received pooled allogeneic DC ( complete mismatch as compared to Teff and Tregs ) utilized in the MLR ( DC dose , 0 . 5 or 1 . 5×106 cells/recipient ) to maintain constant ratios; allogeneic DC were either not conditioned or conditioned with ex vivo–generated Tregs or control T cells . For blocking experiments , conditioned DC were incubated with anti-PD-L1 ( 20 µg/ml ) or isotype control antibody prior to adoptive transfer . In some cases , anti-PD-L1 was injected following cell transfer ( i . p . ; 100 µg/recipient ) . After adoptive transfer , human engraftment was calculated using flow cytometry data from splenic single-cell suspensions ( % huCD45+ = [huCD45+ ( huCD45++ mCD45+ ) ]×100% ) . Surface or intracellular flow cytometry was performed at indicated days after adoptive transfer to assess in vivo modulation of responder human CD4 and CD8 T cells and human DC . Flow cytometry and cytokine data were analyzed using Student 2-tailed t-tests . Comparison values of p<0 . 05 were considered statistically significant . Survival was determined using Kaplan-Meier test . For three pairwise cohort comparisons , statistical analyses was performed using the Holm method [56] .
|
Graft-versus-host disease ( GVHD ) is the most serious complication of bone marrow transplants between individuals ( so-called allogenic transplants ) . The class of suppressor immune cells called regulatory T cells ( Tregs ) inhibit GVHD by dampening the effects of donor immune cells in the grafted tissue . The cellular and molecular mechanisms involved in this process have not been fully characterized , particularly for human cells . In this study , we report that human Tregs , which we generated from precursor cells ex vivo , express high levels of a cell surface protein called PD-L1 ( programmed death ligand-1 ) that is known to mediate immune suppression . Coculture of these Tregs with allogeneic antigen-presenting cells ( APCs ) , which are known to initiate GVHD , increased , in turn , the amount of PD-L1 on the APCs . The Treg-conditioned APCs were then less able than unconditioned APCs to provoke GVHD in a mouse model of the condition , preventing the death of the animals after transplantation . We found that an antibody against PD-L1 blocked the immunosuppressive effects of Tregs or Treg-conditioned APCs , indicating that this protein is an important part of the molecular mechanism . These findings are potentially important for attempts to modulate immune responses in disease by transplanting T cells into patients .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"immunology/immunomodulation",
"hematology/bone",
"marrow",
"and",
"stem",
"cell",
"transplantation",
"immunology/immune",
"response"
] |
2010
|
Regulatory T Cells and Human Myeloid Dendritic Cells Promote Tolerance via Programmed Death Ligand-1
|
The earliest immune responses activated in acute human immunodeficiency virus type 1 infection ( AHI ) exert a critical influence on subsequent virus spread or containment . During this time frame , components of the innate immune system such as macrophages and DCs , NK cells , β-defensins , complement and other anti-microbial factors , which have all been implicated in modulating HIV infection , may play particularly important roles . A proteomics-based screen was performed on a cohort from whom samples were available at time points prior to the earliest positive HIV detection . The ability of selected factors found to be elevated in the plasma during AHI to inhibit HIV-1 replication was analyzed using in vitro PBMC and DC infection models . Analysis of unique plasma donor panels spanning the eclipse and viral expansion phases revealed very early alterations in plasma proteins in AHI . Induction of acute phase protein serum amyloid A ( A-SAA ) occurred as early as 5–7 days prior to the first detection of plasma viral RNA , considerably prior to any elevation in systemic cytokine levels . Furthermore , a proteolytic fragment of alpha–1-antitrypsin ( AAT ) , termed virus inhibitory peptide ( VIRIP ) , was observed in plasma coincident with viremia . Both A-SAA and VIRIP have anti-viral activity in vitro and quantitation of their plasma levels indicated that circulating concentrations are likely to be within the range of their inhibitory activity . Our results provide evidence for a first wave of host anti-viral defense occurring in the eclipse phase of AHI prior to systemic activation of other immune responses . Insights gained into the mechanism of action of acute-phase reactants and other innate molecules against HIV and how they are induced could be exploited for the future development of more efficient prophylactic vaccine strategies .
Although human immunodeficiency virus type 1 ( HIV-1 ) induces a chronic infection ultimately culminating in the development of an acquired immunodeficiency syndrome , it is now recognized that critical damage to the host immune system is mediated during the acute phase of infection , when an exponential burst of viral replication takes place , associated with massive depletion of the central memory CD4+ T cell pool [1] , [2] . Prophylactic strategies to combat HIV-1 infection thus need to modulate events in the earliest stages of infection leading up to and impacting on this acute viral burst – which prompts an urgent need to understand the virus-host interactions occurring during this “window of opportunity” . Adaptive responses are known to play an important role in containment of the acute burst of viral replication in AHI [3] , but events in the earliest stages of infection are also likely to be heavily influenced by components of the innate immune system [3] . These include cellular determinants of the efficiency of viral entry into and replication within host cells such as the CCR5-delta32 allele , CCR2 , CCL5 ( RANTES ) , CX ( 3 ) CR1 , CXCL12 , or TRIM5 , all of which can influence host resistance or susceptibility to HIV infection [3] . Interaction of virions with dendritic cells ( DCs ) early after virus transmission can have outcomes including virion destruction ( following binding to langerin on resting Langerhans cells ) , efficient viral transmission to CD4+ T cells ( following binding to DC-SIGN on sub-epithelial DCs ) , or the triggering of DCs ( via interaction with TLRs ) to produce cytokines/chemokines that may mediate antiviral activity , but may also drive immunopathological immune activation including cellular apoptosis [4] , [5] . Genetic studies have helped to cast light on the in vivo importance of certain components of the innate immune system in acute/early HIV infection . These include associations between expression of certain KIRs and their cognate HLA alleles and resistance to , and/or control of HIV replication , implicating NK cells in control of HIV replication [6] , [7] , [8] . Furthermore , β-defensins , secreted from oral and mucosal epithelial cells appear to inhibit HIV-1 infection [9] . More recently , a peptide fragment derived from alpha–1 antitrypsin ( AAT ) , a serine protease inhibitor and acute phase protein present in blood plasma , was shown to inhibit HIV host cell infection by blocking gp41 mediated cell entry [10] . Other natural factors exist that modulate HIV infection , such as a proteolytic product of the prostate phosphatase that is present in semen , which has the ability to dramatically enhance HIV infection [11] . Much of our current picture of events in the eclipse and earliest viremic phases of acute HIV-1 infection is derived from in vitro studies and work carried out in non-human primate simian immunodeficiency virus ( SIV ) infection models , as the critical initial stages of infection are very difficult to study in humans . The availability of plasma sample series collected over a time-frame spanning the eclipse and viral expansion phase of HIV infection provide a unique opportunity to gain insight into the systemic activation of immune responses during this time . Previous reports have quantified an array of cytokines and markers of apoptosis in plasma panels and described a massive systemic “cytokine storm” occurring during the viral ramp-up phase , associated with an increase in plasma levels of apoptotic microparticles [12] , [13] , [14] . Importantly however , no systemic elevation in apoptosis markers or cytokine levels was detected during the eclipse phase when virus is being amplified at local infection sites prior to systemic dissemination . In this study , we used a proteomics-based approach combined with biochemical and cell biological assays to characterize factors that are elevated in plasma during the earliest stages of acute HIV-1 infection in humans . We describe increases in plasma levels of acute-phase reactants and proteolytically processed fragments that have anti-HIV activity during the eclipse phase prior to detection of HIV viremia or the first increases in systemic cytokine levels , which may represent the earliest systemic host antiviral response activated following infection .
Samples collected at sequential time points spanning the eclipse and viral ramp-up phases from 19 US plasma donors who acquired HIV-1 infection were studied to gain insight into the kinetics of the earliest systemic anti-viral defenses activated in the acute phase of infection . Plasma was typically obtained from each donor at intervals of 2–5 d . Plasma panels were tested for HIV-1 by RT-PCR analysis of viral RNA titers , and time courses from different donors were aligned relative to the time point ( T0 ) when viremia first reached levels detectable by conventional assays ( >100 RNA copies ml−1; Fig . 1A ) . Most panels covered a time-frame from around d −20 to d +20 relative to T0 . It is currently thought that the eclipse phase in HIV-1 infection is in the range of 7–10 d [12] , [15] hence most panels likely included samples collected from time points prior to the acquisition of infection onwards . In order to determine whether there are detectable changes in plasma proteins or peptides accompanying the emergence of viremia , an initial mass spectrometry-based screen was performed on three plasma donor panels ( Fig . 1B ) . Analysis of the longitudinal MALDI-TOF data revealed mass peaks that were elevated at viremic time points ( Fig . 2A ) . One mass peak with a molecular mass of 2178 Dalton [M+H]+ was found to be considerably elevated in HIV-1-positive plasma ( Fig . 2A ) . Sequencing by MALDI-TOF/TOF and LC-MS/MS identified this mass as peptide 86–105 derived from A-SAA ( Fig . 2B and Fig . S1A ) . A semi-quantitative analysis of mass peak intensities of the 2178Da [M+H]+ peptide mass revealed that this peptide was elevated coincident with the increase in viremia , and in 2 of the 3 subjects , immediately prior to the detection of viremia ( Fig . 2C ) , suggesting that A-SAA protein levels are elevated at these times . A second mass peak with a molecular mass of 2213 Dalton [M+H]+ was identified as peptide 960–979 of complement C3 ( Fig . S1B ) . This peak was also elevated prior to as well as during viremia ( Fig . 2C ) . A-SAA was shown previously to be elevated in patients with AIDS [16] , and is commonly used as a general marker for inflammation [17] , [18] . A recent study demonstrated that A-SAA has anti-viral activity in vitro [19] . We therefore examined a larger set of plasma donor panels ( 19 ) by ELISA to test whether elevation of A-SAA levels may be a general feature associated with acute HIV-1 infection , and how its induction is related to the increase in plasma viral RNA titers . As a control and to establish a baseline for use in statistical analysis of the data , we also measured A-SAA levels in plasma panels from five control plasma donors who did not become infected with HIV ( Fig . S2A ) . Baseline levels of A-SAA ( calculated as described in the methods section ) varied between individuals and were generally between 600–3800ng/ml . Analysis of A-SAA levels in the plasma panels from HIV-infected donors confirmed that A-SAA was elevated relative to baseline prior to and/or concurrent within the earliest detection of viremia . Importantly , significant A-SAA elevations ( i . e . falling above a 90% prediction interval ) were observed prior to T0 ( viral RNA >100 copies ml−1 ) in 15 out of 19 subjects ( Fig . 3A ) . In the subject group as a whole , A-SAA levels were thus elevated significantly prior to T0 , the time of first detection of plasma viremia ( p = 0 . 02 , as determined using a Binomial test ) . To monitor alterations of the acute form ( A-SAA ) as well as the constitutively expressed form ( C-SAA ) of serum amyloid A , we performed immunoblot assays with specific antibodies ( Fig . 3B ) . In 4 out of 10 plasma donor panels tested by immunoblotting ( 9011 , 9013 , 9016 and 9018 ) , we confirmed the initial increase in A-SAA levels prior to viral ramp-up . In addition to this first wave of A-SAA induction that occurred prior to detection of viremia , immunoblotting also confirmed a second , more intense phase of elevation , observed in six of the ten donors tested ( 9011 , 9012 , 9016 , 9018 , 64012 , 64122 ) that coincided with the increase in viral load . Other individuals either had very low or consistently high levels of A-SAA over the time-frame analyzed . By contrast to A-SAA , a major acute phase reactant that is inducible during the infection process , C-SAA , which was observed in an unmodified and glycosylated form , was detected with only minor inductions in all panels examined . To further explore whether elevation of A-SAA was specifically attributable to HIV-1 infection , we examined 6 plasma panels from donors who became positive either for hepatitis C virus ( HCV ) or hepatitis B virus ( HBV ) during the period of sample collection [14] . We detected increased levels of A-SAA over the time course of infection in 1/3 panels from subjects with acute HBV infection and 3/3 panels from subjects with acute HCV infection ( Fig . 3C and Fig . S2 B and C ) , indicating that A-SAA induction may represent a common host response to microbial infection . A more comprehensive analysis by liquid chromatography tandem mass spectrometry ( LC-MS/MS ) revealed a number of other plasma components including complement factors , apo-lipoproteins and alpha-1-antitrypsin ( AAT ) that are present in plasma during AHI ( Table S1 ) . The factors found to be elevated in plasma during AHI included a C-terminal peptide derived from AAT , residues 377–396 , referred to as VIRIP ( Table S1 and Fig . 4A ) , which was shown to inhibit HIV-1 entry into host cells by targeting the gp41 fusion peptide [10] . Seven plasma panels from HIV-infected individuals and five panels from uninfected controls were evaluated for the presence of VIRIP by tandem mass spectrometry in a semi-quantitative fashion . Ion counts detected for the precursor ion representing the expected molecular mass of VIRIP were correlated to viremia , and revealed an elevation of VIRIP coincident with and after the initial increase in viremia in two of the seven plasma donor panels from infected individuals , but none in the five controls ( Fig . 4B ) . A semi-quantitative titration of VIRIP peptide by mass spectrometry indicated that the amount detected corresponds to an estimated value of 0 . 1–0 . 3µM of VIRIP in plasma at peak concentrations ( Fig . 4C ) . Considering the sample loss during the isolation of VIRIP peptide from plasma , the effective VIRIP concentration will likely be in the range of low µg/ml , which is approximating the IC50 value at which VIRIP interferes with HIV-1 entry [10] ( see also below ) . The appearance of VIRIP in plasma samples from HIV-1-infected subjects raised the question of how proteolytic processing of AAT may be mediated under these conditions . Previous studies indicated that MMPs ( collagenases and elastases ) interact with and cleave AAT [20] . Inspection of the regions of AAT flanking the VIRIP sequence predicted cleavage sites for MMP-1 , -6 , -7 , -8 , -9 , -12 , and MMP-26 at the N-terminal , and MMP-7 at the C-terminal end of VIRIP ( Fig . 5A ) . In vitro digestion of purified AAT with recombinant MMP-7 revealed a degradation product at 5 kDa detectable by immunoblotting consistent with the C-terminal AAT fragment 377–418 containing VIRIP ( Fig . 5B ) . Subsequent analysis was carried out using LC-MS/MS , which identified peptides containing both cleavage sites required for the formation of VIRIP at Phe376-Leu377 and Phe396-Leu397 ( Fig . 5C ) . A third cleavage within the VIRIP sequence at Pro381-Met382 was also observed . Importantly , formation of VIRIP itself was detected at the 2 h time point , confirming MMP-7 as a candidate protease involved in generation of the peptide in vivo . In order to test for potential interference with the infection process we evaluated the ability of acute phase proteins AAT , A-SAA and C-reactive protein ( CRP ) and the C-terminal AAT fragments 397–418 and 377–396 ( VIRIP ) to inhibit HIV-1 replication using in vitro peripheral blood mononuclear cell ( PBMC ) and dendritic cell ( DC ) infection models . PHA-activated PBMCs or monocyte-derived dendritic cells ( MDDCs ) were incubated with the test analytes both prior to and during infection with either an R5- or an X4-tropic virus and subsequent HIV-1 replication was monitored by the analysis of supernatant p24 levels or reverse transcriptase activity . VIRIP , derived from near the C-terminus of AAT , markedly inhibited the replication of both the R5 and the X4 virus , consistent with previous findings [10] ( Fig . 6A ) . In contrast , no effect was observed with either full-length AAT or the C-terminal 22mer 398–418 ( Fig . 6A ) . In addition , CRP did not inhibit the replication of either the R5 or the X4 virus in PBMCs ( Fig . 6C ) . A-SAA did not exhibit any inhibitory activity in the PBMC infection system , but did inhibit the replication of the R5 virus in MDDCs as early as 12–24 h after infection ( Fig . S3 ) and to a greater extent after 7 d ( Fig . 6B ) . Importantly , inhibition of MDDC infection was still greater than 50% at a 1µgml−1 A-SAA concentration , which is well in the range of the levels detected in infected individuals ( Fig . 3A ) . A-SAA was recently reported to inhibit MDDC infection by an X4/R5 dual tropic virus via down-regulation of CCR5 expression [19] . We conclude that components of the acute phase response indeed have the capacity to interfere with HIV-1 infection , thereby potentially helping to control viral dissemination in the eclipse phase .
The availability of sequential samples from plasma donors who became infected with HIV-1 provides a unique opportunity to study changes in plasma in the eclipse and viral expansion phases of acute infection . Recent studies described the induction of a “cytokine storm” and massive cellular apoptosis during the phase of exponential viral replication [12] , [13] , [14] . Here a proteomics-driven approach was used to demonstrate for the first time that acute phase proteins , some of which exhibit antiviral activity , are induced systemically even prior to the first detection of viremia and also before any detectable increase in plasma cytokine levels . The factors elevated included A-SAA , a protein primarily synthesized by cells in the liver , high-level production of which is known to be induced during the acute phase response to infection , trauma or stress by pro-inflammatory cytokines including TNF-alpha , IL-6 and IL-22 [21] . Notably , A-SAA was frequently elevated with biphasic kinetics in the plasma of subjects acquiring HIV-1 infection , an initial elevation occurring during the eclipse phase and a second elevation during the viral ramp-up phase . The latter was temporally coincident with elevations in circulating levels of multiple pro-inflammatory cytokines [14] and likely reflected a hepatic response to this systemic stimulus . The initial elevation in plasma A-SAA levels was found to occur significantly prior to detection of viral RNA in the plasma , well before any systemic elevations are detected in plasma cytokine/chemokine levels [14] . Following sexual transmission of HIV-1 , virus replicates in the local genital or rectal mucosa , then spreads to the draining lymph nodes and subsequently to the gut-associated lymphoid tissue ( GALT ) [22] . The mechanism by which acute phase protein production is triggered during this process is currently unknown , but it may involve transfer of inductive factors to the liver ( including pro-inflammatory cytokines produced at local sites of viral replication ) , triggering production of acute phase reactants prior to widespread virus dissemination and systemic increases in cytokine levels . Alternatively , A-SAA can be produced at extra hepatic sites: A-SAA expression has been reported in macrophages , adrenal glands , kidney and intestine , albeit at lower levels as compared to hepatocytes [17] . The initial burst of A-SAA levels may also contribute to the subsequent “cytokine storm” observed later at the systemic level , since A-SAA was shown to induce an array of immunomodulatory cytokines including of the Th1-type in monocytes , macrophages and lymphocytes [23] , [24] . Activation of acute phase reactants may represent a very early line of anti-viral defense in HIV-1 infection , since A-SAA , AAT and a C-terminal peptide derived from AAT ( referred to as VIRIP ) were each shown to exert anti-viral activity in vitro [10] , [19] , [25] , [26] ( Fig . 6A , B ) . However , not all acute phase proteins that are induced in response to inflammation have anti-viral properties as was observed with CRP ( Fig . 6C ) , which was also elevated systemically in plasma of some donors prior to viremia ( data not shown ) . We found selective inhibition of HIV-1 infection of MDDCs but not PBMCs by A-SAA . Its ability to inhibit HIV-1 replication was previously proposed to be mediated by down-regulation of CCR5 expression [19] . The capacity to inhibit R5 virus replication in MDDCs but not PBMCs may thus be due to the fact that MDDCs express only low levels of CCR5 , whereas CCR5 expression on CD4+ T cells is much higher . It is unlikely that the inhibitory effects of A-SAA are limited to specific pathogens only , as this acute phase reactant was described to be up-regulated in a number of pathological processes [17] including inflammation and diverse bacterial and viral infections [27] , [28] , [29] . Consistent with this , we observed marked elevation of A-SAA in plasma panels from donors acutely-infected with HCV . Notably , A-SAA has been shown to mediate antiviral activity against this viral pathogen too , an effect proposed to be mediated by mechanisms distinct from its inhibitory effect on HIV-1 infection [30] , [31] . In acute HCV infection , A-SAA induction may be stimulated as a consequence of viral replication in the liver , and the associated cytokine response . In acute HBV infection little A-SAA elevation was observed . This may relate to the different replication kinetics of HBV as compared to HIV-1 and HCV , and/or the relatively muted cytokine response activated in acute HBV infection [14] . Genetic studies suggest that polymorphisms observed in AAT , another component of the acute phase response , are linked to susceptibility to HIV-1 infection [32] , [33] . AAT was initially reported to block the activity of HIV-1 protease [26] , [34] , [35] . Subsequently , an AAT-derived 26mer C-terminal peptide was shown to inhibit HIV-1 LTR gene expression in vitro [36] , [37] . However , the strongest effect reported so far is exerted by VIRIP generated as a proteolytic product from AAT , which inhibits viral entry by binding to the gp41 fusion peptide and is active in the low micromolar range [10] . For both A-SAA and VIRIP , we were able to detect endogenous levels in plasma that are in the range capable of mediating viral inhibition , which raises the possibility that such factors play a role in combating viral replication , particularly in the earliest stages of infection . No strong correlation was found between the initial timing of A-SAA elevation and either R0 ( the viral reproductive rate ) , the slope of viral ramp-up or the highest recorded viral load in the 19 subjects studied here . Nevertheless , the relationship between the magnitude and dynamics of early acute-phase protein production and the acute viral burst and subsequent efficiency of control of viremia should be addressed in a future study on a larger cohort from whom samples were collected over a longer time-frame extending into early infection . Interestingly , we also noted that levels of AAT proteolytic fragments in plasma from subjects chronically-infected with HIV-1 were elevated as compared to those in healthy controls ( data not shown ) , suggesting that acute phase proteins may also play a role in chronic viral infection . It is possible that components of the acute phase response may not only contribute to the control of viral replication in infected individuals , but may also be involved in mediating resistance to infection . For instance , HIV-1-exposed uninfected individuals have been shown to have elevated levels of cleaved forms of A-SAA [19] , which suggests that their anti-viral activity contributes to resistance against infection . Evidence from the literature and our own data suggest that MMPs are responsible for AAT proteolysis , in particular MMP-7 , MMP-9 and MMP-26 [38] , [39] . Altered levels of MMP-9 have been reported to correlate with HIV-1 infection , and breakdown of extracellular matrix has been suggested to aid dissemination of the virus [40] , [41] . Our in vitro experiments confirmed the known MMP-7 cleavage sites on the C-terminal part of AAT and also demonstrated that VIRIP can be generated despite presence of an additional cleavage site within the sequence of the peptide . This cleavage site between Pro381-Met382 , directly neighbouring the active site Ser383 , has been shown previously to be sensitive to the oxidation state of the Met382 residue [42] , thereby protecting VIRIP from further degradation . In addition , our results support the notion that the biological function of cleavage of AAT by MMP-7 may not solely be inactivation of its serine protease function , but also to generate new proteolytic peptides that have additional activities in themselves . The abundant acute phase proteins that we were able to detect elevations in during AHI may be only selected examples of constitutively-produced or inducible analytes that play a role in combating infections . There is an increasing amount of evidence suggesting that “endogenous factors” exist that have inherent inhibitory activity towards infectious pathogens [43] . These include antimicrobial polypeptides [44] and antiproteases such as cystatins that have also been detected in cervical mucosa [45] . Insights gained into the mechanism of action of innate factors and acute phase reactants against HIV-1 and how they can be induced should be considered for novel vaccine strategies and therapeutics .
Plasmapheresis samples from the US plasma donor cohort used in this study were purchased from Zeptometrix Corporation and SeraCare Life Sciences . Donors were recruited via the SeptaCare Special Donor Program and enrolled for plasma donation after a medical examination and an interview with medical staff , in which it was explained that the donated plasma will be used by dedicated drug and vaccine researchers to perform research to help others . This study was approved by the Oxford Tropical Research Ethics committee ( OXTREC , University of Oxford ) and the NIH Office of Extramural Research ( Nr . 201029 to B . M . K ) . Panels of sequential samples obtained by plasmapheresis from US plasma donors who became infected with HIV-1 , HBV or HCV and control subjects were purchased from Zeptometrix Corporation and SeraCare Life Sciences and were stored at −80°C before use . Details of the panels and methods used for analysis of HIV-1 , HBV and HCV viral loads are as described [14] . Each plasma panel included samples collected prior to detection of plasma viremia through to seroconversion . The plasma panels from HIV-infected individuals were temporally aligned relative to a common time origin ( T0 ) , defined as the time point when the viral load first reached detectable levels ( >100 viral RNA copies ml−1 ) , as described previously [14] . The median and range duration of observation prior to the first detectable HIV RNA level ( T0 ) for the 19 panels used in this study was 21 days with a range of 7 to 58 days . Plasma donor samples were fractionated using weak anion exchange ( WAX ) magnetic beads ( Bruker Daltonics , Bremen , Germany ) according to the manufacturer's recommendation ( Text S1 ) . For analysis by MALDI-TOF/TOF , a solution of α-cyano-4-hydroxycinnamic acid ( matrix , 3 mg ) in ethanol∶acetone ( 10 ml , 2∶1 ) was prepared freshly . Samples processed as described above and matrix solution were mixed in a ratio of 1∶4 , and 1 µL aliquots were spotted in triplicates on an Anchor Chip MALDI plate ( Bruker Daltonics , Bremen , Germany ) . Data was acquired on an UltraFlex MALDI-TOF/TOF instrument ( Bruker Daltonics , Bremen , Germany ) at 60% laser power until a total intensity of 2×105 ion counts was reached . Spectra were analyzed using Bruker Daltonics FlexAnalysis software ( version 2 . 4 ) . For LC-MS/MS tandem mass spectrometry , analysis was carried out by injection of 3 µl of sample prepared as described above on a nanoAcquity UPLC system coupled to a Waters Q-TOF Premier tandem mass spectrometer in MSE ( high/low collision switching ) mode as described previously [46] . Processing of raw data was performed using the ProteinLynx Global Server 2 . 2 . 5 software and the data were interrogated on an in-house MASCOT server ( version 2 . 2 ) . Alternatively , samples were analyzed on a Bruker HCTplus Ion Trap tandem mass spectrometer ( Bruker Daltonics , Bremen , Germany ) as described [47] . Semi-quantitative analysis of MS data was based on measuring peak heights observed for the peptide masses that were assigned to A-SAA86–105 [M+H]+ 2178Da , C3960–979 [M+H]+ 2213Da ( analysis by MALDI-TOF ) and VIRIP377–396 [M+3H]3+ ( methionine oxidized form ) 773 . 75Da ( analysis by LC-MS/MS ) . Seven plasma donor panels of infected individuals ( 61 time points total ) and five uninfected control plasma donor panels ( 40 time points total ) were examined . For an approximate estimation of VIRIP quantities in plasma , synthetic VIRIP was synthesized using Fmoc chemistry based solid-phase peptide synthesis on an automated synthesizer ( Advanced ChemTech , Louisville , KY , USA ) and different concentrations measured by LC-MS/MS by monitoring the [M+3H]3+ precursor ion intensities under the same conditions used for the analysis of plasma samples . It should be noted that sample loss during plasma sample processing may lead to an underestimation of the effective VIRIP concentration in plasma . Recombinant active MMP-7 ( EMD Biosciences , Gibbstown , NJ ) ( 7 . 2 µg , 4 µl ) was added to a solution of AAT ( Sigma Aldrich , St . Louis , MO ) ( 100 µl , 2 . 9 µg µl−1 ) in 100 mM ammonium bicarbonate buffer and the mixture was incubated at 37°C . Aliquots ( 5 µl ) were removed for immunoblotting at 0 , 5 , 20 , 60 and 120 min , mixed with RSB ( 5 µl ) and retained for analysis . Immunoblotting was carried out following gel electrophoresis on 4–12% Bis-Tris or 16% Tris–Tricine gels ( Invitrogen , Carlsbad , CA ) for increased resolution of low molecular weight species . LC-MS/MS analysis was carried out with the 5 min and 120 min time point samples following desalting and concentration by methanol chloroform precipitation [48] . Analysis of plasma samples using an anti-SAA ELISA assay was performed using a commercial SAA ELISA kit ( Abazyme , Needham , MA ) and conducted according to the manufacturer's instructions . Plasma samples were diluted with assay buffer to obtain final concentrations within the linear range of the assay of 1–80 ng ml−1 . Inclusion of recombinant protein standards demonstrated that A-SAA was sensitively detected by the assay while C-SAA was non-reactive up to final concentrations of 750 ng ml−1 . HIV-1 infection assays used viruses derived from the infectious molecular clones pNL4 . 3-BaL . ecto ( R5-tropic ) and pNL4 . 3 ( X4-tropic ) , supplied by John Kappes and Christina Jambor ( University of Alabama at Birmingham , USA ) . Cryopreserved healthy donor PBMCs were stimulated with 2 . 5ug ml−1 PHA ( Sigma Aldrich , St . Louis , MO ) for 2 d in R10 medium ( Invitrogen , Carlsbad , CA ) and 1 d with 50U ml−1 IL-2 ( Roche , Nutley , NJ ) in R10 medium adjusted to contain 20% serum . MDDCs were generated by culturing CD14+ cells isolated from buffy coats ( National Blood Service , London , UK ) using a Human Monocyte Isolation Kit II ( Miltenyi Biotec , Cologne , Germany ) for 6 d in 10ng ml−1 IL-4 and 100ng ml−1 granulocyte-macrophage colony stimulating factor ( GM-CSF ) ( eBioscience , San Diego , CA ) . The proportion of CD14loCD11chi MDDC was typically 95% . Inhibitors included CRP ( ProSpec-Tany TechnoGene LTD . , Rehovot , Israel ) , AAT ( Sigma , Gillingham , UK ) , A-SAA ( MBL International , Woburn , MA ) , VIRIP and the 22 amino acid AAT fragment C-terminal to this peptide ( both synthesized by Fmoc chemistry as described above ) . CRP , AAT and A-SAA were dissolved in serum-free OptiMEM ( Invitrogen , Carlsbad , CA ) , VIRIP and the 22 amino acid AAT fragment were dissolved in DMSO ( Sigma-Aldrich , St . Louis , MO ) and diluted to a 2% solution in OptiMEM . Inhibitors were diluted into cell culture medium to give the indicated final concentrations . The final concentration of DMSO in the culture medium was never more than 0 . 2% . Control cultures were treated with a similar volume of 2% DMSO OptiMEM or plain OptiMEM ( vehicle only ) to match the addition of inhibitor . PBMCs were preincubated with inhibitors for 2 h prior to infection , then were infected with R5 or X4 virus at a MOI of 0 . 01 for 16 h at 37°C in the presence of inhibitor . Excess virus and inhibitor were washed out and cells were cultured for 7 d in R10 medium plus 50U ml−1 IL-2 . Supernatants were harvested and p24 levels were determined by ELISA ( Advanced BioSciences Laboratories , Inc . , Kensington , MD ) . MDDCs were preincubated with inhibitors for 1 h prior to infection at a MOI of 0 . 1 for 2 h at 37°C in the presence of inhibitor; and after washout of inhibitor and excess virus , cells were cultured for 4 before reading out supernatant p24 levels as described above . Statistical analysis of the data from the viral inhibition assays was performed using a one-way ANOVA . Two statistical tests were performed on transformed data from the sample time-courses from the five control non-HIV-infected plasma donors and the time points prior to D-15 in the time-courses from HIV-infected plasma donors: a Wilcoxon test for a two-group comparison ( control vs . HIV-infected , p = 0 . 28 ) and an ANOVA test in a linear mixed model for the group difference ( p = 0 . 4 ) . As no significant differences were found between the controls and pre-D-15 data from the HIV-infected subjects , both were used to determine baseline A-SAA levels in the HIV-infected subjects . Linear mixed-effects models were fit to these data to estimate the subject-specific baseline level of A-SAA in each HIV-infected individual . A-SAA values above the 90% upper prediction bound was considered significantly elevated ( see Fig . 3A and S2A ) . A two-sided Binomial test was conducted to examine whether the first elevation of A-SAA occurred significantly before T0 . Additional tests of associations between the timing of first A-SAA elevation and three different parameters of viral replication ( viral reproductive rate R0 , slope of viral ramp-up and the highest recorded viral load ) were also performed using linear models .
|
Acquired immune deficiency syndrome ( AIDS ) remains a major health problem worldwide , affecting predominantly the adult population in the western world and in developing countries in particular . Despite a tremendous effort to develop a cure or a vaccine that confers protection against human immunodeficiency virus ( HIV-1 ) infection , this has not been achieved in a satisfactory manner to date . Recent research efforts have suggested that the earliest immune responses activated after exposure to the virus have an influence on virus spread , containment and disease progression . In this study , a panel of donors who provided plasma samples collected over a time-frame spanning the period before and immediately after detection of HIV-1 infection permitted an insight into the activation of the earliest systemic immune responses . We describe increases in plasma levels of acute-phase reactants and proteolytically processed fragments that have anti-viral activity in vitro . These inductions occur prior to detection of HIV-1 virus in the blood and before the first increases in systemic cytokine levels , which may represent the earliest systemic host antiviral response activated following infection .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biochemistry",
"cell",
"biology/cell",
"signaling",
"cell",
"biology",
"biochemistry/chemical",
"biology",
"of",
"the",
"cell",
"chemical",
"biology/protein",
"chemistry",
"and",
"proteomics",
"chemical",
"biology/chemical",
"biology",
"of",
"the",
"cell"
] |
2010
|
Elevation of Intact and Proteolytic Fragments of Acute Phase Proteins Constitutes the Earliest Systemic Antiviral Response in HIV-1 Infection
|
Thymus is crucial for generation of a diverse repertoire of T cells essential for adaptive immunity . Although thymic epithelial cells ( TECs ) are crucial for thymopoiesis and T cell generation , how TEC development and function are controlled is poorly understood . We report here that mTOR complex 1 ( mTORC1 ) in TECs plays critical roles in thymopoiesis and thymus function . Acute deletion of mTORC1 in adult mice caused severe thymic involution . TEC-specific deficiency of mTORC1 ( mTORC1KO ) impaired TEC maturation and function such as decreased expression of thymotropic chemokines , decreased medullary TEC to cortical TEC ratios , and altered thymic architecture , leading to severe thymic atrophy , reduced recruitment of early thymic progenitors , and impaired development of virtually all T-cell lineages . Strikingly , temporal control of IL-17-producing γδT ( γδT17 ) cell differentiation and TCRVγ/δ recombination in fetal thymus is lost in mTORC1KO thymus , leading to elevated γδT17 differentiation and rearranging of fetal specific TCRVγ/δ in adulthood . Thus , mTORC1 is central for TEC development/function and establishment of thymic environment for proper T cell development , and modulating mTORC1 activity can be a strategy for preventing thymic involution/atrophy .
The thymus is the primary organ for T cell development and generation of a diverse repertoire of T cells that are crucial for host defense but are also self-tolerated . Thymic epithelial cells ( TECs ) are essential for thymopoiesis and establish an environment that properly nurtures T cell development [1] . TECs include cortical and medullary subsets that reside in different localizations in the thymus and perform distinct functions . While cortical thymic epithelial cells ( cTECs ) are important for positive selection of conventional TCRα/β T ( cαβT ) cells , medullary thymic epithelial cells ( mTECs ) induce negative selection of highly self-reactive T cells and generation of regulatory T cells ( Tregs ) [2–5] . Interestingly , TECs are dynamically regulated by the cTEC to mTEC ratios being highest in the fetus and progressively lower as the mouse matures . In adult mice , mTECs substantially outnumber cTECs [6] . Although several transcription factors such as Foxn1 and Aire and receptors such as RANK , CD40 , and LTβR are found important for TEC development/function [7–9] , mechanisms that control thymopoiesis and mTEC/cTEC ratios are poorly understood . During T cell ontogeny , early T cell progenitors ( ETPs ) enter into the thymus at the corticomedullary junction . Following initial migration toward the cortex , ETPs , which are Lin−CD4−CD8− double negative ( DN ) cells that express CD44 , cKit , and CD24 but not CD25 , undergo maturation sequentially through the DN2 , DN3 , and DN4 stages [10] . TCRγδ T ( γδT ) -cells arise from these DN stages and can further differentiate into effector lineages such as IFNγ-producing γδT1 and IL-17-producing γδT17-cells within the thymus [11] . Interestingly , γδT17 differentiation occurs predominantly in the fetal thymus [12 , 13] . Additionally , several TCRVγ ( Vγ5 and Vγ6 ) and TCRVδ1 segments recombine only in the fetal thymus [14 , 15] . Whether and how TECs may nurture a thymic environment to confer such temporal control of γδT17 differentiation and fetal-specific TCRVγ/Vδ usage has been unclear . DN thymocytes uncommitted to the γδT fate but with in-frame rearranged TCRβ may overcome the developmental checkpoint between DN3 and DN4 to reach the CD4+CD8+double positive ( DP ) stage and adopt the αβT fate . Expression of a functional TCRα/β that recognizes self-peptide-major histocompatibility complex ( MHC ) complexes presented by cTECs triggers positive selection for maturation to the CD4+CD8− or CD4−CD8+ single positive ( SP ) stage [16] . After positive selection , SP thymocytes migrate to the thymic medulla in a CCR7-dependent manner [17] . In the medulla , mTECs present promiscuously expressed tissue-specific antigens ( TRAs ) to SP thymocytes to trigger negative selection of cells that express TCR with high affinity to TRAs [18] . Small subsets of αβTCR-expressing thymocytes adopt Treg and NKT cell fates . Both negative selection and Tregs are critical for self-tolerance to prevent autoimmune diseases . Abnormal TEC development and function can cause severe consequences such as immunodeficiency or autoimmune diseases in both humans and animals , exemplified by deficiencies in Foxn1 or Aire [7 , 19 , 20] . The serine/threonine kinase mammalian/mechanistic target of rapamycin ( mTOR ) has the ability to integrate various environmental and intracellular stimuli and cues to control cell growth , proliferation , survival , autophagy , and metabolism . Mammalian/mechanistic target of rapamycin complex 1 ( mTORC1 ) , one of the two complexes , which contains a crucial and unique adaptor molecule Raptor , phosphorylates multiple substrates such as S6K1 and 4E-BP1 to promote protein , nucleic acid , and lipid synthesis , which is crucial for cell growth and proliferation [21] . mTOR is activated in thymocytes following TCR engagement via both PI3K-Akt and RasGRP1-Ras-Erk1/2 pathways [22] and intrinsically controls the development and/or function of iNKT cells , Tregs , and cαβT cells [23–28] . However , whether mTOR plays a role in TECs to extrinsically control T cell development is unknown . In this report , we demonstrate that mTORC1/Raptor signaling in TECs is crucial for thymopoiesis and proper generation of multiple T cell lineages . Deficiency of mTORC1/Raptor in TECs causes severe thymic atrophy , altered thymic structure , decreased mTEC/cTEC ratios , and severely reduced production of cαβT cells , Tregs , iNKT cells , and γδT cells correlated with decreased recruitment of ETPs in the thymus . Moreover , fetal thymus restricted γδT17 differentiation and TCRVγ5/6Vδ1 recombination occur in adult thymus in the absence of mTORC1 in TECs , suggesting that TECs and thymic environment rather than hematopoietic stem cells confer temporal control of γδT cell development .
We first examined S6 phosphorylation , an mTORC1/S6K1 dependent event , in TECs from mice aged at 9 d , 3 wk , and 10 wk . S6 phosphorylation was the strongest in TECs from 9-d-old mice but the lowest in cells from 10-wk-old adult mice ( Fig 1A ) . Further comparison between 6-wk- and 6 . 5-mo-old mice revealed decreased S6 phosphorylation in TECs in aged mice ( Fig 1A ) . Thus , mTORC1 activity in TECs appeared high in young mice but decreased with older age . To determine the importance of mTORC1/Raptor signaling in thymus homeostasis , we examined 6–8-wk-old Rptorf/f-Rosa26-ERCre ( Rptf/f-ERCre or eKO ) and control Rptorf/f ( Rptf/f or wild-type; WT ) mice following tamoxifen injections on days 1 , 2 , and 5 . On day 8 , thymi in Rptf/f-ERCre mice were much smaller than Rptf/f mice ( Fig 1B ) , accompanying a substantial decrease in total thymocyte numbers ( Fig 1C ) . The percentage of CD4+CD8+DP thymocytes was decreased , but the percentages of CD4−CD8−DN , CD4SP , and CD8SP thymocytes were increased in Rptf/f-ERCre mice ( Fig 1D and 1E ) . The absolute number of DP thymocytes was severely decreased; CD4SP cell number was decreased by 50%; but DN and CD8SP thymocyte numbers were not obviously affected in tamoxifen-treated Rptf/f-ERCre mice ( Fig 1E ) . This was correlated with increased death in DP thymocytes ( Fig 1F ) . Hematoxylin and eosin ( H&E ) staining of thymus thin sections from tamoxifen-treated Rptf/f-ERCre mice revealed abnormal thymus architecture: shrinkage of the cortex and increased presence of vacuous/cyst-like structures in medulla ( Fig 1G ) , which were confirmed by immunofluorescence staining of cortex and medulla with anti-Keratin 8 ( KRT8 ) and anti-Keratin 5 ( KRT5 ) antibodies , respectively ( Fig 1H ) . The shrinkage of cortex in Rptf/f-ERCre mice was correlated with a sharp decrease in the number of DP thymocytes , which normally account for most cells in the cortex . In Rptf/f-ERCre mice , although CD45−EpCAM+ TEC percentages were increased 6-fold ( Fig 1I ) , their total numbers were similar to that of WT controls ( Fig 1J ) . The predominance of UEA-1+Ly51− mTECs over UEA-1−Ly51+ cTECs was not substantially disturbed in both percentages and numbers ( Fig 1K and 1L ) . Thus , acute systemic mTORC1 deletion rapidly caused severe thymic atrophy , altered thymic architecture , and selectively reduced number of DP thymocytes . Although systemic deletion of mTORC1 caused severe thymic atrophy and decrease of DP thymocytes , T cell-specific ablation of mTORC1 in Rptf/f-LckCre or Rptf/f-CD4Cre mice did not cause obvious thymic atrophy or abnormal distribution of DN , DP , and SP populations in the thymus [28] . To test our hypothesis that mTORC1 might play a critical role in TECs for thymopoiesis and function , we generated and analyzed Rptf/f-Foxn1Cre ( knockout; KO ) and Rptf/f ( WT ) mice . Foxn1Cre mice contain an IRES-Cre cassette inserted into the 3’ untranslated region in the Foxn1 locus to direct Cre expression starting on embryonic day 11 . 5 in TECs [29] . Strikingly , fetus ( embryonic day 20 , E20 ) , newborn ( 1d ) , young ( 10 days , 10d; 18 days , 18d; 3 weeks , 3w ) , and adult ( 6–8 weeks , 6w ) Rptf/f-Foxn1Cre mice displayed apparent thymic atrophy ( Fig 2A ) , accompanied by severely decreased total thymocyte numbers compared with Rptf/f controls ( Fig 2B ) . Additionally , normal medullary and cortical structure was lost in Rptf/f-Foxn1Cre thymus , which had a severely atrophied medulla ( Fig 2C and 2D ) . Thus , mTORC1 in TECs was crucial for normal thymopoiesis . Severe thymic atrophy in Rptf/f-Foxn1Cre mice suggested that mTORC1/Raptor signaling might be required for TEC development and/or function . Although TEC percentages from Rptf/f-Foxn1Cre mice were not notably altered at indicated ages ( Fig 3A and 3B ) , total TEC numbers were obviously decreased in Rptf/f-Foxn1Cre thymi ( Fig 3C ) . The magnitude of reduction was smaller at E20 but was progressively exacerbated as mice matured . Both Ly51−UEA-1+ mTEC and Ly51+UEA-1− cTEC proliferation reflected by BrdU incorporation was decreased ( Fig 3D ) , but their survival was not impaired in Rptf/f-Foxn1Cre mice ( Fig 3E ) . The impaired TEC proliferation was correlated with decreased S6 phosphorylation and , thus , reduced mTORC1 activity ( Fig 3F ) and decreased glucose uptake ( Fig 3G ) . Thymic atrophy and reduced TECs in Rptf/f-Foxn1Cre mice were not caused by expression of Cre protein itself in TECs , as thymi in Rpt+/+-Foxn1Cre mice and Rptf/f mice were similar in sizes , total thymic cellularity , thymocyte subsets , and TEC numbers ( S3A–S3H Fig ) . Together , these observations demonstrated that mTORC1 was critical for normal TEC development at least through promoting TEC expansion and glucose uptake . In WT mice , Ly51+UEA-1− cTECs were about 2-fold more than Ly51−UEA-1+ mTECs in E20 and newborn thymi but accounted for about or less than 10% of total TECs after 3 wk of age ( Fig 4A and 4B ) . In Rptf/f-Foxn1Cre mice , cTEC percentages were 2–9-fold higher than WT controls from embryos to adulthood with the biggest difference at 3 wk of age ( 8 . 76 ± 1 . 40% WT versus 79 . 36 ± 4 . 75% KO ) , resulting in substantial decreases of mTEC/cTEC ratios after 3 wk of age ( Fig 4C ) . Noticeably , due to severe decreases of total TECs , cTEC numbers were also decreased in Rptf/f-Foxn1Cre mice throughout their life span except at 3 wk of age ( Fig 4D ) . Expression of Aire , a transcription factor critical for mTEC maturation [20] , was not decreased but rather slightly increased in Rptf/f-Foxn1Cre mTECs ( Fig 4E ) , ruling out decreased Aire expression as a causal factor of impaired TEC maturation/maintenance . Both cTECs and mTECs can be defined into MHCIIlowCD40low immature and MHC-IIhiCD40hi mature stages [30 , 31] . From embryos to 3-wk-old mice , fewer Rptf/f-Foxn1Cre cTECs reached mature stage than WT controls ( Fig 4F and 4G ) . Such phenotype was not observed in adult mice . For mTECs , under-representation of mature stage was only observed in embryos but not after birth . The relatively unimpaired mTEC maturation was correlated with elevated Aire expression . However , due to a severe decrease of total TECs , mature mTECs and cTECs were considerably decreased throughout the life span of Rptf/f-Foxn1Cre mice ( Fig 4H ) . Together , these observations demonstrated that mTORC1 is important for TEC expansion and efficient maturation and for establishing mTEC predominance over cTECs after adolescence . We further examined the impact of mTORC1 deficiency in TECs on T cell development . Percentages of DN , DP , CD4SP , and CD8SP thymocytes in Rptf/f-Foxn1Cre were similar to WT controls at both 10-d and 6-wk of age ( Fig 5A and 5B ) , except that CD4SP thymocyte percentage was decreased by 50% in Rptf/f-Foxn1Cre thymus at 10 d . Within CD4SP and CD8SP cells , the ratios of TCRβ+CD24− mature population were not obviously reduced ( Fig 5C ) , suggesting that maturation of SP thymocytes was unhindered . Due to the drastic decrease of total thymic cellularity , the absolute numbers of all these populations were severely decreased in Rptf/f-Foxn1Cre mice ( Fig 5D ) . BrdU incorporation in DN , DP , and SP thymocytes was similar or only slightly decreased ( Fig 5E ) , but annexin V+ apoptotic cells in these populations were increased ( Fig 5F ) , suggesting that mTORC1 in TECs promotes thymocyte survival . IL-7 , a pro-survival cytokine , was not decreased but actually increased at the mRNA level in Rptf/f-Foxn1Cre TECs ( Fig 5G ) , implying that mTORC1 may not control thymocyte survival via upregulating IL-7 transcription . However , due to the scarcity of TECs , we could not measure IL-7 protein in TECs and thus could not rule out that mTORC1 may promote thymocyte survival via increasing IL-7 translation . Nevertheless , our data demonstrated that mTORC1/Raptor deficiency in TECs led to increased thymocyte death and impaired αβT cell production without causing a developmental blockade at specific developmental checkpoints . T cells generated in the thymus populate peripheral lymphoid organs to perform their functions . Both CD4+ and CD8+ cαβT cell percentages and numbers in the spleen were considerably decreased in Rptf/f-Foxn1Cre mice ( Fig 5H–5J ) , without obviously skewing TCRVβ usage ( S6 Fig ) . There were noticeable increases of CD44+CD62L− or CD44+CD62L+ effector/memory-like CD4 and CD8 T cells but decreases in CD62L+CD44− naïve T cells in Rptf/f-Foxn1Cre mice ( Fig 5K ) , which was likely caused by lymphopenic proliferation . Thus , decreased T cell output from Rptf/f-Foxn1Cre thymus resulted in T cell lymphopenia . Recent studies have implicated mTECs for nTreg development [3 , 4] . The severe reduction of mTECs in Rptf/f-Foxn1Cre mice prompted us to examine whether nTreg development was jeopardized . As with CD4+Foxp3− Teff , CD4+Foxp3+ , nTreg percentages and numbers were substantially decreased in Rptf/f-Foxn1Cre thymi at both 10 d and 6 wk of age ( Fig 6A–6C ) . Moreover , nTreg percentages within CD4+TCRβ+ cells and the Treg/Teff ratios were more than 50% lower than those in WT controls ( Fig 6D–6F ) , suggesting a more severely compromised nTreg development than Teff in Rptf/f-Foxn1Cre mice . One potential mechanism for the severe nTreg developmental defect in Rptf/f-Foxn1Cre mice could be the increased death of these cells . However , Rptf/f-Foxn1Cre nTregs were not obviously prone to death when compared with control mice ( Fig 6G ) . Expression of CD27 , which promotes nTreg survival and generation [5] , was not decreased but slightly increased in nTregs , Foxp3-CD4+CD8- SP , and CD4-CD8+ SP thymocytes from Rptf/f-Foxn1Cre mice ( Fig 6H ) . Additional studies are needed to determine whether an abnormal CD27 costimulatory signal or other mechanisms contribute to severely compromised nTreg generation in Rptf/f-Foxn1Cre mice . In Rptf/f-Foxn1Cre mice , splenic Treg percentages and numbers were also reduced within total splenocytes ( Fig 6I–6K ) . However , Treg/Teff ratios were about 1 . 5-fold higher than WT controls ( Fig 6L ) . These Foxp3+ Treg expressed nTreg marker Helios ( Fig 6M ) , indicating that the relative enrichment of Tregs in the periphery of Rptf/f-Foxn1Cre mice was likely caused by lymphopenia-induce proliferation because Treg expansion is superior to Teff under such condition . Tregs from Rptf/f-Foxn1Cre mice expressed higher levels of several Treg signature molecules such as LAG3 , CTLA-4 , and GITR that facilitate their suppressive activity . Together , these observations indicated that mTORC1 in TECs plays important roles in nTreg differentiation . The invariant Vα14-Jα18 TCR-expressing NKT cells ( iNKT ) cells are also generated in the thymus . Unlike cαβT cells , they are positively selected after engagement of the iVα14TCR with self-lipid ligand-CD1d complex expressed on DP thymocytes [32] . The role of TECs in exogenously controlling iNKT cell development is largely unclear . The percentages of PBS57-loaded CD1d-Tetramer ( CD1Dtet ) + TCRβ+ iNKT cells were decreased in the thymus of 10-d- and 6-wk-old Rptf/f-Foxn1Cre mice ( Fig 7A and 7B ) , with more drastic decreases of iNKT cell total numbers ( Fig 7C ) . Moreover , although the ratio of iNKT to cαβT was comparable between Rptf/f-Foxn1Cre and Rptf/f control mice at 10 d of age , it decreased by 67% in adult Rptf/f-Foxn1Cre thymus ( Fig 7D ) , suggesting more defective iNKT generation than cαβT cells in adult mice . Although IL-15 expressed by mTECs promotes late stage iNKT cell development [33] and the mTEC number was severely decreased in Rptf/f-Foxn1Cre thymus , no obvious late stage iNKT developmental blockade was observed . The relative percentages of stages 1 ( CD24−CD44−NK1 . 1− ) , 2 ( CD24−CD44+NK1 . 1− ) , and 3 ( CD24−CD44+NK1 . 1+ ) iNKT cells in these mice were similar to WT controls ( Fig 7E and 7F ) , although their absolute numbers were markedly decreased ( Fig 7G ) . Just like in the thymus , iNKT cell percentages and numbers were obviously decreased in the spleen ( Fig 7A–7C ) . Together , these observations indicate that iNKT cell generation is dependent on mTORC1 signaling in TECs . Unlike cαβT cells , most γδT cells develop independent of the MHC-mediated antigen presentation by TECs [34 , 35] . Although γδT cell percentages in Rptf/f-Foxn1Cre thymi were not obviously altered at 10 d and 6 wk of age ( Fig 8A and 8B ) , total γδT cell numbers were substantially decreased ( Fig 8C ) . However , unlike nTreg and iNKT cells , γδT to cαβT ratios were not decrease or even slightly increased in Rptf/f-Foxn1Cre mice ( Fig 8D ) , suggesting that γδT cell generation was dependent on mTORC1 in TECs but appeared less severely affected than nTregs and iNKT cells . Considerable numbers of γδT cells are programmed to differentiate to distinct effector lineages within the thymus [34–36] . γδT17 cells are developed mostly in fetal thymus [12 , 13] . Similarly , TCRVγ5 , Vγ6 , and Vδ1 only recombine in fetal thymus [11 , 14] . Mechanisms that enforce such temporal controls are unknown . In WT thymus , γδT17 cells accounted for about 40% and 5% of total γδT cells at birth and 6 wk of age , respectively . In Rptf/f-Foxn1Cre mice , γδT17 percentages did not differ greatly at birth but increased 5-fold in adults compared to WT controls ( Fig 8E and 8F ) . Although total thymic γδT17 numbers were decreased in newborn Rptf/f-Foxn1Cre thymi , they were similar to WT controls when 6 wk old ( Fig 8G ) . Because adult Rptf/f-Foxn1Cre thymi were much smaller than WT controls , similar total γδT17 cell numbers in these mice suggested that γδT17 generation was greatly favored in adult Rptf/f-Foxn1Cre thymi , although thymic γδT1 ratio was not altered ( Fig 8E and 8H ) and γδT1 numbers were obviously decreased in Rptf/f-Foxn1Cre mice ( Fig 8I ) . Additionally , CD44+CD27− γδT cells , which were enriched with γδT17 [13] , were also increased in Rptf/f-Foxn1Cre thymi ( Fig 8J ) . Concordantly , γδT17 but not γδT1 percentages in splenic , lung , and liver γδT cells were also increased in Rptf/f-Foxn1Cre mice ( Fig 8K and 8L ) . To determine whether Cre protein expression in TECs was sufficient to cause enhanced γδT17 generation , we compared γδT1/17 differentiation in Rpt+/+-Foxn1Cre mice and Rpt+/+ mice . As shown in S10A–S10D Fig , Rpt+/+-Foxn1Cre thymic γδT cells were not obviously different from Rpt+/+ thymic γδT cells in percentages and numbers as well as in γδT1 and γδT17 percentages , suggesting that Cre expression in TECs per se was not able to cause the abnormalities in Rptf/f-Foxn1Cre mice . Together , these observations revealed that mTORC1/Raptor in TECs prevented γδT17 generation in the adult thymus . Coinciding with the loss of temporal control of γδT17 differentiation , Vγ5 , Vγ6 , and Vδ1 recombination , which occurs only in fetal thymi in WT mice , occurred at high levels in adult Rptf/f-Foxn1Cre thymi ( Fig 9A and 9B ) . Although γδT subsets defined by TCRVγ usages were similar in newborn WT and Rptf/f-Foxn1Cre thymi ( Fig 9C , S12 Fig ) , they were obviously different in adult thymi ( Fig 9D and 9E ) . Both Vγ6Vδ1+ and Vγ5+ subsets were increased , but Vγ4+ subsets were decreased in relative ratios in adult Rptf/f-Foxn1Cre thymi compared with WT controls; thus , mTORC1/Raptor in TECs enforced strict restriction of fetal-specific Vγ5/6/Vδ1 recombination . In newborn thymi , we observed similarly high percentages of γδT17 cells that were either Vγ4+ or Vγ6+ , which accounted for most of γδT17 cells in a normal fetal thymus [12] , in both WT and Rptf/f-Foxn1Cre mice ( Fig 9F and 9G ) . There were no obvious differences in γδT17 percentages in Vγ1+ , Vγ4+ , Vγ5+ , and Vγ6Vδ1+ populations of γδT cells between newborn WT and Rptf/f-Foxn1Cre thymi . However , γδT cells in adult Rptf/f-Foxn1Cre thymi showed substantially increased Vγ6Vδ1+ and Vγ5+ γδT17 cells ( Fig 9H and 9I ) . Most newborn γδT17 cells were either Vγ4+ or Vγ6Vδ1+ in both WT and Rptf/f-Foxn1Cre mice . However , Vγ6Vδ1+ γδT cells accounted for the majority of γδT17 cells in Rptf/f-Foxn1Cre thyme , and this population of γδT17 cells was 4-fold greater in Rptf/f-Foxn1Cre thymi than in WT control ( Fig 9J ) . A significant portion of γδT17 cells in adult Rptf/f-Foxn1Cre thymi was Vγ4−Vγ6Vδ1− , which could be other Vγ subsets or poorly stained Vγ6Vδ1+ cells . However , they were unlikely αβT cells because they were TCRβ− . Moreover , γδT cells sorted from adult WT and Rptf/f-Foxn1Cre thymi expressed similarly low levels of TCRα compared with αβT cells ( Fig 9K ) . One possible reason for increased γδT17 cells in Rptf/f-Foxn1Cre adult thymi could be selective expansion of these cells in the thymus after they were generated in the fetus and in neonate . However , thymic γδT cells or CD44−CD27+ , CD44+CD27+ , and CD44+CD27− γδT subsets from Rptf/f-Foxn1Cre mice incorporated BrdU at similar rates as their respective WT controls ( Fig 10A–10C ) . Since γδT17 cells reside mainly in the CD44+CD27− subset , these observations suggested that the relative increase of γδT17 cells was not likely caused by selective expansion of these cells in Rptf/f-Foxn1Cre mice . Another potential but not mutually exclusive possibility was that mTORC1 deficiency in TECs caused selective retention of Vγ6+Vδ1+ γδT cell/γδT17 cells generated in the fetal thymus , leading to an increase of these cells in the adult Rptf/f-Foxn1Cre thymus . To address this possibility , we generated chimerical mice by reconstituting lethally irradiated Rptf/f and Rptf/f-Foxn1Cre mice with bone marrow from CD45 . 1+CD45 . 2+ WT mice . Five to six weeks after transfer , Rptf/f-Foxn1Cre recipient mice displayed a small thymus ( Fig 10D ) , decreased total thymic total cellularity ( Fig 10E ) and comparable percentages , but decreased numbers of donor-derived CD45 . 1+ thymocyte subsets based on CD4 and CD8 staining ( Fig 10F–10H ) . Donor-derived CD45 . 1+ γδT cell percentages were similar , but the numbers were decreased in Rptf/f-Foxn1Cre recipients compared with Rptf/f recipients ( Fig 10I–10K ) . Importantly , γδT17 cell percentages were increased 4–8-fold in donor-derived γδT cells in the Rptf/f-Foxn1Cre recipient thymus and spleen when compared with Rptf/f recipients ( Fig 10L ) . Thus , generation of fetal restricted γδT17 cells from WT hematopoietic stem cells from adult bone marrow was enhanced in adult thymi when mTORC1 was absent in TECs . Furthermore , percentages of Vγ5+ and Vγ6Vδ1+ cell in donor-derived γδT cells from Rptf/f-Foxn1Cre recipient thymi appeared higher than those in donor-derived γδT cells from Rptf/f recipient mice , while Vγ1 . 1+ and Vγ4+ γδT cell percentages were similar between these two groups ( Fig 10M ) , suggesting that generation of fetal restricted Vγ5/Vγ6 γδT cells from WT hematopoietic stem cells from adult bone marrow might be increased in adult TEC specific mTORC1 deficient thymi . T cell development initiates after ETPs take residence in the thymus [10] . Severe impairment of multilineage T cell generation without obvious blockade at specific developmental stages prompted us to examine if ETPs in the thymus were altered in Rptf/f-Foxn1Cre mice . The relative ratios of ETP ( Lin−cKit++CD25−CD24+CD44+ ) , DN2 ( cKit+CD44+CD25+ ) , DN3 ( CD44−CD25+ ) , and DN4 ( CD44−CD25− ) subsets within Rptf/f-Foxn1Cre Lin− thymocytes did not deviate greatly from E16 fetus , 3-wk- and 6-wk-old mice in the control group ( Fig 11A and 11B ) . This suggests there was no obvious developmental blockade from ETP to DN4 . However , ETPs and other DN subsets in fetal and postnatal Rptf/f-Foxn1Cre thymi were decreased in numbers ( Fig 11C–11E ) . Rptf/f-Foxn1Cre ETPs showed no obvious defect in survival ( Fig 11F ) . However , ETPs but not other DN subsets , displayed impaired BrdU incorporation ( Fig 11G and 11H ) , suggesting decreased in vivo expansion of ETPs when mTORC1 was absent in TECs . Migration of ETPs to the thymus requires signals from CXCR4 , CCR7 , and CCR9 , and their cognate ligands [37–40] . In E16 thymi ( Fig 11I ) , as well as in sorted postnatal d10 TECs ( Fig 11J ) , expression of these chemokines was considerably decreased in the absence of mTORC1 . Together , these observations suggested that mTORC1 might , at least in part , enhance expression of multiple thymotropic chemokines in TECs for recruitment of ETPs to the thymus and promote ETP expansion for efficient T cell generation .
We demonstrated here that mTORC1 in TECs is pivotal for normal thymopoiesis and for establishing a thymic environment to foster proper T cell generation . Deficiency of mTORC1 in TECs resulted in severe thymic atrophy , decreased TEC numbers , abnormal thymic architecture , and decreased mTEC/cTEC ratios , leading to reduced ETPs in the thymus , impaired generation of virtually all T cell lineages , and defective temporal control of γδT17 differentiation and fetal restricted TCRVγ/Vδ recombination . Using Foxn1Cre-mediated deletion , we have revealed that mTORC1 may control multiple aspects of TEC biology . First , mTORC1 is important for TEC expansion and its deficiency leads to decreased TEC numbers in both fetal and adult Rptf/f-Foxn1Cre mice . Such function of mTORC1 in TECs is consistent with its role in cell cycle entry and synthesis of building blocks critical for cell growth and expansion [21] . Second , mTORC1 promotes late-stage TEC maturation indicated by decreased relative ratios of MHC-IIhiCD40hi TECs in Rptf/f-Foxn1Cre mice . Although not examined in the current study , decreased mature TEC numbers likely affect thymic selection and T cell repertoire . Third , mTORC1 controls the balance between mTECs and cTECs and ensures establishing predominance of mTECs over cTECs . Finally , mTORC1 may augment the recruitment of ETPs to the thymus by increasing CXCL12 , CCL21 , and CCL25 expression in TECs and promote ETP expansion in the thymus through mechanism ( s ) yet to be defined . Using Rptf/f-ERCre mice , we have also shown that acute deletion of mTORC1 in adult mice quickly causes severe thymic atrophy , altered thymic architecture , and substantial reduction of DP thymocytes within one week . Our data are consistent with previous observations that rapamycin treatment induces thymic atrophy and DP thymocyte death in mice [41] . Since TEC numbers are not obviously reduced and thymocyte-specific deletion of Raptor does not affect thymus size and total cellularity [28] , we propose that mature TEC function relies on mTORC1 activity to ensure DP thymocyte survival . Due to shortened life span of Rptf/f-ERCre mice after tamoxifen injection , our results do not rule out a potential role of mTORC1 for mature c/mTEC homeostasis after prolonged deletion . It is important to point out that a recent study has found severe thymic atrophy in the same strain of mice 2–3 wk after tamoxifen injection and has attributed thymic atrophy to early T cell developmental blockade [28] . TECs and thymic architecture were not evaluated in that study . Our conclusion is not in conflict with that study , as hematopoiesis can be greatly impacted even after one week of tamoxifen treatment in these mice [42] . Nevertheless , to firmly establish the role of mTORC1 in mature TECs and how it promotes DP thymocytes survival , selective deletion of Raptor in mature TECs is required . An important question that remains to be addressed is how mTORC1 controls TEC development and function . The transcription factor Foxn1 is essential for TEC development and thymopoiesis [7 , 19] as well as for thymus maintenance [43 , 44] . Aire is required for mTEC development and function [45] . Expression of these two molecules in TECs is not decreased in mTORC1 deficient mice ( Fig 4E and S15 Fig ) . However , mTORC1 also controls nuclear translocation of multiple molecules [24] , so our data do not rule out that mTORC1 may regulate the localization and function of these molecules . mTORC1 regulates the expression/activity of many other molecules [21] . The impairment of thymopoiesis and T cell development in mTORC1-deficient mice likely compounds the effects of multiple abnormalities . Although the generation of virtually all T cell lineages is impaired in Rptf/f-Foxn1Cre mice , individual T cell lineages appear to display differential sensitivities to mTORC1 deficiency in TECs . iNKT cells appear most stringently dependent on mTORC1 signaling in TECs , which is surprising , because their positive selection relies on engagement of the iVα14TCR with self-lipid ligands presented by CD1d expressed on DP thymocytes in the cortex rather than TECs [32] . CD1d expression on and Vα14-Jα18 recombination in DP thymocytes were not obviously affected in Rptf/f-Foxn1Cre mice ( S16 Fig ) . A recent study found that mTECs produce IL-15 to promote late stage iNKT cell development [33] . In Rptf/f-Foxn1Cre mice , there was no obvious iNKT cell developmental blockade at a late stage , suggesting that mTORC1 in TECs may function through other mechanism ( s ) to promote early iNKT cell development . Similar to iNKT cells , nTregs are more sensitive than cαβT cells to mTORC1 deficiency . nTreg differentiation depends on self-peptide-MHC-II presented by or derived from mTECs [3 , 4] . The disproportional decrease in the number of mTECs as well as reduction of mature mTECs in Rptf/f-Foxn1Cre thymus may contribute to a more severe impairment of nTreg generation . Furthermore , a thymic environment such as local TGFβ and CD80/86-mediated costimulation modulates nTreg generation [46 , 47] . Altered thymic environment in Rptf/f-Foxn1Cre mice could also contribute to nTreg deficiency . Immune cells undergo specific switches during development [48] . Among them are γδT17 cell differentiation and TCRVγ5 , Vγ6 , and Vδ1 recombination , which predominantly or strictly occur in fetal thymus and are switched off in adult thymus [12–15] . Mechanisms that enforce such temporal controls or developmental switch are unknown . We demonstrated that TEC-specific deletion of mTORC1 results in a loss of fetal restriction on γδT17 differentiation and recombination of Vγ5/6/Vδ1 , leading to uncontrolled γδT17-cell generation and Vγ5/6Vδ1 recombination in adulthood . Our data suggest that mTORC1 controls TECs to enforce such temporal Vγ5/6/Vδ1 recombination and γδT17 generation . Interestingly , a recent report has also found impaired temporal control of γδT17 differentiation and Vγ/Vδ recombination in adult β5t mutant mice [49] . Thus , although it has been previously suggested that fetal hematopoietic stem cells contain yet unknown properties that confer fetus specificity of Vγ5/6/Vδ1 usages in an in vitro culture system [50] , our data and those from Nitta et al . [49] suggest that thymic environment , particularly TECs , rather than fetal bone marrow hematopoietic stem cells ( HSCs ) ensures temporal control of γδT development . Interestingly , β5t mutation selectively impairs cTEC development to cause dysregulation of γδT17 cell generation in adult mice [49] . Because cTECs are relatively enriched in adult Rptf/f-Foxn1Cre mice , it is possible that mTORC1 may play a crucial role in cTECs to restrain Vγ6+ γδT17 cell generation in adult thymi . Important issues to be addressed in the future are how specific determinants in TECs dictate temporal control of γδT cell development and how mTORC1 signaling impact on these determinants . A potential possibility is that fetal and adult TECs are qualitatively different in a way that only fetal TECs confer an environment suitable for γδT17 differentiation and for opening Vγ5/6Vδ1chromatin for recombination . If this is true , mTORC1 may play an important role in the transition of TECs from fetal stage , which may be young and permissive for γδT17 differentiation and TCRVγ5/6/Vδ1 recombination , to adult TECs , which are aged and impermissive for γδT17 differentiation and TCRVγ5/6/Vδ1 recombination . γδT17 differentiation requires transcription factor RORγt and signals from Notch , TGFβ , and LTβR but not TCR [13 , 51 , 52] and is opposed by IL-15Rα signaling [53] . Whether mTORC1 acts on TECs to influence these signal mechanisms to confer temporal control of γδT17 differentiation is unknown at present . Nevertheless , our data provide genetic evidence that TECs nurture a thymic environment that limits postnatal γδT17 differentiation and Vγ5/6Vδ1 recombination in an mTORC1-dependent manner . Despite its importance , thymus undergoes involution or atrophy with advanced age or under certain pathological conditions . Thymic involution leads to a decrease in T cell production and shrinking of the T cell repertoire , which can result in impairment of adaptive immunity and propensity for autoimmunity [54 , 55] . Altered TECs can either cause or prevent thymic involution/atrophy [43 , 44 , 56] . Given the roles of mTORC1 in TECs for thymopoiesis and thymus homeostasis , and the declination of mTORC1 signaling in TECs with age ( Fig 1A ) , it is reasonable to speculate that gradual decreases of mTORC1 activity in TECs may contribute to thymic involution , and increases of mTORC1 activity might delay or prevent thymic involution . These hypotheses warrant further investigation . Additionally , rapamycin and its derivatives are utilized extensively in organ transplantation and cancer therapy . Their potential effects on thymic function should be taken into consideration .
Mouse experiments described were approved by the Institutional Animal Care and Use Committee of Duke University . Mice were euthanized with CO2 for experiments . Rptorf/f mice [57] were purchased from the Jackson laboratory and further backcrossed to C57Bl/6J background for at least four generations . Foxn1Cre mice [29] were kindly provided by Dr . Nancy Manley at the University of Georgia . Rptf/f-Rosa26-ERCre mice were previously reported [24 , 58] . For ERCre mediated deletion , mice were i . p . injected with 200 μl 10 mg/ml tamoxifen on day 1 , 2 , and 5 and euthanized on day 8 . Mice after overnight mating with virginal plug in the next morning were designated as gestation day 1 . All animals were housed in specific pathogen-free conditions . Experiments described were approved by the Institutional Animal Care and Use Committee of Duke University . TECs were prepared according to a published protocol with modifications [59] . In brief , thymi were cut into 2 mm pieces and directly digested in 2 ml digestion buffer ( 250 μl 10mg/ml collagenase type IV ( Worthington ) , 40 μl 50mg/ml DNase I ( Worthington ) and 1 . 71ml FBS-free RPMI-1640 ) at 37°C with constant orbital shaking at 150–200 rpm for 15 min . After gentle vortex , remnants were allowed to settle down; the supernatants were collected and kept on ice; settled remnants were digested similarly two more times . After the last digestion , cells were combined and filtered through 70 μm nylon mesh . After centrifuged at 472 g for 5 min , pellets were resuspended in 10 ml RPMI-containing 10% FBS ( RPMI-10 ) , spun again , and resuspended in either cold FACS buffer ( 5 Mm EDTA , 2%FBS in PBS ) or RPMI-10 . Newborn and fetal thymi were treated similarly except that 500 μl of digestion buffer was used . TECs used for sorting were enriched by EasySep APC positive selection Kit ( Stemcell Technologies ) after staining with an APC-conjugated anti-EpCAM antibody . Total lung cells were isolated according to a published protocol with modifications[60] . Briefly , lung was cut into approximately 1–2 mm pieces and then digested in 2 ml digestion buffer ( 500 μl 10 mg/ml collagenase type IV , 10 μl 50 mg/ml DNase I and 1 . 5ml 5% FBS IMDM ) at 37°C for 1 h with shaking every 10 min . Cells were washed with , and resuspended in , IMDM containing 5%FBS . Liver mononuclear cells were isolated using gradient centrifugation as previously described [61] . Fluorochrome-conjugated anti-CD45 . 2 ( clone 104 ) , CD45 ( clone 30-F11 ) , CD45 . 1 ( clone A20 ) , EpCAM/CD326 ( clone G8 . 8 ) , Ly51 ( clone 6C3 ) , MHCⅡ-I-A/I-E ( clone M5/114 . 15 . 2 ) , CD40 ( clone 3/23 ) , CD4 ( clone GK1 . 5 ) , CD8 ( clone 53–6 . 7 ) , TCR-β ( clone H57-597 ) , TCRγδ ( clone GL3 , NK1 . 1 ( clone PK136 ) , CD44 ( clone IM7 ) , CD24 ( clone M1/69 ) , CD25 ( clone PC61 . 5 ) , CD27 ( clone LG . 3A10 ) , c-Kit/CD117 ( clone 2B8 ) , CD62L ( clone MEL-14 ) , Gr1 ( clone RB6-8C5 ) , CD11b ( clone M170 ) , CD11c ( clone N418 ) , F4/80 ( clone BM8 ) , CD1d ( clone 1B1 ) , CD1d tetramer ( NIH tetramer facility ) , B220 ( clone RA3-6B2 ) , CD19 ( clone 6D5 ) , TER119/Erythroid Cells ( clone TER-119 ) , CD3e ( clone 145-2C11 ) , Annexin-V , Streptavidin , CCR6 ( clone 29-2L17 ) , IFN-γ ( clone XMG1 . 2 ) , IL-17A ( clone TC11-18 H10 . 1 ) , Foxp3 ( clone FJK-61s , eBioscience ) , Anti-TCR-Vγ1 . 1 ( clone 2 . 11 , BioXcell ) , Vγ4 ( clone UC3-10A6 , BioXcell ) , Vγ5 ( clone 536 ) , Vγ7 ( clone F2 . 67 ) , and Vδ6 . 3 ( clone 17C , kindly provided by Dr . Pablo Pereira , Institut Pasteur , France ) were purchased from Biolegend unless indicated otherwise . Ulex Europaeus Agglutinin I ( UEA-1 , clone B-1065 ) was purchased from vector laboratories . Phospho-S6 ( Ser235/236 , d57 . 2 . 2E ) antibody was from Cell Signal Technology . FITC-conjugated TCR-Vβ usage kit , including anti-TCRβ2 ( clone B20 . 6 ) , β3 ( clone KJ25 ) , β4 ( clone KT4 ) , β5 . 1/5 . 2 ( clone MR9-4 ) , β6 ( clone RR4-7 ) , β7 ( clone TR310 ) , β8 . 1/8 . 2 ( clone MR5-2 ) , β8 . 3 ( clone IB3 . 3 ) , β9 ( clone MR10-2 ) , β10b ( clone B21 . 5 ) , β11 ( clone RR3-15 ) , β12 ( clone MR11-1 ) , β13 ( clone MR12-3 ) , β14 ( clone 14–2 ) , and β17a ( clone KJ23 ) were purchased from BD Pharmingen . The 17D1 ( anti-Vγ5Vδ1 ) monoclonal antibody was prepared from cultured hybridoma supernatant after ammonium sulfate precipitation and affinity purification with a goat anti-rat IgM ( Jackson ImmunoResearch Laboratories ) conjugated Sepharose-4B beads ( Amersham Pharmacia Biotech AB ) and was conjugated with biotin ( ProteoChem ) according to a manufacturer’s protocol . The 17D1 antibody detects Vγ6Vδ1 when pretreated with the anti-TCRγδ ( GL3 ) antibody [62] . Cells were stained for cell surface molecules using 2% FBS-PBS . Cell death was identified by using the Violet Live/Dead cell kit ( Invitrogen ) or annexin-V and 7-AAD . Intracellular staining for Foxp3 was performed using the eBioscience Foxp3 Staining Buffer Set . Phospho-S6 staining was performed using the BD Biosciences Cytofix/Cytoperm and Perm/Wash solutions . Stained samples were acquired on a FACS Canto-II ( BD Biosciences ) flow cytometer . Data was analyzed with FlowJo software ( Tree Star ) . All FCS files associated with data presented in this study have been deposited in the zenodo website ( http://zenodo . org/record/34843 or DOI URL: http://dx . doi . org/10 . 5281/zenodo . 34843 ) . Thymocytes and single cell suspensions from other organs were stimulated with phorbol 12-myrustate 13-acetatae ( PMA , 50 ng/ml ) plus ionomycin ( 500 ng/ml ) in the presence of GolgiPlug ( 1 ng/ml ) for 4 h . After cell surface staining , intracellular staining for IL-17A and IFN-γ was performed by using the BD Biosciences Cytofix/Cytoperm and Perm/Wash solutions . For thymocytes , mice were i . p . injected with 5-bromo-2-deoxyuridine ( BrdU , Sigma; 1 mg/mouse or 50 mg/kg in 100–200 μl PBS ) and were stained 4 h after injection to assess BrdU incorporation . For embryonic ETPs , mice pregnant for 16 d were i . p . injected with 3 mg BrdU , and fetal thymi were collected 6 h after injection to assess BrdU incorporation . For TECs , mice were i . p . injected with 1 mg BrdU 3 times every 24 h , and thymi were collected 14 h after the last injection for TEC preparation . After cell surface staining , cells were intracellularly stained for BrdU using a BrdU Flow Kit ( BD Biosciences ) according to the manufacturer’s protocol . Thymus for H&E staining were fixed in 10% formalin solution for 1 d , and then changed into 70% ethanol . Paraffin-thin sections were stained with H&E according to standard protocols . Thymus for immunofluorescent staining were embedded in OCT ( Leica Biosystems Richmond Inc , Richmond ) and frozen immediately in −80°C . Frozen thin sections ( 5 μm ) were fixed in a 1:1 mixture of acetone and methanol at −20°C for 8 min . After blocking with PBS containing 3% BSA with 0 . 1% Tritonx-100 for 30–45 min at room temperature ( RT ) , samples were stained using primary rat-anti-mouse-keratin 8 ( Troma-1 , DSHB , University of Iowa , 1:50 dilution ) and rabbit-anti-mouse-keratin 5 ( PRB-160P , Covance; 1:200 dilution ) , followed by secondary Rhodamine-conjugated-donkey anti-rabbit IgG ( 1:400 dilution ) and FITC-conjugated-goat anti-rat IgG ( 1:400 dilution , Santa Cruz Biotechnology ) . Samples were mounted with Vector mounting solution containing DAPI ( Vector ) and allowed to dry overnight at RT or 4°C in the dark before imaging . Images were acquired using a Zeiss ApoTome Microscope and analyzed using PhotoshopCS4 software . CD45 . 1+CD45 . 2+ WT bone marrow cells were depleted off T cells using a PE-conjugated anti-CD3 antibody , anti-PE-antibody conjugated magnetic beads , and LD columns ( Miltenyi Biotec ) according to the manufacturer’s protocol . CD45 . 2+ Rptf/f and Rptf/f-Foxn1Cre mice were lethally irradiated ( 1000 rad ) and were intravenously injected with 1 . 0 x 107 T cell depleted bone marrow cells 4 h after irradiation . Recipient mice were analyzed 5–6 wk later . Single cell suspension made for TEC preparations from 3-wk-old Rptf/f-Foxn1Cre and WT littermates were plated at 1 x 107 cells/well in 96-well U-bottom plates and treated with or without fluorescent 100 μM 2- ( N- ( 7-Nitrobenz-2-oxa-1 , 3-diazol-4-yl ) Amino ) -2-Deoxyglucose ( 2-NBDG; Life Technologies ) in PBS and incubated at 37°C with 5% CO2 for 30 min . The 2-NBDG uptake reaction was stopped by removing culture medium and washed with pre-clod PBS two times . Cells were stained for surface molecules before analysis with flow cytometry . Total RNAs were extracted from E16 thymi or sorted 10 d-old mTECs ( Epcam+CD45−UEA-1+Ly51− ) and cTECs ( Epcam+CD45−UEA-1+Ly51− ) using the Trizol reagent ( Sigma ) . The first strand cDNAs were reversely transcribed using an iScript cDNA Synthesis Kit ( Bio-Rad ) . Quantitative real-time PCR ( qRT-PCR ) was performed using a Mastercycler Realplex ( Eppendorf ) and the SensiMix SYBR No-ROX Kit ( Bioline ) . Data were analyzed using the 2-Ct method after normalization to β-actin expression and shown as relative expression levels . For semi-quantitative PCR , cDNA template in 1:4 serial dilutions from each sample was used . For Vα14 to Jα2 , Jα18 , and Jα56 recombination , genomic DNA isolated from sorted CD4+CD8+ DP thymocytes from Rptf/f and Rptf/f-Foxn1Cre mice was used as the template for semi-quantitative PCR according to a previously published protocol except that TSC1 was used as loading control [61] . Primer pairs used for the amplification are summarized in Table 1 . Data were presented as mean ± SEM and analyzed for statistical differences using the Prism 5/GraphPad software . Comparisons were made using two-tailed Student’s t test . p-Values less than 0 . 05 were considered significant .
|
The thymus is the primary organ for T cell generation . Abnormal thymus function profoundly affects host immunity and numerous diseases . Thymopoiesis and thymus function rely on orchestrated interaction between multiple cell types representing different origins . Among them , thymic epithelial cells ( TECs ) are crucial for thymus development and maintenance and T cell generation . How TEC development and function are regulated is poorly understood . The mammalian/mechanistic target of rapamycin ( mTOR ) , a serine/threonine kinase , signals with two complexes , mTORC1 and mTOC2 , to control metabolism , growth , proliferation , and survival . Using a mouse model with mTORC1 selectively ablated in TECs , we demonstrate that mTORC1 in TECs plays critical roles in thymopoiesis and thymus function . Absence of mTORC1 results in impaired TEC maturation and function , altered thymic architecture , severe thymic atrophy , and impaired development of virtually all T-cell lineages . Moreover , it also causes increased generation of IL-17–producing γδT ( γδT17 ) cells and fetal-specific γδT subsets in adult thymus , revealing that mTORC1 in TECs is central for temporal control of γδT17 differentiation and TCRVγ/δ recombination . Our results establish mTORC1 as a central regulator for TEC development/function and for the establishment of normal thymic environment for proper T cell development . We suggest modulating mTORC1 activity as a strategy for preventing thymic involution/atrophy .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"blood",
"cells",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"pathology",
"and",
"laboratory",
"medicine",
"atrophy",
"immunology",
"cloning",
"age",
"groups",
"adults",
"signs",
"and",
"symptoms",
"stem",
"cells",
"molecular",
"biology",
"techniques",
"cytotoxic",
"t",
"cells",
"research",
"and",
"analysis",
"methods",
"specimen",
"preparation",
"and",
"treatment",
"staining",
"white",
"blood",
"cells",
"animal",
"cells",
"t",
"cells",
"thymus",
"molecular",
"biology",
"hematopoietic",
"progenitor",
"cells",
"immune",
"system",
"people",
"and",
"places",
"thymocytes",
"cell",
"staining",
"cell",
"biology",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"population",
"groupings"
] |
2016
|
mTORC1 in Thymic Epithelial Cells Is Critical for Thymopoiesis, T-Cell Generation, and Temporal Control of γδT17 Development and TCRγ/δ Recombination
|
Establishing the sources of reinfestation after residual insecticide spraying is crucial for vector elimination programs . Triatoma infestans , traditionally considered to be limited to domestic or peridomestic ( abbreviated as D/PD ) habitats throughout most of its range , is the target of an elimination program that has achieved limited success in the Gran Chaco region in South America . During a two-year period we conducted semi-annual searches for triatomine bugs in every D/PD site and surrounding sylvatic habitats after full-coverage spraying of pyrethroid insecticides of all houses in a well-defined rural area in northwestern Argentina . We found six low-density sylvatic foci with 24 T . infestans in fallen or standing trees located 110–2 , 300 m from the nearest house or infested D/PD site detected after insecticide spraying , when house infestations were rare . Analysis of two mitochondrial gene fragments of 20 sylvatic specimens confirmed their species identity as T . infestans and showed that their composite haplotypes were the same as or closely related to D/PD haplotypes . Population studies with 10 polymorphic microsatellite loci and wing geometric morphometry consistently indicated the occurrence of unrestricted gene flow between local D/PD and sylvatic populations . Mitochondrial DNA and microsatellite sibship analyses in the most abundant sylvatic colony revealed descendents from five different females . Spatial analysis showed a significant association between two sylvatic foci and the nearest D/PD bug population found before insecticide spraying . Our study shows that , despite of its high degree of domesticity , T . infestans has sylvatic colonies with normal chromatic characters ( not melanic morphs ) highly connected to D/PD conspecifics in the Argentinean Chaco . Sylvatic habitats may provide a transient or permanent refuge after control interventions , and function as sources for D/PD reinfestation . The occurrence of sylvatic foci of T . infestans in the Gran Chaco may pose additional threats to ongoing vector elimination efforts .
Disease eradication or elimination programs depend on time-limited intensive campaigns and are likely to fail if resistance to insecticides or drugs ( i . e . , malaria ) or sylvatic transmission cycles ( i . e . , yellow fever ) occur . Chagas disease is the most important vector-borne disease in Latin America in terms of disability-adjusted lost years , with an estimated 10–18 million people infected with Trypanosoma cruzi [1] . Elimination of domestic or peridomestic ( hereafter abbreviated D/PD ) populations of the insect vectors of T . cruzi through residual spraying with insecticides has shown varying degrees of success depending on the species and the occurrence of sylvatic foci . Several vector species occupy sylvatic habitats and show different degrees of domestication , such as T . dimidiata in Central America , Panstrongylus megistus , T . brasiliensis and T . pseudomaculata in Brazil , Rhodnius ecuadoriensis in northern Peru and Ecuador , and T . pallidipennis and related species in Mexico [2]–[4] . Species of sylvatic or peridomestic triatomines that were not recognized as control targets have emerged as primary vectors of T . cruzi in geographically defined areas over the last two decades [e . g . , 5] . For species such as R . prolixus , house reinfestations may also be driven by invasion from peridomestic or sylvatic foci [6] . Triatoma infestans historically is the main vector of human T . cruzi infection . In 1991 , this species was the target of a regional elimination program ( the Southern Cone Initiative ) that interrupted vector- and blood-borne transmission to humans in Chile , Uruguay , Brazil , eastern Paraguay and parts of Argentina [7] . However , only limited success in the elimination of T . infestans and interruption of vector-borne transmission has been achieved in the Gran Chaco region due to repeated reinfestations even in areas under intensive professional vector control [8] . The Gran Chaco , an ecoregion of 1 . 3 million km2 mainly spanning northern Argentina , Bolivia and Paraguay , has high levels of poverty and is hyperendemic for Chagas disease [9] . Recurrent reinfestation after residual spraying with insecticides and lack of a sustainable vector surveillance program result in renewed parasite transmission 3–5 years after community-wide vector control campaigns [10]–[13] . The obstacles to the elimination of T . infestans in the Gran Chaco may stem from different processes yet to be identified conclusively . The Southern Cone Initiative for the elimination of T . infestans was based on two major assumptions with wide consensus and limited supporting evidence [14] , [15]: ( i ) the species was restricted to D/PD habitats [16]–[19] , with true sylvatic foci only occurring in rock piles associated with wild guinea pigs in the Cochabamba and Sucre Andean valleys in Bolivia [20]–[22] , and ( ii ) T . infestans had low genetic variability and therefore was very unlikely to develop resistance to modern pyrethroid insecticides . Rare findings of T . infestans in sylvatic habitats up to the early 1980 s were judged to be of little relevance by several investigators ( reviewed in [23] , [24] ) . The surprising finding of melanic forms ( “dark morphs” ) in isolated dry forests in the Bolivian [23] , [25] and Argentine Chaco [24] , and more recently in the Paraguayan Chaco [26] , combined with the discovery of sylvatic foci with normal phenotypes in Chile and Bolivia [27]–[29] challenged the highly domesticated status of T . infestans . In addition , recent evidence showed T . infestans had richer genetic variability than previously assumed [30]–[33] , with strong chromosomal and DNA content differences between T . infestans from different sources [34] , whereas pyrethroid resistance emerged in northwestern Argentina and throughout Bolivia since the late 1990 s [35] , [36] . Understanding the ecological dynamics of reinfestation in insecticide-treated villages and untangling the mechanisms underlying the observed patterns is crucial for devising improved vector control tactics and the eventual elimination of T . infestans and other major triatomine vectors [18] , [37] . Genetic [38] and phenetic [39] , markers combined with carefully georeferenced bug samples collected before and after control interventions , a geographic information system ( GIS ) and spatial statistics [41] provide the means to better understand reinfestation dynamics . Here we first integrate the use of all these tools to investigate house reinfestation dynamics in the context of control interventions . As part of a longitudinal project on the eco-epidemiology and control of Chagas disease in a well-defined rural area in the dry Argentine Chaco [8] , we detected isolated findings of adult T . infestans and recently established , very low-density D/PD colonies during two years after a community-wide residual spraying of pyrethroid insecticides of all houses . To identify the putative sources for such occurrences and the sylvatic vectors of T . cruzi [42] , we conducted intensive surveys for triatomine bugs in diverse sylvatic habitats after interventions and surprisingly found various sylvatic foci of T . infestans . Using fine-resolution satellite imagery , GIS , spatial statistics , genetic markers and wing geometric morphometry , we investigated the relatedness between sylvatic and D/PD populations of T . infestans and the threat that they may represent to vector control and elimination attempts in the Argentinean Chaco . Based on previous findings of sylvatic T . infestans in the Bolivian Chaco [43] and of an isolated adult specimen of T . infestans infected with T . cruzi in semi-sylvatic habitats of our study area in the mid-1980 s [44] , we speculated that similar foci might exist in the Argentinean Chaco and that Triatoma guasayana was a likely candidate sylvatic vector of T . cruzi given its high abundance , widespread occurrence and occasional infection with the parasite [43] , [45] , [46] .
Field studies were carried out in Amamá ( 27° 12′ 30″S , 63° 02′ 30″W ) and neighboring rural villages in a 650 km2 area situated in the Moreno Department , Province of Santiago del Estero , Argentina ( Figure 1 ) . This area is located in the dry Chaco ecoregion [42] and its history of infestation since the mid-1980 s has been described elsewhere [8] . Based on the history of control interventions , the study area was subdivided into core ( 5 villages , 143 domiciles and 790 peridomestic sites ) and peripheral ( 7 villages , 132 houses and 709 peridomestic sites ) areas with all sites georeferenced . In April 2004 , community-wide residual spraying with 2 . 5% deltamethrin ( K-Othrin , Bayer ) of nearly all houses was conducted by professional vector-control personnel using a standard insecticide dose in domiciles ( 25 mg/m2 ) and standard or double dose in peridomestic sites for enhanced impact . Here we only report results from the core area ( villages of Amamá , Trinidad , Mercedes , Villa Matilde and Pampa Pozo; Figure 1 ) because no systematic searches for bugs were performed in sylvatic habitats around the peripheral communities . Timed manual searches for triatomine bugs with a dislodging spray ( 0 . 2% tetramethrin , Espacial ) were conducted in all domestic ( 0 . 5 person-hour ) and peridomestic sites ( one person-hour per house compound ) from all study villages in October 2004 , April and December 2005 , and November 2006 as described before [10] . In the core area , 143 domiciles and 764 peridomestic sites were inspected for triatomine bugs at least once between 2004 and 2006 . All detected foci were immediately sprayed with deltamethrin using the same procedures . As part of an ongoing monitoring program , discriminant dose assays demonstrated that no pyrethroid resistance occurred in local populations of T . infestans ( María Inés Picollo , unpublished results ) . We conducted four intensive surveys of triatomine bugs in sylvatic habitats using mouse-baited ( Noireau ) traps fitted with adhesive tape ( Plasto® , Brazil ) [47] in October and November 2005 , April and November-December 2006 as described before [24] . Mean temperatures varied between 23°C and 26°C in October-December ( spring ) surveys , and were below 20°C in April ( fall ) . Searches for sylvatic triatomine foci were conducted in 15 sampling areas that included representative forest sections with different degrees of disturbance ( i . e . , degraded forest under logging operations , cleared sections , ecotones , and implanted grasslands preceded by selective deforestation ) and in all sorts of refuges potentially suitable for triatomine bugs . The total capture effort was 598 trap-nights ( range per survey , 129 to 169 ) . Traps were usually placed far from houses in holes of fallen or standing trees ( live or dead ) , trunks or tree stumps and in between terrestrial bromeliads ( Bromelia serra and Bromelia hieronymi ) , cacti ( Opuntia quimilo and Opuntia ficus-indica ) or piles of shrubs ( Figure S1 ) . Traps were deployed when the weather was warm and not rainy approximately between 17 . 00–18 . 00 hs and retrieved before 10 . 00 hs to protect mice from exposure to extreme temperatures . All trap locations were georeferenced using a GPS ( Garmin , Etrex Legend C ) . All sylvatic sites surveyed in October and November 2005 were different except one , and 98% of them were re-inspected with mouse-baited traps on April 2006 to assess bug occurrence , persistence and invasion . The survey conducted in November-December 2006 only included sites that had not been surveyed previously . Flight-dispersing triatomine bugs were collected using black-light traps [48] placed in 36 georeferenced sylvatic sites where concurrent searches with mouse-baited traps were made ( i . e . , in the same areas ) . Light traps were deployed away from houses in habitats where there was a wide opening in the forest that allowed at least a 100 m visibility . Light traps were operated from approximately 19:45 ( i . e . , 15 min before sunset ) to 22:00–23:00 hs because the flight activity of T . infestans peaks during the first hour after sunset , and is more likely to occur when air temperature exceeds 20°C and wind speed is <5 km/h [48]–[50] . Suitable conditions for flight initiation of T . infestans occurred during the surveys conducted in October-November 2005 but not in April 2006 . All collected bugs were kept alive in plastic vials with folded filter paper , identified to species following Lent and Wygodzinsky [19] and counted . Species identification of very small first- or second-instar nymphs sometimes was considered tentative depending on the integrity of the material; no such doubts remained for third-instars or later stages . Fourth- or fifth-instar nymphs and adult bugs collected in 2005 were individually weighed on an electronic balance ( OHAUS , precision , 0 . 1 mg ) and total body length ( L ) measured from the end of the clipeus to the end of the abdomen with a vernier caliber ( precision , 0 . 02 mm ) to estimate a weight-to-length ratio ( W∶L ) –a quantitative index of nutritional status . The qualitative nutritional status of nymphs was determined by a cross-sectional view of the abdomen and cuticle distension and classified into four categories that ranged from unfed to large blood contents [51] . Feces from live third-instars and larger stages were examined microscopically for T . cruzi infection at 400× magnification . DNA from bugs assigned to T . infestans ( based on morphological evidence ) was obtained , PCR-amplified , and sequenced for a 661 bp fragment of the mitochondrial genes cytochrome oxidase I ( mtCOI ) [32] and a 572 bp fragment of the cytochrome B ( mtcytB ) gene [52] . Sequences from sylvatic T . infestans were compared with Triatoma spp sequences available at Genbank and from previous surveys on the instraspecific variability of T . infestans [32] , [53]–[56] . Sylvatic T . infestans mtCOI plus mtcytB composite haplotypes were compared with previously recorded haplotypes of D/PD T . infestans from the study villages ( collected in 2001–2002 ) , from other more distant ( 40 km ) localities within Santiago del Estero Province ( Quilumpa , Km 40 , La Loma and Invernada Norte , collected in 2003–2004 ) , and from other Argentinean Provinces more than 300 km apart ( Salta , La Rioja , Tucumán and Formosa , collected in 2000–2005 ) . A detailed description of the source localities was published elsewhere [32] . Genetic variability was estimated as the mean number of pairwise differences per site ( π ) , Watterson's estimator ( θW ) and the haplotype diversity ( Hd ) with DnaSP 5 . 0 [57] and a statistical parsimony haplotype network was built with TCS 1 . 21 [58] . For higher resolution of the relationships between sylvatic and D/PD populations of T . infestans , the multilocus ( ML ) genotype for 10 microsatellite loci was obtained for sylvatic T . infestans using primers and PCR conditions previously described [59] . ML genotypes were compared with those from T . infestans captured in D/PD sites from Amamá and neighboring villages in October 2002 and April 2004 before full-coverage insecticide spraying [60] . Inter-individual genetic distance based on the complement of the proportion of shared alleles [61] was estimated with MICROSAT 1 . 5d ( http://hpgl . stanford . edu/projects/microsat/ ) , and a neighbor-joining ( NJ ) tree was built with MEGA 3 . 0 [62] . Using the genotypes of local D/PD T . infestans as reference populations , we applied the Bayesian based assignment-exclusion test implemented in GENECLASS 2 [63] to individually assign sylvatic individuals to the local pre-spraying D/PD populations ( defined as the total gene pool at a given community in each capture date ) . No post-spraying reference groups could be formed because after community-wide insecticide spraying ( 2004–2006 ) most bug collections contained one or a few insects per site that were sparsely distributed throughout the communities ( i . e . , no established populations of T . infestans were detected ) . Reference populations were not excluded as the putative origin of the sylvatic insects when the marginal probability exceeded 0 . 05 . We used 100 , 000 replications and a simulation algorithm [64] . Sibship of T . infestans bugs collected in traps with more than one individual ( TN-92 and TN-139 ) was inferred with the maximum likelihood approach implemented in COLONY 2 . 0 [65] performing two independent runs and assuming a probability of null alleles of 0 . 05 in loci ms42 , ms64 and ms65 due to departures from Hardy-Weinberg expectations . The wing geometric morphometry of the only sylvatic T . infestans male collected was compared with T . infestans males captured in D/PD sites from Amamá and neighboring study villages in October 2002 ( n = 87 ) and April 2004 ( n = 74 ) as described elsewhere [66] . The geometric coordinates of 11 type-I landmarks ( venation intersections ) from all right wings were digitized by the same user ( JSB ) . After performing the generalized Procrustes superposition ( GPA , [67] ) , the residual coordinates of the total sample ( including the sylvatic specimen ) were transformed into partial warps ( PW ) . These shape variables allow standard statistical analyses such as principal component ( PCA ) or discriminant analyses ( DA ) . To cope with small sample sizes in some villages , the first nine principal components of the PW were used as input for a DA performed on the village samples ( excluding the sylvatic specimen ) . These principal components are also called relative warps ( RW ) . The sylvatic specimen was then used as supplementary data and its position in the morphospace examined in terms of Mahalanobis distances . Digitization , GPA , PCA and DA were performed using the corresponding modules of the CLIC package [68] . Global positioning system readings from all sampling sites ( with mouse-baited and light traps ) were integrated into a Geographic Information System ( ArcGIS 9 . 1 , ESRI , Redlands , CA , U . S . A . ) of the study communities containing a georeferenced satellite image ( Ikonos2 , Space Imaging Inc . , Atlanta , GA , U . S . A . ) and the position of all houses and peridomestic sites sprayed with insecticides in 2004 . Cartesian coordinates ( Universal Transverse Mercator , UTM , Zone 20S ) were calculated for each D/PD site and trapping location in order to perform spatial analysis . A focal spatial statistic ( Gi ( d ) ) [69] was used to determine the presence and extent of spatial clustering of T . infestans D/PD abundance ( average of timed manual catches of bugs per site in 2002 and 2004] around each T . infestans-positive sylvatic focus ( point i ) . This local statistic is additive in the sense that it focuses on the sum of the j values in the vicinity of point i . Hence , we took each T . infestans-positive sylvatic focus , one at a time , and searched the nearby area for occurrences of more or fewer D/PD T . infestans bugs collected before full-coverage insecticide spraying than expected by random . This procedure identified specific trap locations as members or non-members of infestation clusters . We used a binary weight wij based on a distance threshold ( d ) scheme . Clustering of D/PD T . infestans abundance around a positive sylvatic site occurred when the observed Gi was higher than 2 . 32 ( the expected value at P<0 . 01 ) . We evaluated the value of Gi up to 3 km from each sylvatic site with T . infestans –a tentative upper bound of the flight range of T . infestans . Analyses were performed using the software Point Pattern Analysis [70] . Humane care and use of laboratory animals were performed according to Institutional Animal Care and Use Committee ( IACUC , CICUAL in Spanish ) guidelines at UBA's Faculty of Exact and Natural Sciences . Animal care and use is guided by the International Guiding Principles for Biomedical Research Involving Animals developed by the Council for International Organizations of Medical Sciences .
A total of 13 ( 9 . 1% of 143 domiciles ) domestic foci of T . infestans with 23 bugs and 38 ( 5 . 0% of 764 sites ) peridomestic foci with 223 bugs were detected between 2004 and 2006 after full-coverage spraying with deltamethrin . Nearly 25% of all collected T . infestans were adult bugs . Only 30 ( 5% ) of 598 mouse-baited traps deployed overnight in sylvatic habitats were positive for triatomine bugs ( Table 1 ) . Six sylvatic foci of T . infestans with normal chromatic characters ( totaling 23 nymphs and 1 male; range per site , 1–17 ) were found in tree holes or trunks ( Figures S1 and S2 ) . Another probable sylvatic foci of this species with two first- or second-instar nymphs was conservatively excluded because the morphological identification of these stages was uncertain and mtDNA markers did not amplify; this probable focus occurred in the vicinity of the largest sylvatic colony of T . infestans ( trap TN-139 , Figure S2 ) . The apparent density of sylvatic T . infestans was 4 per 100 trap-nights ( 24 bugs in 598 trap-nights; mean ± SD , 3 . 8±6 . 4 bugs per site ) . One sylvatic focus located west of Amamá ( trap TN-139 ) was infested both in October ( 1 male ) and November 2005 ( 14 first- or second-instar nymphs and 2 fourth-instars ) and was taken as one colony . No T . infestans bugs were collected with mouse-baited traps in April or November 2006 . T . guasayana occurred more frequently ( 3 . 0% of mouse-baited traps in all surveyed habitats ) than T . infestans ( 1 . 2% , Table 1 ) . Feces and hairs of Didelphis opossums were found in one T . guasayana focus . All first- or second-instars of Triatoma sp . not identified to species level most likely were T . guasayana based on morphology , size and type of habitat . Light-trap collections yielded 110 adult T . guasayana , one specimen of T . garciabesi ( female ) and one of T . platensis ( male ) , and no T . infestans in 41 light-trap-nights ( Table 1 ) . Of the 41 light-trap nights , 28 ( 68 . 3% ) were positive for triatomine bugs . The adult sex ratio in T . guasayana was 1∶2 . 2 ( male to female ) . Sylvatic foci of T . infestans occurred at 5 sampling areas located 2 . 0–11 . 5 km apart ( Figure 2 ) . Most triatomines ( 17 or 70 . 8% of 24 T . infestans and 18 or 64 . 3% of 28 T . guasayana ) caught with mouse-baited traps occurred in areas that had been deforested selectively ( totalling 40 bugs at 11 sites ) ; the other seven T . infestans were caught in secondary forest with medium-sized or a few large-sized trees . The only three T . garciabesi found were caught in mature forest under active deforestation . The remaining triatomine bugs were caught in secondary forest with medium- or large-sized trees . The main identified micro-habitats of T . infestans were in holes of fallen trees and decaying tree trunks lying on the ground ( 21 or 87% of 24 bugs collected ) , a tree stump and a live standing tree . These ecotopes included 4 ‘quebracho colorado’ ( Schinopsis lorentzii ) and 2 ‘mistol’ ( Zizyphus mistol ) trees ( Figure S1 ) . Nearly all triatomine bugs caught with mouse-baited traps and examined for qualitative nutritional status ( n = 36 ) were unfed ( 61 . 1% ) or had very little remnants of a blood meal ( 33 . 3% ) and very low W/L ratios ( Table S1 ) . Of 140 sylvatic triatomine bugs examined microscopically ( 10 T . infestans , 21 T . guasayana and 3 T . garciabesi caught with mouse-baited traps and 106 T . guasayana collected with light traps ) none was found microscope-positive for T . cruzi . The morphological identification of 20 sylvatic bugs as T . infestans was confirmed by DNA sequencing of mtCOI and/or mtcytB fragments; DNA from six other bugs ( all first- to third-instars identified as T . infestans based on morphological characters ) could not be amplified . The two third-instar nymphs not amplified were taken as T . infestans because a morphological misidentification ( relative to the locally known species ) was considered very unlikely . None of the sylvatic T . infestans bugs carried the T_C change at position 556 , which is characteristic of T . platensis and is absent in a large sample of T . infestans from Argentina , Bolivia , Peru , and Uruguay [32] . Sylvatic T . infestans with mtCOI and mtcytB composite haplotypes ( n = 16 , Table S2 ) exhibited high nucleotide variability ( θW = 0 . 006 , π = 0 . 007 ) and haplotype diversity ( Hd = 0 . 901 ) . No shared haplotypes were found among bugs from different traps , whereas traps with more than one bug had one ( TN-92 , n = 3 ) and five ( TN-139 , n = 11 ) different haplotypes ( Table S2 ) . Of eight sylvatic haplotypes identified , six were exclusive of sylvatic bugs whereas two haplotypes were recorded in local peridomestic populations of T . infestans and elsewhere in Argentina ( Figure 3 ) . Sylvatic haplotypes were spread along the entire statistical parsimony network; they did not form a unique cluster separated from the rest and were more closely related to D/PD than to other sylvatic haplotypes ( Figure 3 ) . One sylvatic haplotype was highly divergent ( am-XIV ) but also was closely connected to an Amamá peridomestic haplotype ( haplotype b-XIV ) . The multilocus ( ML ) genotype for 10 microsatellite loci was obtained for 21 sylvatic T . infestans . We identified a total of 86 different alleles for the 10 loci , of which only 15 ( 17 . 5% ) and 17 ( 19 . 8% ) were private alleles not detected in the local D/PD populations in 2002 and 2004 , respectively . Sylvatic T . infestans clustered among D/PD bugs with no sharp discontinuity ( Figure 4 ) . T . infestans bugs captured concurrently at trap TN-139 clustered together whereas bugs collected there at different times were more closely related to different clusters of Amamá peridomestic bugs ( i . e . , the closest village ) . In addition , insects from trap TN-139 had five different mtCOI-mtcytB haplotypes ( Table S2 ) . Sibship microsatellite analyses showed that bugs that shared a mitochondrial haplotype ( or that had consistent haplotypes because of missing data for mtCOI or mtcytB ) were most likely full- or half-sibs whereas bugs with different haplotypes were not ( Tables S3 and S4 ) . Bugs from trap TN-92 clustered together and closely to bugs from Mercedes village ( where the trap was located ) and from another village at ∼5 km ( Pampa Pozo ) . These three bugs were full- or half-sibs and shared the same mitochondrial haplotype ( Tables S3 and S4 ) . The bug from site trap TN-182 was grouped with bugs from the nearest village ( Mercedes ) located at ∼8 km . The bug collected at trap TN-101 ( close to Villa Matilde , Fig . 1 ) clustered with bugs from Amamá and Pampa Pozo . The Bayesian-based assignment-exclusion test indicated that 18 of 21 sylvatic ML genotypes were not excluded from one or more of the D/PD reference populations ( Table 2 ) . D/PD populations were excluded as putative sources for three sylvatic insects captured in two different sites ( traps TN-182 and -139 ) . The mtCOI-mtcytB haplotype from the bug in trap TN-182 ( al VII , Figure 4 ) was also genetically distant from the local D/PD populations and was closely related to D/PD populations from La Rioja , more than 400 km far from the study area ( Figure 4 ) . Wing geometric morphometry was used to compare the only sylvatic T . infestans male collected ( trap TN-139 ) with T . infestans males captured in local PD sites in 2002 and 2004 . The factorial map showed that the sylvatic bug clearly overlapped with 2002 PD bugs from Amamá –the closest village to its capture site ( Figure 5 ) and it was also assigned to 2004 PD bugs from Amamá ( not shown ) . All sylvatic foci of T . infestans were located 110–2 , 300 m from the nearest D/PD sites ever found to be infested by this species after full-coverage insecticide spraying ( i . e . , detected during the preceding 18 months ) ( Figure 2 ) . Trap location TN-182 included two T . infestans-positive sites ( TN-180 and TN-182 ) that were analyzed together because their separation ( 13 m ) was smaller than the distance resolution of the Gi ( d ) test ( 50 m ) . The distance between traps positive for T . infestans to the nearest house varied from 125 to 1 , 900 m . Spatial analysis showed a statistically significant association ( Gi ( d ) >2 . 32 , P<0 . 01 ) between two sylvatic foci of T . infestans found within three km of a D/PD site and the average timed-manual catch of bugs before insecticide spraying ( Figure 2 ) . Significant clustering occurred up to 1 . 2 km in Amamá ( trap TN-139 , with 17 insects ) and up to 150 m in Mercedes ( trap TN-101 , with one third-instar nymph ) ( Figure 2 ) . The remaining three sylvatic foci of T . infestans ( TN-182 , TN-180 and TN-92 ) were located at 430–1 , 846 m from the nearest infested house , but did not appear to be significantly associated with any of them ( Gi ( d ) >1 . 96; P>0 . 05 ) .
A long-standing , key scientific question with vast implications for vector control is what is the source of the triatomine bugs appearing after community-wide insecticide spraying [18] , [37] , [81] . Are they ( i ) survivors or the offspring of previously existing bugs; ( ii ) immigrants from untreated D/PD or sylvatic foci; or ( iii ) migrants brought by passive transport from other villages or elsewhere ? This issue is applicable to all major triatomine vector control programs throughout Latin America and the responses may differ between settings and even within the same species , as with T . dimidiata in Central America and Mexico or T . brasiliensis and P . megistus in Brazil –all of which display substantial within-species differences in habitat distribution , invasive capacity and other relevant traits . As with other species of triatomine bugs , T . infestans adults and nymphs are attracted to lights [48] , [82] . Sylvatic populations of T . infestans are much more widespread than assumed in the past [23]–[29] and have recently been discovered in the Paraguayan Chaco [26] . Because sylvatic habitats are not targeted for vector control operations , they may provide hidden refuges for T . infestans from which they may reinvade houses in search of more suitable conditions and resources . Our results suggest that in areas with recurrent reinfestation , vector control programs should consider the potential occurrence of external sources ( semi-sylvatic or sylvatic ) around the target community . The role that sylvatic populations of T . infestans ( either with melanic or normal phenotype ) play in the process of recolonization of insecticide-treated villages and their invasive capacity needs to be more widely investigated to evaluate the risk they pose to effective vector control and eventual elimination in the Gran Chaco and elsewhere .
|
Triatoma infestans , a highly domesticated species and historically the main vector of Trypanosoma cruzi , is the target of an insecticide-based elimination program in the southern cone countries of South America since 1991 . Only limited success has been achieved in the Gran Chaco region due to repeated reinfestations . We conducted full-coverage spraying of pyrethroid insecticides of all houses in a well-defined rural area in northwestern Argentina , followed by intense monitoring of house reinfestation and searches for triatomine bugs in sylvatic habitats during the next two years , to establish the putative sources of new bug colonies . We found low-density sylvatic foci of T . infestans in trees located within the species' flight range from the nearest infested house detected before control interventions . Using multiple methods ( fine-resolution satellite imagery , geographic information systems , spatial statistics , genetic markers and wing geometric morphometry ) , we corroborated the species identity of the sylvatic bugs as T . infestans and found they were indistinguishable from or closely related to local domestic or peridomestic bug populations . Two sylvatic foci were spatially associated to the nearest peridomestic bug populations found before interventions . Sylvatic habitats harbor hidden foci of T . infestans that may represent a threat to vector suppression attempts .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"spatial",
"and",
"landscape",
"ecology",
"medicine",
"infectious",
"diseases",
"public",
"health",
"and",
"epidemiology",
"chagas",
"disease",
"ecology",
"epidemiology",
"genetics",
"population",
"genetics",
"biology",
"population",
"ecology",
"parasitic",
"diseases",
"genetics",
"and",
"genomics"
] |
2011
|
Hidden Sylvatic Foci of the Main Vector of Chagas Disease Triatoma infestans: Threats to the Vector Elimination Campaign?
|
Enteropathogenic Escherichia coli ( EPEC ) use a needle-like injection apparatus known as the type III secretion system ( T3SS ) to deliver protein effectors into host cells . Effector translocation is highly stratified in EPEC with the translocated intimin receptor ( Tir ) being the first effector delivered into the host . CesT is a multi-cargo chaperone that is required for the secretion of Tir and at least 9 other effectors . However , the structural and mechanistic basis for differential effector recognition by CesT remains unclear . Here , we delineated the minimal CesT-binding region on Tir to residues 35–77 and determined the 2 . 74 Å structure of CesT bound to an N-terminal fragment of Tir . Our structure revealed that the CesT-binding region in the N-terminus of Tir contains an additional conserved sequence , distinct from the known chaperone-binding β-motif , that we termed the CesT-extension motif because it extends the β-sheet core of CesT . This motif is also present in the C-terminus of Tir that we confirmed to be a unique second CesT-binding region . Point mutations that disrupt CesT-binding to the N- or C-terminus of Tir revealed that the newly identified carboxy-terminal CesT-binding region was required for efficient Tir translocation into HeLa cells and pedestal formation . Furthermore , the CesT-extension motif was identified in the N-terminal region of NleH1 , NleH2 , and EspZ , and mutations that disrupt this motif reduced translocation of these effectors , and in some cases , overall effector stability , thus validating the universality of this CesT-extension motif . The presence of two CesT-binding regions in Tir , along with the presence of the CesT-extension motif in other highly translocated effectors , may contribute to differential cargo recognition by CesT .
Enteropathogenic and enterohemorrhagic Escherichia coli ( EPEC and EHEC ) cause acute gastroenteritis in humans and are a common source of outbreaks [1] . EPEC is a significant pathogen in the pediatric population , especially in areas with limited access to healthcare and clean water , whereas EHEC is a common food- or water-borne contaminant in industrialized nations [1] . EPEC and EHEC contain a genomic island called the locus of enterocyte effacement ( LEE ) that encodes a type III secretion system ( T3SS ) [2] necessary for the formation of attaching and effacing ( A/E ) lesions on epithelial cells [3] . The T3SS is a needle-like protein injectisome used by Gram-negative bacteria to deliver effector proteins into host cells directly from the bacterial cytosol [4] , where they target specific host processes to allow for attachment , survival , and propagation of the bacteria [5 , 6] . Enteric pathogens that use a T3SS for host attachment , infection , and/or colonization are significantly attenuated when lacking their encoded T3SS [7] , identifying it as a key mediator of host-pathogen interactions . Structural biology efforts have advanced our understanding of the assembly , structure , and function of the T3SS [8 , 9] . The T3SS contains ~25 proteins assembled into distinct structures including , ( i ) an extracellular needle filament capped at the distal end by hydrophobic translocon proteins , ( ii ) a basal body comprised of inner and outer membrane-spanning rings , ( iii ) an ATPase-containing sorting platform complex at the cytoplasmic face of the basal body , and ( iv ) cytosolic chaperones that bind , protect , deliver , and control effector secretion [10] . Three classes of non-flagellar T3SS chaperones have been described . Class I chaperones bind translocated effectors , class II chaperones bind the hydrophobic translocators , and class III chaperones escort and prevent cytosolic polymerization of the extracellular needle filament [11 , 12] . The class I chaperones are further subdivided into class IA and IB . Class IA chaperones are usually specific for one effector and are located adjacent to the gene that encodes the cognate effector [11 , 13] . Class IA chaperones that bind multiple effectors have been reported , including EPEC CesT and Salmonella SrcA [14 , 15] , and are referred to as multi-cargo chaperones [16] . Class IB chaperones bind multiple effectors and are usually encoded within large operons that contain structural components of the T3SS instead of being adjacent to a specific effector gene [11 , 13] . Class IB chaperones appear to be functionally interchangeable between species and recognize a specific sequence motif [17] . Multi-cargo chaperones play a significant role in T3SS-dependent infection biology as mutants lacking these proteins are attenuated in animal and plant models of infection [16] . CesT from EPEC and EHEC was originally thought to be a class 1A chaperone specific for the translocated intimin receptor ( Tir ) [18 , 19] . However , it was later reclassified as a multi-cargo chaperone because it interacts with at least 9 other effectors [14 , 20 , 21] , most of which require CesT for translocation into host cells [22 , 23] . Recent work has indicated that the effector binding and secretion activities mediated by CesT can be functionally separated . For example , mutants in the C-terminal domain of CesT retain their ability to bind effector cargo , yet exhibit reduced effector secretion [24] . This C-terminal domain was also identified as a site for tyrosine phosphorylation in a phosphotyrosine-proteome study [25] , in which tandem tyrosine phosphosites ( Y152 and Y153 ) influenced NleA or global effector secretion , respectively [26] . Furthermore , host-cell contact has been proposed to liberate free CesT in the bacterial cytosol that can then bind and antagonize CsrA repression of the nleA 5’UTR [27] . This interaction is facilitated by the C-terminal domain of CesT [28] , allowing for greater control over the timing and translocation efficiency of the NleA effector . Notwithstanding the requirement of CesT for effector secretion , Tir has been implicated in effector secretion hierarchy . Deletion of tir in the hyper-secreting ΔsepD strain of EPEC significantly reduced the level of at least 6 effectors in culture supernatants [21] . A similar but more modest effect was also seen for this subset of effectors translocated into host cells when only tir was deleted [23] . Given the primary importance of the Tir-CesT complex in orchestrating secondary effector secretion in E . coli , we initiated structural studies to characterize the Tir-CesT interaction and to delineate the role that this effector-chaperone pair plays in protein translocation . Here , we present the co-crystal structure of a C-terminal truncation of CesT in complex with an N-terminal fragment of Tir . This structure allowed us to define a CesT-extension motif , leading to the identification of a second CesT-binding region in the C-terminal domain of Tir , which we verified using biochemical and molecular assays . Furthermore , we identified the CesT-extension motif in the N-terminus of a subset of other effectors and demonstrated the function of this motif in effector translocation .
The first ~20 amino-terminal residues of E . coli T3SS effectors contain a T3SS-specific secretion signal that can be predicted bioinformatically [29] . Downstream of the T3SS secretion signal , but within the first ~100 residues , is an unspecified CesT-binding domain that has been identified in Tir , Map , and NleH [18 , 20 , 21] . Despite the fact that CesT binds to the N-terminus of these effectors , sequence alignments have not identified a consensus motif within this region . To determine the minimal recognition sequence of the CesT-binding region , various His6-tagged N-terminal Tir constructs were tested for their ability to co-purify CesT ( Fig 1 ) . Tir fragments containing residues 23–80 , 32–80 , and 35–77 co-purified CesT as seen by Ni2+-affinity pull-down and immunoblotting ( Fig 1B and S1A Fig ) , whereas CesT alone was never pulled-down in the absence of Tir by the Ni2+-affinity resin , thus confirming specificity of our assay . When these Tir fragments were truncated further to residues 32–73 , 37–80 , and 37–73 , they lost the ability to co-purify with CesT ( Fig 1B ) . To determine the molecular basis of the Tir-CesT interaction , we carried out crystallization trials with the three Tir fragments that co-purified CesT ( S1A Fig ) , however none of the complexes produced crystals . CesT contains a unique C-terminal extension that is not conserved among closely related chaperones , such as SrcA from Salmonella [15] . This C-terminal extension was shown to be important for effector secretion but was dispensable for effector binding [24] . We hypothesized that this C-terminal region of CesT was either disordered or heterogeneous from differential phosphorylation , possibly preventing favourable crystallization contacts . To address this , we truncated CesT at residue 138 ( CesT138 ) and tested whether this variant could co-purify the same Tir peptides as full-length CesT . Tir peptides 23–80 , 32–80 , and 35–77 retained their ability to co-purify with CesT138 indicating that the C-terminus of CesT was not required for this interaction ( Fig 1C and S1B Fig ) . However , the shorter Tir peptide 32–73 was now able to co-purify CesT138 ( Fig 1C and S1B Fig ) , whereas the Tir peptides 37–80 and 37–73 were unable to co-purify CesT138 . Gel filtration chromatography confirmed that both CesT and CesT138 existed in a dimeric configuration ( S2 Fig ) , which is the functional unit of T3SS chaperones [30] . Taken together , these data suggest that the minimal CesT-binding region is located between residues 35–77 and that the C-terminus of CesT may interfere with binding of Tir residues 73–80 . The structural basis for how CesT binds and interacts with multiple T3SS effectors is not known . To determine the molecular determinants behind this interaction we conducted crystallization trials for all of the successful Tir-CesT co-purifications ( S1 Fig ) . The Tir32-80-CesT138 complex produced crystals in the trigonal space-group P322 with one molecule of the complex in the asymmetric unit . Diffraction data were collected to 2 . 74 Å resolution and the structure was determined by molecular replacement ( Table 1 ) . Structural refinement produced a final model with good geometry and R factors ( Rwork and Rfree of 20 . 8% and 25 . 5% , respectively ) ( Table 1 ) . Residues 130–138 of CesT138 , the N-terminal histidine tag , and residues 32–34 , 54–64 , and 76–80 of Tir were not included in the final model due to poor or absent electron density . Tir32-80 adopts minimal regular secondary structure that is limited to two small β-strands , β1’ and β2’ ( Fig 2A ) . The Tir32-80 fragment binds CesT138 in two distinct locations and is separated by a break in the peptide chain likely due to residue mobility in the crystal . Tir residues 35–53 adopt a β-hairpin-like fold and extend the 5-stranded β-sheet core of CesT138 , while also being pinched between α1 and orthogonally below by α3 of CesT138 ( Fig 2A ) . The interaction between Tir32-80 β2’ and CesT138 occurs through a conserved 3-amino acid β-motif , adopting the consensus sequence of Φ-X4-Φ-x-Φ where x is any amino acid and Φ represents a hydrophobic residue . The β-motif was originally identified in the SipA-InvB complex [31 , 32] , but appears to be a conserved mode of binding present in all class I chaperone-effector complexes [33] . Slight differences have been observed in the β-motif , most notably that one to four residues can separate the first and second hydrophobic residues ( ie . Φ- ( X1-4 ) -Φ-x-Φ ) . Tir residues 65–75 are bound along the concave surface of the β-sheet core of CesT138 ( Fig 2A ) . Despite CesT138 having a global acidic surface potential ( Fig 2B ) , Tir32-80 binding is mediated through distinct hydrophobic-hydrophobic contacts ( Fig 2C ) . Specifically , Tir residues I38 ( purple ) , L44 ( cyan ) , and L49 ( cyan ) anchor the β-hairpin-like peptide to CesT138 ( Fig 2D ) ; and L69 plus three additional proline residues make a second point of contact with the β-sheet core of CesT138 ( Fig 2E ) . CesT and CesT138 form a dimer in solution ( S2 Fig ) consistent with previous reports [30] , and is a property conserved among T3SS class I chaperones . Although only one molecule of CesT138 was present in the crystallographic asymmetric unit , the dimer interface is clearly present along the principle 2-fold axis of symmetry ( S3A Fig ) . Recently , the structure of CesT in complex with CsrA was reported [28] . CesT in the CsrA-CesT complex also adopts the same dimer orientation as observed in the Tir32-80-CesT138 complex , providing further evidence that the domain swapped dimer of the previous unladen EHEC CesT structure is likely a crystallographic artifact ( S3 Fig ) . Furthermore , structural alignment of Tir32-80-CesT138 to the CsrA-CesT complex reveals significantly different binding modes for Tir and CsrA to CesT ( Fig 3A; a monomer of CesT is shown for simplicity ) . CsrA binds CesT predominantly through residue contacts along CesT α3 and α4 ( red ) , in which the latter comes from the second molecule of the CesT dimer ( Fig 3A ) . In contrast , Tir32-80 binds the cleft formed between CesT α1 and β1 ( Fig 3A ) . CsrA doesn’t directly occlude binding of Tir residues 35–53 to CesT , but residues K26 and R31 come within very close proximity , 2 . 9 Å and 2 . 2 Å , from Tir residues G45 and S46 , respectively ( Fig 3B ) . The C-terminus of CesT ( cyan ) from the CsrA-CesT complex , which is absent from CesT138 , self-associates by binding along the concave surface of CesT ( Fig 3A and 3C ) . Furthermore , the C-terminus of CesT also forms α4 ( red ) that interacts with CsrA , locking the C-terminal CesT peptide ( residues I32 to Y153 ) in place ( Fig 3A and 3C ) . Interestingly , the CesT C-terminus occupies the same binding surface as Tir residues 65–75 ( pink , Fig 3C ) , despite significantly different sequences . These findings likely explain why none of the Tir peptide-CesT complexes crystallized , as CesT residues I32 to Y153 would compete for the same binding grove as Tir residues 65–75 , and thus required the truncation of the CesT C-terminus ( CesT138 ) . Furthermore , this explains why only CesT138 could co-purify with Tir32-73 , as the self-associated C-terminus of CesT likely out-competes the smaller Tir peptide for binding along the same concave surface . Taken together , the structural data from the Tir32-80-CesT138 and CsrA-CesT complexes suggest that ( i ) CesT exists as an unswapped dimer , ( ii ) the Tir binding region of CesT exhibits significant plasticity that could accommodate the binding of multiple effectors with varying sequences , and ( iii ) the C-terminal extension of CesT is required for CsrA interaction that in turn could also prevent CesT from binding Tir . Tir32-80 binds the same hydrophobic surface in each monomer of the CesT138 dimer , producing a Tir32-80:CesT138 stoichiometry of 2:2 in solution , that was validated by gel filtration chromatography ( S2 Fig ) . This crystal packing orientation was observed for the chaperone-effector fragment complexes of SycH-YscM2 [34] , SycH-YopH [35] , and ShcA-HopA1 [33] . However , gel filtration chromatography of the Tir23-550-CesT complex suggests that only a monomer of Tir23-550 binds a dimer of CesT ( S2 Fig ) , consistent with a 1:2 Tir:CesT stoichiometry reported recently [36] . Since one molecule of Tir binds a dimer of CesT , but our crystallographic data suggest that two Tir32-80 fragments can bind a dimer of CesT , these data could be reconciled if full-length Tir contained a second uncharacterized CesT-binding region . Co-expression pull-down assays support this hypothesis as Tir81-550 , which lacks the N-terminal CesT-binding region , retained the ability to pull-down CesT ( Fig 4A ) . This was consistent with previous data showing that CesT can interact with N-terminal truncations of Tir in EHEC [37] . Furthermore , bacterial adenylate cyclase two hybrid ( BACTH ) assays showed that fusion of T18 to Tir23-80 , Tir23-550 , and Tir81-550 all had a positive interaction with CesT and CesT138 fused to T25 ( blue colonies ) , but not to T25 alone ( white colonies ) ( Fig 4B ) . We also observed CesT-CesT and Tir-Tir interactions in these assays consistent with previous reports [38 , 39] . To identify the second CesT-binding region of Tir we used Tir residues 29–80 to conduct sequence alignments on a sliding window of ~50–80 amino acids . The carboxy-terminus of Tir ( residues 490–550 ) had 20% sequence identity to Tir 29–80 and contained a TGRLIGT sequence similar to the sequence that forms β1’ in the N-terminal region of Tir ( Fig 4C ) . T18 fused to Tir490-550 showed a strong interaction with CesT fused to T25 by BACTH assays ( Fig 4B ) , and was able to pull-down CesT in co-expression pull-down experiments ( Fig 4D ) , confirming this site as a second CesT-binding region . Taken together these data suggest that Tir is unique among E . coli effectors in that it contains a second carboxy-terminal CesT-binding region that is sufficient for interaction with CesT . Furthermore , both of the Tir CesT-binding regions have a conserved sequence motif distinct from the known chaperone binding β-motif . Previous studies on the Salmonella effectors SipA and SptP showed that disruption or deletion of the β-motif prevented chaperone binding and subsequent effector secretion through the T3SS [31 , 40] . To probe the function of the N- and C-terminal CesT-binding regions in Tir , we constructed leucine to glutamate mutants , L49E and L514E , within the N- and C-terminal β-motifs ( cyan , Fig 4C ) . Since we showed that Tir lacking the N-terminal CesT-binding region ( Tir81-550 ) could still interact with CesT , we first tested if β-motif variants of the individual N- and C-terminal Tir-peptide fragments retained CesT binding . Despite CesT being abundant in the soluble lysate , the Tir23-80 L49E and Tir490-550 L514E peptide variants were unable to pull-down CesT indicating that disruption of either β-motif prevented capture by CesT ( Fig 5A ) . Consistent with this , T18 fusions of the Tir23-80 L49E and Tir490-550 L514E β-motif variants also showed little to no interaction in the presence of CesT and CesT138 fused with T25 in BACTH assays ( Fig 5B , white colonies ) . However , the wild type T18 fusions of Tir23-80 and Tir490-550 constructs interacted with CesT and CesT138 fused to T25 ( Fig 5B , blue colonies ) . As a control , we tested if the individual Tir23-550 L49E and Tir23-550 L514E variants were stable and could pull-down CesT , as both individual variants still have an intact CesT-binding region . Tir23-550 L49E and Tir23-550 L514E were both able to pull-down CesT suggesting that the mutations did not alter the global stability and structure of Tir , and that the L49E and L514E variants only locally disrupt CesT binding to Tir at the N- and C-terminal CesT-binding regions , respectively ( Fig 5C ) . However , the Tir23-550 L514E variant showed lower levels of CesT pull-down compared to wild-type and L49E Tir23-550 . We also tested the Tir23-550 L49E L514E double variant , as we predicted that disrupting both the N- and C-terminal CesT-binding regions would disrupt interaction with CesT . Indeed , the Tir23-550 L49E L514E double variant was significantly impaired for the ability to pull-down CesT ( Fig 5C ) . As a functional readout , we tested whether the individual CesT-binding regions were required for Tir secretion by complementing a Δtir mutant with tir L49E , tir L514E , or the tir L49E L514E double mutant . As a positive control , Tir secretion was restored to wild type levels in the Δtir strain complemented with tir under its native LEE5 promoter ( Fig 5D and S4A Fig ) . The Tir-L49E variant , which retained the functional C-terminal CesT-binding region , also restored Tir secretion to wild-type levels ( Fig 5D and S4 Fig ) . However , the Tir-L514E variant that retains the N-terminal CesT-binding region , but disrupts the C-terminal CesT-binding region , drastically reduced Tir secretion to levels similar to the Tir-L49E L514E double variant ( Fig 5D and S4 Fig ) . To test whether these same phenotypes were observed for the biologically relevant process of effector translocation , we analyzed infected HeLa cells for effector translocation using these same strains . Similar to the secretion assays , complementation of the Δtir strain with tir or tir L49E restored Tir secretion and resulted in the host-cell modification of tyrosine phosphorylation as observed by the upper band in Tir immunoblots ( Fig 5E ) . The Δtir strain complemented with tir L514E or tir L49E L514E significantly reduced host-cell translocation ( Fig 5E ) . Considering that host tyrosine phosphorylation of Tir is important for actin polymerization within the infected cell , we analyzed infected HeLa cells for pedestal formation . Consistent with our phenotypic data , the Δtir strain complemented with empty plasmid was impaired for pedestal formation , whereas the Tir and Tir L49E-complemented strains formed thick actin pedestals underneath microcolonies of bacteria in >90% of all analyzed cells ( Fig 5F and 5G ) . Bacterial cells expressing Tir L514E or Tir L49E L514E were significantly impaired for the formation of actin pedestals ( Fig 5F and 5G ) , which is in agreement with the host translocation results . As a secondary proxy for effector secretion and translocation , we also probed the levels of NleA , as it is highly dependent on free CesT available in the cell . We observed reduced levels of endogenous NleA secretion and host translocation in the Δtir strain complemented with plasmids carrying the tir wt , L49E , and L514E variants ( Fig 5D and 5E ) , likely due to the higher cellular levels of Tir from trans-complementation that would reduce the levels of free CesT . However , the tir L49E L514E double mutant displayed higher levels of NleA secretion and translocation ( Fig 5D and 5E ) , correlating with reduced binding of the Tir L49E L514E mutant to CesT ( Fig 5C ) . Together these data suggest that the C-terminal CesT-binding region of Tir is required for efficient Tir secretion , host translocation , and the formation of actin pedestals . Complementation studies showed that perturbation of CesT binding to the C-terminal region of Tir impaired Tir secretion and host translocation . Since the second CesT-binding region appears to be unique to Tir among all other CesT cargo , we tested the effect of N- and C-terminal domain truncations of Tir for secretion and translocation efficiency . Truncations of tir were constructed on the chromosome to produce Tir fragments encompassing residues 1–391 ( tir NT ) and 320–550 ( tir CT ) ( Fig 6A ) . Tir NT contains the two elements predicted to be required for secretion , a T3SS signal sequence and a CesT-binding region . On the other hand Tir CT lacks a type III secretion signal sequence but contains the novel CesT-binding region in the C-terminal domain . Secretion assays conducted with the tir NT and tir CT-expressing strains revealed that , although Tir NT was present at similar levels as full-length Tir in the bacterial cytosol , Tir NT was secreted at a very low level compared to full-length Tir ( Fig 6B ) . This strain also displayed similar levels of NleA secretion but drastically reduced cytosolic levels of NleA compared to wild type ( Fig 6B ) . These results are consistent with the Tir L514E variant that displayed impaired in vitro secretion ( Fig 5D ) . The tir CT strain did not secrete Tir CT as expected , since it lacks a type 3 secretion signal , but it was present in the whole cell lysate albeit at a very low level compared to wild type Tir . Interestingly the tir CT strain displayed a marked increase in cytosolic and secreted NleA compared to wild type or tir NT strains ( Fig 6B and S4B Fig ) . Next , we tested whether the tir mutant strains were functional for effector translocation into HeLa cells , and the formation of actin pedestals . Similar to the in vitro assays , the tir NT strain translocated lower levels of Tir NT compared to wild type EPEC , but had similar levels of NleA ( Fig 6C ) . In contrast , the tir CT strain was unable to translocate Tir CT as expected , but now had low levels of NleA secretion similar to the Δtir strain ( Fig 6C ) . Other than wild type , none of the strains tested were able to form highly polymerized actin pedestals ( Fig 6D and 6E ) . Taken together , these data indicate that the Tir C-terminal domain is important for Tir secretion and translocation into host cells , and that cellular levels of Tir in the bacterial cytosol affect the production and secretion of NleA . Since the TGRLISS sequence in Tir was highly conserved between the N- and C-terminal CesT-binding regions ( Fig 4C ) , we conducted a sequence search of this motif in other CesT cargo . We found a similar sequence , which we termed the CesT-extension motif , at the N-terminus of NleH1 , NleH2 , and EspZ ( Fig 7A ) . The CesT-extension motif was much less conserved or not apparent in other CesT-dependent effectors ( Fig 7A ) , suggesting that only a subset of CesT cargo contain this motif . We constructed disruptive glutamate variants in either the isoleucine or leucine residue present in the CesT-extension motif ( residues starred purple in Fig 7A ) . This conserved isoleucine or leucine residue was chosen because , based on our structure , I38 in Tir32-80 extends into the hydrophobic CesT-binding pocket ( purple , Fig 2D ) and mutation to glutamate would disrupt the formation of β1’ . Tir was tested first , showing that the individual Tir23-550 I38E and Tir23-550 I500E variants were stable and could pull-down CesT similarly to wild type Tir ( Fig 7B ) . The Tir23-550 I38E I500E double variant was also able to pull-down CesT , albeit at slightly lower levels , suggesting that the CesT-extension motif is not obligatory for CesT binding ( Fig 7B ) . To determine if the individual CesT-extension motifs were required for Tir secretion , the Δtir strain was complemented with either tir I38E , tir I500E , or tir I38E I500E and Tir secretion was tested in secretion assays . Tir secretion was restored to wild type levels in the Δtir strain complemented with tir under its native promoter and with the Tir-I38E and Tir-I500E single substitution variants ( Fig 7C and S4C Fig ) . However , the Tir I38E I500E double variant had a slight reduction in Tir secretion compared to the wild type-complemented strain in these secretion assays ( Fig 7C and S4C Fig ) . This is consistent with the reduced levels of CesT observed in the pull-down assays using the Tir I38E I500E double variant , but does not excluded the possibility that other factors contribute to the reduce secretion of the Tir I38E I500E variant . To further our in vitro secretion studies , we also analyzed infected HeLa cells for effector translocation and actin polymerization with the same strains . Complementation of the Δtir strain with tir , tir I38E , and tir I500E , restored Tir translocation and tyrosine phosphorylation ( Fig 7D ) , albeit Tir-I38E had slightly lower Tir levels . The Δtir strain complemented with tir I38E I500E had the lowest levels of translocation and host modification ( Fig 7D ) . Consistent with these results , we observed a significant reduction in actin pedestal formation in the strain expressing the Tir I38E I500E double variant ( Fig 7E and 7F ) . We also probed the levels of NleA and similar to the Tir β-motif variants , we observed reduced levels of NleA secretion and translocation in the tir complemented strains ( Fig 7C and 7D ) . However , the tir I38E I500E double mutant did not show higher levels of NleA secretion and translocation like the tir L49E L514E strain ( Fig 7C and 7D ) , correlating with the pull-down data showing the Tir I38E I500E mutant retains CesT interaction ( Fig 7B ) . To expand these observations with other effectors , we made the equivalent glutamate substitutions in NleH1 ( L28E ) , NleH2 ( L28E ) , and EspZ ( L45E ) . We also included NleA ( V44E ) in this analysis because it is a highly translocated effector but contains a very divergent sequence to the CesT-extension motif ( Fig 7A ) . We tested if His6-tagged effectors and their putative CesT-extension motif variants were stable and could pull-down CesT . NleH1 , NleH2 , EspZ , and their respective glutamate variants were all able to pull-down CesT , however reduced CesT pull-down was observed for NleH2 L28E and EspZ L45E ( Fig 8A ) . Interestingly , we observed little to no CesT in the NleA pull-downs , suggesting that CesT may not act as a chaperone for NleA but is only required to antagonize CsrA repression of the nleA 5’-UTR [27] . To determine if the CesT-extension motif affects secretion of these effectors ( excluding NleA ) , secretion assays were conducted using EPEC carrying a plasmid expressing FLAG-tagged versions of NleH1 , NleH2 , EspZ , and their glutamate variants ( Fig 8B ) . NleH1 L28E had a slight reduction in secretion and NleH2 L28E had little to no reduction in secretion compared to wild type . EspZ L45E was not detected in either the supernatant or whole cell lysate suggesting the mutation affected overall effector stability , possibly due to reduced CesT binding in the cytosol , which would be consistent with the pull-down data . Finally , we tested if the glutamate mutations in the CesT-extension motif of each effector affected translocation into infected HeLa cells . Under infection conditions , there was reduced translocation of NleH1 L28E , NleH2 L28E , and EspZ L45E compared to the wild type effectors ( Fig 8C ) . Together these data suggest that the presence of a CesT-extension motif , in addition to the canonical β-motif , contribute to cargo recognition by CesT in a subset of effectors .
Tir drives the committal step of intimate attachment between EPEC and the host cell through an extracellular interaction with intimin on the bacterial cell surface [41 , 42] . Thus , despite at least 12 effectors having full or partial dependence on CesT for translocation into host cells , Tir is the first effector to be released [22 , 23] . This ultimately leads to attaching and effacing lesions from Tir-dependent signaling cascades that cause actin polymerization at the site of attachment , followed by Tir-independent effects on the host cell resulting from the release of secondary effectors [43] . Despite almost two decades of work on Tir and CesT , the mechanism that discriminates Tir secretion over that of other effectors remains unclear . Early transcriptional activation of tir is one possible mechanism to ensure Tir is available first for secretion . However , this does not seem to be the driving mechanism because LEE5 ( which contains tir , cesT , and eae ) is activated concurrently with LEE2 , LEE3 , LEE4 , and LEE7 approximately 70 minutes after exposure to T3SS inducing conditions [44] . Furthermore , Tir secretion occurs approximately 30 minutes after transcriptional activation [44] , suggesting a post-translational mechanism might drive preferential Tir secretion , as a number of other LEE-encoded effectors would also be present in the cytosol that would require discrimination within the cell . The nature in which Tir interacts with CesT is a possible mechanism by which this discrimination occurs . In this work we identified a second CesT-binding region in the C-terminal domain of Tir , and identify a CesT-extension motif , distinct from the known chaperone-binding β-motif , that is present in the CesT-binding regions of Tir and other highly translocated effectors . Our data raise the possibility that the presence of these features contribute to cytosolic discrimination by CesT , however formal assessment of this hypothesis in the context of the bacterial cell remains to be tested . CesT has been the focus of structural and functional studies since its initial discovery as a Tir-specific chaperone [18] . Following structural determination of EHEC CesT [30] , the domain swapped dimer has been a topic of debate as to whether it represents a biologically relevant conformation [45] . Our structural data on Tir32-80-CesT138 , and work from others on the CsrA-CesT complex [28] , indicates that CesT adopts the same dimer conformation even though it binds different substrates in separate locations . These data , along with solution state structural data from NMR [45] , provide evidence that the domain swapped EHEC CesT dimer is most likely an artifact of crystallization , possibly arising from plasticity in the effector-binding region . Interestingly , the structure of the SrcA chaperone from Salmonella also exhibited plasticity in the effector-binding region [15] , which suggests this could be a conserved property among multi-cargo chaperones to accommodate binding of multiple effectors . Previous studies predicted that Tir residues 39–83 contained a degenerate CesT-binding domain [21] . Our Tir-peptide pull-down and structural data show that the minimal CesT-binding region of Tir is residues 35–75 . In addition to providing the first structural view of an effector bound to CesT , the structure of the Tir32-80-CesT138 complex was critical in identifying the second CesT-binding region within the carboxy-terminus of Tir ( residues 490–550 ) . This chaperone binding arrangement appears unique to Tir , as there is no evidence for a secondary CesT-binding region in other CesT effector cargo . For example , the first 101 residues of Map were sufficient to interact with CesT , whereas the C-terminal 103 residues showed no interaction [20] . Furthermore , the structural data lead to the identification of a putative CesT-extension motif present in the N-terminal region of NleH1 , NleH2 , EspZ , and in both the amino- and carboxy-terminal regions of Tir . This region was so named because , according to our structure , it appears to extend the β-sheet core of CesT in its cargo-laden state . Mutation of the CesT-extension motif showed that it was most important for efficient host translocation . A commonality among effectors containing the CesT-extension motif ( Tir , NleH1 , NleH2 , and EspZ ) is that they are among the most highly translocated effectors among CesT cargo [23] . The presence of this motif in a subset of effectors ( and its duplication in the case of Tir ) may help to understand how multi-cargo chaperones recognize and possibly discriminate between effectors in the cell . Efficient Tir-CesT complex formation is likely driven by the presence of two CesT-binding regions in Tir , however additional factors likely contribute to the observed preference for initial Tir secretion [21–23] . One possible contributing factor could be posttranslational modification of the C-terminus of CesT , which contains a site for tyrosine phosphorylation ( Y152/153 ) [25] . A recent study showed that EPEC expressing a CesT Y153F variant exhibited a global increase in effector secretion , whereas EPEC expressing CesT Y152F was attenuated for NleA translocation into HeLa cells [26] . The latter result could be explained by the recent structure of CsrA in complex with CesT , which shows that CesT Y152 forms critical hydrogen bonds along the CsrA binding interface [28] . Therefore , CesT Y152F likely has reduced binding for CsrA and in turn is unable to depress the nleA 5’-UTR . We also observed that NleA had little to no binding of CesT in our pull-down assays . This suggests that the requirement of CesT for NleA secretion dynamics may be indirect and relate more to CsrA antagonism of the nleA 5’-UTR . This would be consistent with other work showing that NleA is only partially dependent on CesT for translocation into host cells [35] . Alternatively , it is possible that NleA or CesT may need post-translational modifications to facilitate interaction , require a third unknown co-chaperone , or NleA follows a chaperone-independent secretion pathway recently reported for a subset of Shigella T3SS effectors [46] . In addition to driving the formation of the Tir-CesT complex , the second CesT-binding region in Tir might stabilize a distinct conformation of the C-terminal domain of Tir that may be required for efficient targeting of the Tir-CesT complex to the T3SS sorting platform . Our data with the CesT-binding region mutants within the C-terminal domain of Tir ( L514E variant ) , along with the chromosomal tir NT mutant ( C-terminal domain truncation ) support this possibility . For example , the Tir L514E and Tir NT constructs both contain a functional CesT-binding region and a type III secretion signal , but show significantly reduced secretion and translocation efficiency into HeLa cells . Furthermore , deletion of residues 519–524 in the C-terminal domain of EHEC Tir also showed significantly reduced Tir secretion [47] . Interestingly , EHEC Tir residues 519–524 align with EPEC Tir residues 511–516 , which overlap with the predicted β-motif in the second CesT-binding region ( ie . the Tir L514E mutant we tested ) . Recently , it was shown that an affinity switch controls substrate secretion hierarchy in the T3SS of EPEC . The SepL-SepD complex engages EscV ( translocase ) to ensure efficient targeting and secretion of the translocators , while simultaneously inhibiting effector targeting [36] . SepD release from the complex disrupts SepL-EscV crosstalk , leading to equivalent targeting of translocators and effectors for secretion . This is followed by the eventual release of SepL that results in inhibition of translocators and exclusive targeting of late effectors . This study also showed that the Tir-CesT complex had a two-fold increase in affinity for wild-type inner membrane vesicles that contain SepD , SepL , and EscV , over CesT alone [36] . This increased affinity for the Tir-CesT complex is probably due to a SepL-Tir interaction , which is supported by previous pull-down data in EHEC where the C-terminal 48 residues of SepL interact with Tir [37] . This particular Tir interaction site on SepL in EHEC also overlaps with one of the two EscV binding patches observed on EPEC SepL from peptide-binding array data [36] . Considering these studies with our findings , we provide an extended version of the affinity-switch model proposed by Portaliou et al . that includes differential secretion of late effectors ( Fig 9 ) . In this model , the SepD-SepL ( and likely CesL ) complex interacts with EscV and allows for strict translocator secretion ( EspA , EspB , EspD ) while simultaneously preventing effector secretion . It is also noteworthy that EscP binds SepL in a calcium-dependent manner and also contributes to the blocking of late effectors for secretion [48] . At this point it is plausible that CesT dimers are predominantly loaded with Tir in the cytosol since Tir contains two CesT-binding regions . After SepD dissociation from the EscV membrane complex , the Tir-CesT complex might compete with SepL for EscV binding . This competition could be mediated by the C-terminal domain of Tir leading to strict docking of the Tir-CesT complex to EscV over other effectors . Alternatively , the Tir-CesT complex may compete for EscP binding to SepL , leading to early docking of the complex within the sorting platform . Either possibility may explain preferential Tir release over other effectors , but it remains to be shown experimentally . SepL eventually dissociates from EscV , a process that may be directly influenced by Tir-SepL interactions , leading to the inhibition of translocator secretion and strict targeting of late effectors . Rapid release of Tir would result in the accumulation of free CesT in the cell , which in turn can antagonize CsrA repression of the nleA transcript through CesT-CsrA interactions , and increase binding of other highly translocated effectors such as EspZ , NleH1 , and NleH2 . Further depletion of the effector pool liberates more CesT , allowing for binding and secretion of lower translocated effectors .
Bacterial strains , plasmids , and primers used in this study are described in Tables A and B in SI Text . Phusion or Phire II polymerase ( Thermo Fisher Scientific ) were used for all PCR reactions , oligonucleotide primers were synthesized by Sigma , and site-directed mutants were constructed by using the Q5 site-directed mutagenesis kit ( NEB ) with the mutation encoded in the amplification primer . For protein expression , CesT and the C-terminal truncation construct encoding residues 2–138 ( CesT138 ) were cloned into pET28a using the NheI/XhoI restriction sites . To allow for subsequent sub-cloning of CesT the internal NdeI site was removed by introducing a silent mutation in the coding region of His138 ( CAT::CAC ) . CesT and CesT138 were then sub-cloned into the second multiple cloning site of pCOLADuet-1 using the NdeI/XhoI restriction sites . CesT-FLAG and CesT138-FLAG were also cloned into pCOLADuet-1 using the NdeI/XhoI site with the FLAG-tag encoded within the PCR amplification primer . Tir and EspZ; NleH1 and NleH2; and NleA constructs were cloned using the BamHI/HindIII , BamHI/SalI , and BamHI/NotI sites , respectively , into the first multiple cloning site of pCOLADuet-1 with CesT or CesT138 ( and/or their FLAG-tag versions ) in the second multiple cloning site for co-expression and pull-down experiments . For T3SS complementation assays , the LEE5 promoter encompassing nucleotides -7 to -323 from the translational start site of tir was cloned into pWSK29 using the XhoI/HindIII restriction sites . Subsequently , tir was cloned into the pWSK29-Ptir plasmid using the HindIII/NotI restriction sites . For effector secretion studies , effectors carrying a C-terminal FLAG-tag encoded in the primer were cloned into pGEN-luxCDABE using the SnaBI/SacI sites . For BACTH experiments , CesT and Tir constructs were cloned into the pKNT25 and/or pUT18C plasmids using the XbaI/SacI sites . All plasmids were verified by sequencing . Protein fractions were separated by SDS-PAGE and transferred to a polyvinylidene fluoride membrane , and blocked in Tris-buffered saline with 0 . 1% ( w/v ) Tween ( TBST ) containing 5% skim milk . Membranes were then probed using the following primary antibodies: mouse monoclonal anti-Tir ( 1:2000 ) for full-length and C-terminal fragments , rat polyclonal anti-Tir from C . rodentium ( 1:2000 ) [49] for N-terminal fragments , rat polyclonal anti-NleA from EHEC ( 1:2000 ) [50] , mouse α-DnaK ( Stressgen , 1:5000 ) , mouse α-FLAG M2 ( Sigma , 1:5000 ) , mouse α-His6 ( GE Healthcare , 1:3000 ) , or goat α-GAPDH ( R&D Systems Inc . , 1:5000 ) . The blots were then developed using the following secondary antibodies: goat anti-mouse ( 1:5000 , Jackson ) , goat anti-rat ( 1:2000 , EMD Millipore ) , or donkey α-goat ( Santa Cruz Biotechnology , 1:5000 ) conjugated to horseradish peroxidase , and imaged using the Clarity Western ECL ( BioRad ) or SuperSignal West Femto Maximum Sensitivity ( ThermoFisher ) substrates and a ChemiDoc XRS+ ( BioRad ) . E . coli BL21-CodonPlus ( DE3 ) cells were transformed with the appropriate co-expression plasmid ( pCOLADuet-1 containing N-terminal His6-tagged effector and C-terminal FLAG-tagged CesT ) , grown overnight in LB media with 50 μg/mL kanamycin at 37°C with shaking , sub-cultured 1:50 in 50 mL LB media with 50 μg/mL kanamycin to an OD600 of ~0 . 4–0 . 5 , and moved to 30°C . When the cultures reached OD600 of ~0 . 6–0 . 7 protein expression was induced by the addition of isopropyl-D-1-thiogalactopyranoside ( IPTG ) to a final concentration of 0 . 5 mM . The cells were incubated for 3 h at 30°C , harvested by centrifugation at 5000 g for 10 min , and frozen on dry ice . Cell pellets were thawed and re-suspended in 2 mL of lysis buffer ( 50 mM Tris-HCl pH 7 . 5 , 300 mM NaCl , 20 mM imidazole , 5% ( v/v ) glycerol , 2 mM 2-mercaptoethanol ) . Re-suspended cells were lysed by sonication and cell debris was removed by centrifugation at 16000 g for 30 min . The resulting supernatant was passed over a gravity column containing 0 . 2 mL Ni-nitrilotriacetic acid ( NTA ) agarose resin ( Qiagen ) that was pre-equilibrated with lysis buffer . Bound protein was washed with 100 column volumes of lysis buffer , and eluted with 5 column volumes of lysis buffer with 250 mM imidazole . Soluble lysate and elution fraction samples were mixed with equal parts of 2X SDS-PAGE loading dye and analyzed by SDS-PAGE and western blotting . The following protocol was used to express and purify CesT , CesT138 , and all the Tir-CesT complexes for crystallization . E . coli BL21-CodonPlus ( DE3 ) cells were transformed with the appropriate plasmid , grown overnight in LB media with 50 μg/mL kanamycin at 37°C with shaking , sub-cultured 1:50 into 1 L LB media with 50 μg/mL kanamycin to an OD600 of ~0 . 4–0 . 5 , and moved to 18°C . When the cultures reached OD600 of ~0 . 6–0 . 7 protein expression was induced by the addition of IPTG to a final concentration of 0 . 5 mM . The cells were incubated overnight at 18°C , harvested by centrifugation at 5000 g for 10 min , and frozen on dry ice . Cell pellets were thawed and re-suspended in 25 mL of lysis buffer ( 50 mM Tris-HCl pH 7 . 5 , 300 mM NaCl , 10 mM imidazole , 5% ( v/v ) glycerol , 2 mM 2-mercaptoethanol , and one complete mini protease inhibitor cocktail tablet ( Roche ) ) . Re-suspended cells were lysed by sonication and cell debris was removed by centrifugation at 31000 g for 30 min . The resulting supernatant was passed over a gravity column containing 3 mL Ni-NTA agarose resin ( Qiagen ) that was pre-equilibrated with lysis buffer . Bound protein was washed with 10 column volumes of lysis buffer , 3 column volumes lysis buffer with 20 mM imidazole , and eluted with 5 column volumes of lysis buffer with 250 mM imidazole . The eluted fraction was concentrated using a 10 or 30 kDa cut-off Amicon ultrafiltration device ( EMD Millipore ) and further purified and buffer exchanged into 20 mM Tris-HCl pH 7 . 5 and 150 mM NaCl by size exclusion chromatography using a HiLoad 16/60 Superdex 200 prep-grade gel-filtration column ( GE Healthcare ) . The purified constructs were >95% pure as judged by SDS-PAGE and stable for at least 1 week at 4°C . Purified His6-Tir32-80-CesT138 was concentrated to ~14 mg/mL and screened for crystallization conditions at 22°C using hanging-drop vapour diffusion in 24-well VDXm plates ( Hampton Research ) and the MCSG 1–4 sparse matrix suites ( Anatrace ) . The best initial crystallization hits were obtained from MCSG-1 condition #17 and MCSG-3 condition #44 . Optimized crystals were grown by mixing 2 μL of 14 mg/mL His6-Tir32-80-CesT2-138 with 1 . 5 μL of precipitant solution ( 0 . 1 M Tris-HCl pH 7 . 5 , 0 . 2 M MgCl2 , and 17% ( w/v ) PEG3350 ) equilibrated against 500 μL of 1 . 7 M MgSO4 . The crystals took 1–3 weeks to reach maximum size and were frozen without cryoprotection in liquid nitrogen . Diffraction data were collected at a wavelength of 0 . 98 Å on beam line 08B1-1 at the Canadian Light Source ( CLS ) ( Table 1 ) . The data were indexed and integrated with iMosflm [51] and scaled using SCALA in the CCP4i suite [52] . The structure was determined by molecular replacement with Phenix Phaser [53] using EHEC CesT residues 38–131 ( PDB ID: 1K3E ) as a search model . The resulting electron density map enabled Phenix AutoBuild [54] to build ~70% of CesT138 . The remaining CesT138 residues and the Tir32-80 fragment were built manually in Coot [55] and alternated with refinement using phenix . refine [56] . Translation/Libration/Screw ( TLS ) groups were used during refinement and determined automatically using the TLSMD web server [57 , 58] . Structure figures were generated using PYMOL Molecular Graphics System ( DeLano Scientific ) , and quantitative electrostatics were calculated using PDB2PQR [59 , 60] and APBS [61] . Elution fractions from various Tir-CesT co-expression pull-down experiments were concentrated to 100 μL , applied to a Superdex 200 10/300 GL column , and eluted using 20 mM Tris-HCl pH 7 . 5 and 150 mM NaCl . Protein standards used to calibrate the column were ferritin ( 440 kDa ) , conalbumin ( 75 kDa ) , ovalbumin ( 44 kDa ) , ribonuclease A ( 13 . 7 kDa ) , and aprotinin ( 6 . 5 kDa ) . E . coli BTH101 cells were co-transformed with the various pKNT25 and pUT18C based plasmids and recovered on LB-Kan50-Amp100 agar plates at 37°C . Single colonies were grown overnight with shaking at 37°C in LB-Kan50-Amp100 , and then 20 μL of each sample was spotted onto LB agar plates containing Kan50 , Amp100 , 0 . 5 mM IPTG , and 40 μg/mL 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside ( X-gal ) . Plates were incubated for 24–48 h at 30°C for the development of blue colonies . Primer pairs with 48 nucleotide homologous tails to escN or tir were used to amplify linear PCR products with pKD3 for generation of the escN or various tir mutants . In-frame marked mutants of EPEC replacing escN residues 9–446 or tir residues 50–319 , 392–535 , and 17–535 with chloramphenicol acetyltransferase ( cat ) were constructed using one-step λ-red inactivation with pKD46 and the transformed linear PCR products [62] . The cat cassette was then removed using plasmid pFLP2 and sucrose selection . All tir and escN deletions were verified by sequencing . Secretion assays were performed similar to those described previously [63] . Standing overnight EPEC cultures grown in LB media ( plus 100 μg/mL Amp as needed ) at 37°C were sub-cultured 1:40 into 4 mL of pre-warmed Dulbecco’s modified eagle medium ( DMEM ) plus 2 mM ethylene glycol-bis ( β-aminoethylether ) -N , N , N' , N'-tetraacetic acid ( EGTA ) in glass tubes . The cultures were incubated standing for 6 h at 37°C in a 5% CO2 incubator ( OD600 of 0 . 7–0 . 9 ) . The cultures were then harvested by centrifugation at 10000 g for 5 min , and the bacterial pellets were washed once in phosphate-buffered saline ( PBS ) and re-suspended in 1X SDS-PAGE loading dye ( normalized by OD600 as necessary ) . The culture supernatant was passed through a low-protein binding 0 . 2 μm filter ( Pall ) , and 1 . 35 mL aliquots were mixed with 150 μL of ice-cold 100% ( w/v ) trichloroacetic acid and incubated overnight at 4 °C . The solutions were centrifuged at 16000 g for 30 min , the supernatant was discarded , and the pellet was washed with 1 mL of ice-cold acetone . The washed pellets were centrifuged at 16000 g for 30 min , the pellet was air dried , and then re-suspended in 10 μL 1X SDS-PAGE loading buffer ( or normalized by OD600 as necessary ) . Samples were then analyzed by SDS-PAGE using coomassie blue G250 stain or by western blotting . HeLa cells ( Coombes lab collection ) were grown in DMEM + 10% fetal bovine serum at 37°C in a 5% CO2 incubator . Cells were routinely grown in 75 mm2 dishes ( VWR ) until confluent , and were then seeded at 2 . 2×106 into 100 mm dishes ( Corning ) and incubated overnight . Prior to infection the HeLa cells in 100 mm dishes were washed with 10 ml of warm PBS . EPEC cultures were grown overnight standing at 37°C , harvested by centrifugation , and resuspended in DMEM . HeLa cells were then infected with EPEC at a multiplicity of infection of 50:1 for 3 h at 37°C in a 5% CO2 incubator . The cells were washed five times with cold PBS , harvested with a cell scraper , centrifuged at 1000 g for 5 min , and resuspended in 250 μL PBS + 0 . 5% ( v/v ) Triton X-100 . Cells were then lysed for 30 min on ice with gentle rocking , centrifuged at 10000 g for 5 min , and the following supernatant was mixed with equal parts of 2X SDS-PAGE loading buffer for SDS-PAGE and western blot analysis . HeLa cells maintained in DMEM + 10% fetal bovine serum were seeded at 1×105 into 24 well tissue culture plates ( VWR ) containing 12 mm circle micro coverglass slips ( VWR ) and incubated overnight at 37°C in a 5% CO2 incubator . Prior to EPEC infection the glass slips were washed with 1 ml of warm PBS . EPEC cultures carrying a GFP expression plasmid for visualization ( pFPV25 . 1 for the tir chromosomal domain mutants and pACYC-GFP for the tir complementation strains with various Tir point mutants ) were grown standing overnight in LB media at 37°C , harvested by centrifugation , and resuspended in DMEM . HeLa cells were then infected with EPEC at a multiplicity of infection of 50:1 for 3 h at 37°C in a 5% CO2 incubator . Infected cells were washed with PBS ( and after each subsequent step ) , fixed with 4% paraformaldehyde in PBS for 15 min , permeabilized with 0 . 3% ( v/v ) Triton X-100 in PBS for 5 min , and blocked with 5% ( w/v ) bovine serum albumin ( BSA ) in PBS for 30 min . F-actin was then stained using Alexa Fluor 568 Phalloidin ( 1:500 , ThermoFisher ) in PBS containing 1% ( w/v ) BSA for 60 min . Stained coverslips were washed in PBS and mounted on glass slides using ProLong Gold Antifade Mountant with 4’ , 6-Diamidino-2-phenylindole dihydrochloride ( DAPI ) for nuclear staining ( Life Technologies ) and allowed to sit overnight before sealing with nail polish . Microscopy was performed using a ZEISS Axio Imager 2 with 40X and 100X oil-immersion lenses with laser excitation . Images were captured using a Hamamatsu ORCA-R2 digital CCD camera and exported TIFF files were processed into their individual and composite color channels using ImageJ2 [64] . Quantification of pedestal formation ( binding index ) was conducted as described previously [65] , where the percentage of infected HeLa cells that contained a microcolony of at least five GFP-positive bacteria associated with F-actin condensation were enumerated ( co-localization , yellow ) .
|
Enteropathogenic Escherichia coli injects effector proteins into host cells using a type III secretion system ( T3SS ) . The translocated intimin receptor ( Tir ) is the first effector delivered into host cells and imparts efficient secretion of other effectors . However , the mechanism for Tir-dependent modulation of the T3SS is poorly understood . We provide evidence that the multi-cargo chaperone CesT binds to two regions in Tir at the N- and C-terminus through a specific recognition motif , and show that CesT binding to the Tir C-terminus is important for host translocation . Furthermore we show that the CesT-specific motif is conserved in a subset of highly translocated effectors . This study highlights the multi-faceted role that T3SS chaperones play in effector secretion dynamics .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Material",
"and",
"methods"
] |
[
"bacteriology",
"dimers",
"(chemical",
"physics)",
"complement",
"system",
"medicine",
"and",
"health",
"sciences",
"immune",
"physiology",
"pathology",
"and",
"laboratory",
"medicine",
"viral",
"transmission",
"and",
"infection",
"hela",
"cells",
"molecular",
"probe",
"techniques",
"pathogens",
"biological",
"cultures",
"immunology",
"microbiology",
"immunoblotting",
"physiological",
"processes",
"secretion",
"systems",
"cell",
"cultures",
"sequence",
"motif",
"analysis",
"molecular",
"biology",
"techniques",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"immune",
"system",
"proteins",
"microbial",
"physiology",
"bioinformatics",
"proteins",
"chemistry",
"cell",
"lines",
"molecular",
"biology",
"immune",
"system",
"physics",
"biochemistry",
"bacterial",
"physiology",
"host",
"cells",
"virulence",
"factors",
"virology",
"physiology",
"database",
"and",
"informatics",
"methods",
"secretion",
"biology",
"and",
"life",
"sciences",
"cultured",
"tumor",
"cells",
"physical",
"sciences",
"chemical",
"physics"
] |
2018
|
Molecular basis for CesT recognition of type III secretion effectors in enteropathogenic Escherichia coli
|
Treponema pallidum subsp . endemicum ( TEN ) is the causative agent of endemic syphilis ( bejel ) . An unusual human TEN 11q/j isolate was obtained from a syphilis-like primary genital lesion from a patient that returned to France from Pakistan . The TEN 11q/j isolate was characterized using nested PCR followed by Sanger sequencing and/or direct Illumina sequencing . Altogether , 44 chromosomal regions were analyzed . Overall , the 11q/j isolate clustered with TEN strains Bosnia A and Iraq B as expected from previous TEN classification of the 11q/j isolate . However , the 11q/j sequence in a 505 bp-long region at the TP0488 locus was similar to Treponema pallidum subsp . pallidum ( TPA ) strains , but not to TEN Bosnia A and Iraq B sequences , suggesting a recombination event at this locus . Similarly , the 11q/j sequence in a 613 bp-long region at the TP0548 locus was similar to Treponema pallidum subsp . pertenue ( TPE ) strains , but not to TEN sequences . A detailed analysis of two recombinant loci found in the 11q/j clinical isolate revealed that the recombination event occurred just once , in the TP0488 , with the donor sequence originating from a TPA strain . Since TEN Bosnia A and Iraq B were found to contain TPA-like sequences at the TP0548 locus , the recombination at TP0548 took place in a treponeme that was an ancestor to both TEN Bosnia A and Iraq B . The sequence of 11q/j isolate in TP0548 represents an ancestral TEN sequence that is similar to yaws-causing treponemes . In addition to the importance of the 11q/j isolate for reconstruction of the TEN phylogeny , this case emphasizes the possible role of TEN strains in development of syphilis-like lesions .
Treponema pallidum subsp . endemicum ( TEN ) is the causative agent of bejel ( endemic syphilis ) , a chronic human infection usually affecting children under 15 years of age . The primary stage of endemic syphilis is often localized to the mucosa of the oral cavity or nasopharynx and frequently remains undetected . Secondary lesions often mimic syphilitic lesions and are found on both mucosal and skin surfaces including the oral cavity , pharynx , and larynx ( for review see [1] ) . The tertiary stage is characterized by gummatous or destructive lesions of mucosa , skin , and bones . Recently reported cases of bejel have come from African countries with dry climates including Mauretania , Niger , Chad , Mozambique and from countries in the Middle East including Turkey , Saudi Arabia , and Iran [1] . Moreover , several imported cases of bejel have been described in France [2] and Canada [3] in children coming from countries where endemic syphilis has been reported . Compared to the syphilis-causing Treponema pallidum subsp . pallidum ( TPA ) and the yaws-causing Treponema pallidum subsp . pertenue ( TPE ) ( reviewed in [4 , 5] ) , TEN is the least well characterized and least studied human pathogenic treponeme . There are few genetic studies on TEN strains [3 , 6–14] , which is likely due to a limited number of available TEN samples . In fact , most studies on TEN strains described one of the two reference strains , i . e . , Bosnia A or Iraq B . The Bosnia A strain was isolated in 1950 in southern Europe ( Bosnia ) from a 35-year old male with several mucosal and skin lesions [15] , and the Iraq B strain was isolated in 1951 in Iraq from a 7-year old girl who had oral mucous lesions and an anal condylomata [15] . Because of the low number of available reference strains , only a single complete genome sequence of TEN Bosnia A has been published to date showing a close relatedness ( higher than 99 . 9% ) to TPE strains and several sequences surprisingly similar to TPA strains [16] . First reported in 2013 , an unusual 11q/j subtype ( defined by enhanced CDC typing ) [17 , 18] was found among samples taken from a syphilis patient in Paris [19] who had returned from Islamabad , Pakistan , where he admitted having had sex with commercial sex worker . Based on a partial sequence type reported for TP0548 , Mikalová et al . [20] pointed out that this sequence was more related to the yaws-causing strains rather than to syphilis-causing strains . Further analyses resulted in the classification of the 11q/j isolate as a TEN treponeme [21] . In this study , we characterized the 11q/j isolate in a set of 44 independent chromosomal regions . Sequencing of these loci revealed that the 11q/j isolate belongs to the T . pallidum subsp . endemicum with two loci having sequences that were related to either TPE ( TP0548 ) or TPA ( TP0488 ) . The relevance of these findings is discussed here .
The study was approved by the institutional review board of the Comité de Protection des Personnes d’Ile de France 3 ( S . C . 3005 ) . The sample ( a swab from an indurated genital ulceration ) was collected from a 42-year-old heterosexual man who attended the outpatient STD clinic of Hôpital Saint-Louis ( Paris ) and was analyzed anonymously . Isolated DNA ( 20 μl ) from this sample ( referred as 11q/j ) was obtained from the National Reference Center for Syphilis in France ( CNR Syphilis , www . cnr-syphilis . fr ) that had performed a routine analysis on DNA from clinical samples [22] . A nested PCR protocol for detection of the polA gene was performed using the previously described outer primers , polA_outer_F1 ( 5´-TTCTGTGCTCACGTCTGGTC-3´ ) and polA_outer_R1 ( 5´-TGCAACCATCGTATCGAAAA-3´ ) , which resulted in a 637 bp amplicon [23–25] and inner primers for nested polA PCR , polA_F1 ( 5´-TGCGCGTGTGCGAATGGTGTGGTC-3´ ) and polA_R1 ( 5´-CACAGTGCTCAAAAACGCCTGCACG-3´ ) , resulting in a 377 bp amplicon , were used as described in Liu et al . [26] . This nested PCR protocol was shown to be able to detect 1–10 copies of treponemal DNA in a 1 μl of sample [23–25] and was used for detection of the number of treponemal genome equivalents in 1 μl of DNA . The original 11q/j DNA sample ( 3 μl ) was randomly amplified using a REPLI-g Single Cell kit ( Qiagen , Hilden , Germany ) according to the manufacturer's instructions . Randomly amplified sample of 11q/j was then used for ( i ) direct nested PCR amplification with specific primers ( listed in S1 Table ) according to a previously published protocol [25 , 27] , ( ii ) whole DNA sequencing using an Illumina MiSeq Next-gen sequencer ( Illumina , San Diego , CA , USA ) , and ( iii ) the subsequent amplification with T . pallidum specific primers used in the pooled segment genome sequencing ( PSGS ) method [16 , 28–30] , which was followed by Illumina sequencing . Regions successfully amplified from the 11q/j isolate were also amplified from another available DNA reference sample , i . e . , TEN Iraq B . The TEN Iraq B DNA was provided by Dr . Kristin N . Harper from the Department of Population Biology , Ecology , and Evolution , Emory University , Atlanta , Georgia , USA , in 2005 . The Iraq B DNA was amplified with PCR or the nested PCR protocol with the same specific primers used for nested PCR amplification of the 11q/j isolate ( S1 Table ) . The obtained partial sequences from the 11q/j isolate and TEN Iraq B were either assembled from Sanger and/or Illumina sequencing reads using SeqMan or SegMan NGen software ( DNASTAR , Madison , WI , USA ) , respectively , with default assembling parameters . Genes were annotated according to the whole genome sequence of TEN Bosnia A ( CP007548 . 1; [16] ) and the 11q/j isolate and the TEN Iraq B genes were tagged with TEND11qj_ and TENDIB_ prefixes , respectively . The resulting sequences of the 11q/j isolate and TEN Iraq B were analyzed and compared to the following genomes: TPA Nichols ( CP004010 . 2; [31] ) , TPA SS14 ( CP004011 . 1; [31] ) , TPE Samoa D ( CP002374 . 1; [29] ) , TPE CDC-2 ( CP002375 . 1; [29] ) , TPE Gauthier ( CP002376 . 1; [29] ) , TPE Fribourg-Blanc ( CP003902 . 1; [30] ) , and TEN Bosnia A [16] . Alignments of treponemal sequences were performed using SeqMan software and MEGA7 software [32] . Phylogenetic trees were constructed in MEGA7 software [32] using the Maximum Likelihood method based on the Tamura-Nei model [33] . The following formula was used to calculate the probability that the observed nucleotide sequences were caused by accumulation of individual mutations instead of a recombination: pmut = ( pmut_gen x pmut_nuc ) n , where pmut = the end probability of mutations resembling recombinant events , pmut_gen = the frequency of mutation per single nucleotide , pmut_nuc = the probability of a nucleotide substitution into the nucleotide sequence in the putative recombinant region , n = the number of mutated nucleotides within the putative recombinant region . pmut_gen was calculated based on the number of variable sites identified within all available sequences of the 11q/j sample ( total length of 29 , 753 bp , except for loci TP0488 and TP0548 ) and the corresponding sequences of TEN Bosnia A and TEN Iraq B . pmut_nuc had a constant value of 0 . 333 reflecting 3 possible substitutions changing the original sequence at each nucleotide site . Different probabilities of transitions and transversions were not considered in this analysis . The “n” was calculated based on the number of different nucleotide positions between the 11q/j isolate and one of the two TEN strains that matched either the TPA or TPE orthologous sequence . The resulting sequences of the TEN 11q/j isolate and the TEN Iraq B with length ≥ 200 bp were deposited in the GenBank under following accession numbers: KY120774-KY120814 for TEN 11q/j isolate; KY120815-KY120855 for TEN Iraq B . The detailed overview of sequenced loci is shown in S2 Table .
The only available DNA-containing sample ( 20 μl ) was obtained from the CNR Syphilis that performed the isolation of DNA from the original swab sample [22] . As revealed by the nested polA PCR reaction [23] with detection limit of less than 10 molecules [26] , the sample contained undetectable amounts of treponemal DNA , i . e . less than 10 molecules of treponemal DNA per 1 μl . Following whole genome amplification with random primers , nested PCR protocol revealed positivity in a 10−2 dilution indicating , at least , 1x102 copies of treponemal genome equivalents per 1 μl in a total of 50 μl of amplified sample . This randomly amplified sample was used for further analyses . The randomly amplified sample was used for direct Illumina sequencing and resulted in 1 , 786 , 712 individual reads . Of those , only 10 reads were mapped to the TEN Bosnia A genome indicating that the ratio of treponemal DNA to DNA from other species ( mostly human ) is less than 1:105 . Subsequently , the randomly amplified sample was used for specific amplification with the PSGS technique [16 , 28–30] and primer pairs from Pool 1 amplifying the first quarter of the treponemal genome ( Fig 1 ) . Specific amplification resulted in a total of 353 , 006 individual reads , of which 41 , 308 reads were mapped to the TEN Bosnia A genome . Consensus sequences from at least 2 individual reads represented sequenced DNA regions of the 11q/j isolate . All regions determined by Illumina sequencing are shown in Fig 1 and S2 Table . Altogether , 15 genomic loci were obtained for the 11q/j isolate with lengths ranging from 63–2 , 455 bp , with total length of 9 , 626 bp and with coverage ranging from 2–15 , 832x . In addition to Illumina sequencing , a nested PCR of 31 chromosomal loci was performed from the randomly amplified 11q/j sample using 1 μl of the starting DNA template . The resulting amplicons were Sanger sequenced . Loci for nested PCR were selected based on whole genome comparisons of published TPE strains ( Samoa D , CDC-2 , Gauthier ) and TEN Bosnia A . Preferentially , loci with accumulated single nucleotide variants ( SNVs ) and/or indels between TPE and TEN strains were selected as well as conservative genes suitable for unambiguous distinction between TPE and TEN subspecies . All regions amplified using nested PCR and sequenced using the Sanger method are presented in Fig 1 and S2 Table . The 16S and 23S rRNA loci were amplified from both genome positions [12] . The length of resulting sequences of the 11q/j isolate ranged from 352–2302 bp and represented a total of 23 , 979 bp . Illumina and Sanger sequencing of the 11q/j isolate resulted in sequences obtained from 44 chromosomal DNA regions covering , altogether , 32 , 635 bp ( 2 . 87% ) of the TEN Bosnia A genome length ( S2 Table ) . Two genomic regions within TP0121 and TP0136 genes , where both sequencing techniques partially overlapped , revealed identical sequences . The average length of sequenced regions in the 11q/j isolate was 742 bp ( range 63–2 , 455 bp ) . The sequenced chromosomal regions were dispersed throughout the entire chromosome with distances ranging from 0 . 1–124 . 7 kb ( Fig 1 ) . All sequenced genomic regions of the 11q/j isolate were also amplified and Sanger sequenced from the TEN Iraq B DNA and these regions are described in S2 Table . Interestingly , sequencing of a short gene fragment ( 548 bp ) of TENDBA_0488 between positions 684–1231 revealed that the sequence of the 11q/j isolate was very similar to the sequence in TPA Nichols , but not to TEN strains ( Fig 2A ) ; suggesting the occurrence of a recombination event at this locus . The minimal size of recombinant DNA sequence was 505 nucleotides ( between coordinates 715–1219; Fig 2A ) . A set of 21 nucleotide positions of the 11q/j isolate were different from TEN Bosnia A as well at TEN Iraq B , but identical to TPA strains . A partial sequence of the TEND11qj_0488 from the 11q/j isolate , representing the recombinant part ( 505 bp long fragment ) , was used for construction of a tree ( Fig 3A ) that revealed clustering of the 11q/j isolate within TPA strains , not within TEN strains . The probability that the observed nucleotide sequence within this locus was caused by an accumulation of individual mutations instead of a recombination was tested using the following formula: pmut = ( pmut_gen x pmut_nuc ) n ( see Materials and Methods ) . pmut_gen was calculated based on the number of variable sites identified within all available sequences of the 11q/j sample ( 22 variable positions in a total length of 29 , 753 bp from the 3 analyzed TEN genomes; 0 . 00074 nt differences per 1 bp ) . Loci TP0488 and TP0548 were not included in this calculation . The “n” was calculated based on the number of variable positions detected in the sequence alignment presented in Fig 2 . In the TP0488 gene , there was a total of 23 nucleotide positions in the 11q/j sequence that differed from TEN strain Bosnia A but were identical to TPA strain SS14 . With the assumption that the 11q/j isolate represents a TEN strain , the probability that the accumulated SNVs within the TP0488 of the 11q/j isolate were due to accumulation of individual mutations would be: pmut = ( 0 , 00074 x 0 . 333 ) 23 , i . e . pmut = 1 . 01987 x 10−83 . To rule out potential co-infection with TPA and TEN in this patient , Illumina sequencing reads of the 11q/j sample , especially in regions with positions that differ between TPA and TEN , were evaluated and revealed 20 informative sites with coverage ≥ 4x ( range 4x–94x ) . However , there was no heterogeneity in these positions , excluding co-infection with multiple strains . As shown previously , the 11q/j isolate within its 86 bp-long fragment of TP0548 gene revealed a new sequence type that is , in fact , related to TPE strains [20] . Analysis of a larger 2 , 302 bp-long region comprising TP0547 , TP0547a , and TP0548 genes ( positions 589 , 926–592 , 227 corresponding to the whole genome sequence of the TEN Bosnia A ) revealed that the sequence of the 11q/j isolate , within the TP0548 gene , was very similar to TPE strains , especially to a sequence from TPE Samoa D ( Figs 2B and 3B ) . The minimal size of recombinant DNA sequence was 613 nucleotides ( between coordinate 69 of the TENDBA_0547a and coordinate 623 of the TENDBA_0548 ) and comprised 56 variable positions ( Fig 2B ) . Thirty-seven of the nucleotide positions of the 11q/j isolate were different from TEN Bosnia A and TEN Iraq B , but identical to at least one of the TPE Samoa D or Gauthier strains ( Fig 2B ) . Both TEN Bosnia A and Iraq B showed 23 nucleotide positions identical to at least one of the TPA strains , i . e . , to Nichols or SS14 , but different from the 11q/j isolate . A partial sequence of TEND11qj_0548 ( 613 bp ) was used for construction of a tree ( Fig 3B ) and reveled clustering of the 11q/j isolate among TPE strains , but not among TEN strains . The probability that the observed nucleotide sequence within this locus was caused by an accumulation of individual mutations instead of a recombination was pmut = 1 . 12245 x 10−65 ( pmut = ( 0 . 00074 x 0 . 333 ) 18 ) , since there was a total of 18 SNVs that differed from TEN strain Bosnia A but were identical to TPE strain Gauthier . Indels were omitted from the calculation . The sequences of 42 chromosomal regions , excluding TP0488 and TP0548 sequences , were concatenated and used to construct a phylogenetic tree to visualize the relatedness of 11q/j isolate to other treponemal genomes ( Fig 3C ) . The corresponding genome regions from the published whole genome sequences of two TPA strains ( Nichols , SS14 ) , four TPE strains ( CDC-2 , Gauthier , Samoa D , and Fribourg-Blanc ) , and TEN Bosnia A were used . Moreover , the dataset was supplemented with sequences of TEN Iraq B . All positions in the alignment containing gaps and missing data were eliminated resulting in a total of 29 , 447 positions in the final dataset having 509 variable sites . Overall , the 11q/j isolate clustered with both TEN Bosnia A and TEN Iraq B , indicating that most chromosomal loci of the11q/j isolate were consistent with TEN classification .
In this work , we analyzed an interesting human clinical isolate , 11q/j , that was first reported in 2013 as a case of syphilis [19] , but due to an unusual sequence pattern at the TP0548 locus , similar to TPE , it was thought to be an imported case of yaws [20] . In 2016 , the 11q/j isolate was further characterized in 7 genomic loci and classified as subspecies TEN [23] . Due to the unusual syphilis-yaws-bejel history of the 11q/j isolate we characterized larger genome regions of this clinical sample using different sequencing approaches . The small amount of treponemal DNA within the only available sample of the 11q/j isolate ( copy number less than 10 molecules of treponemal DNA per 1 μl ) with an excessive amount of contaminating human DNA , which exceeded the treponemal DNA by at least 100 , 000 times , precluded the use of other techniques that have been recently reported to be effective in sequencing treponemal DNA directly from clinical samples [34 , 35] . Efficient enrichment of treponemal DNA requires the number of treponemal copies > 1x104 per 1 μl [34] . In fact , enrichment of the TEN Iraq B DNA sample , containing 104 copies per 1 μl , revealed genome coverage less than 12 . 4% [35] . For these reasons , we mostly used nested PCR in this study . In all cases , amplification was done from samples containing at least 102 copies of treponemal DNA to avoid introduction of sequencing errors . As reported in a previous study based on analysis of 7 chromosomal regions , classification of the clinical isolate 11q/j was consistent with T . pallidum subsp . endemicum [21] . In this work , we confirmed this finding based on analyses of 42 chromosomal regions ( excluding the TP0488 and TP0548 loci ) , which were independently amplified and analyzed . The corresponding phylogenetic tree revealed a clear clustering of the 11q/j isolate with TEN strains ( Fig 3C ) . In addition , the genetic distance between the 11q/j isolate and TEN Bosnia A and Iraq B was greater than the distance between Bosnia A and Iraq B , indicating that the ancestor of the 11q/j isolate diverged before TEN Bosnia A and TEN Iraq B diversified . Although the sequenced portion of the 11q/j isolate represented less than 3% of the total genome length , the number of analyzed nucleotide positions informative for differentiation between TPE and TEN was much larger . Considering the extent of similarity of the genome sequences of available TPE and TEN strains ( i . e . , they are 99 . 91–99 . 94% similar ) , there were relatively few ( 711–970 ) variable sites between TEN Bosnia A and TPE strains Gauthier , CDC-2 and Samoa D [16] . Within the 11q/j isolate , 196 ( 20–28% ) of these variable sites were sequenced and 98% of them revealed sequence similarity to TEN strains . Therefore , it is very likely that the classification of 11qj isolate as TEN strain will remain the same even after acquisition of additional genomic sequences . Sequence analysis of TP0488 of the 11q/j sample revealed a sequence very similar to TPA strains . A similar situation has been previously reported in the genome of Bosnia A , where several chromosomal regions including TP0326 , TP0488 , TP0577 , TP0858 , TP0968 , and TP1031 showed striking similarity to TPA treponemes [16] . However , the TPA-like sequences at the TP0488 locus of Bosnia A and Iraq B strains were different from the TP0488 sequence of the 11q/j sample and were located between positions 1175–1195 ( Fig 2A ) , indicating that the 11q/j recombination event was independent of the recombination event at the TP0488 locus in the ancestor of TEN Bosnia A and TEN Iraq B . Interestingly , in the TPA Mexico A , TP0488 was found to contain a sequence very similar to that found in TPE strains , suggesting that the TP0488 locus is prone to gene recombination [36] . The TP0488 gene encodes a methyl-accepting chemotaxis protein ( Mcp2-1 ) [37] and , as shown by expression profiling of treponemes isolated from rabbit infections , is highly expressed in TPA strains [38] . Moreover , the Mcp2-1 protein has been shown to elicit a humoral response [37] . In the TPA Mexico A genome , 8 out of 18 TPE-like changes were located in the Cache domain ( domain binding small molecules ) [39] . Similarly , 13 out of 21 amino acid replacements resulting from recombination in the 11q/j isolate were also located in the Cache domain , suggesting differences in binding properties of the Mcp2-1 protein . As discussed in a previous work [36] , the observed sequence patterns are consistent with recombination events that have likely occurred during parallel human infections with both TPA and TEN or TPA and TPE treponemes . Gene TP0548 , on the other hand , was for the first time found to be recombinant . TP0548 from the 11q/j isolate appeared to be composed of sequences ( in addition to TEN sequences ) originating from TPE treponemes . Moreover , TEN Bosnia A and TEN Iraq B showed TPA-like sequences within this locus . For this reason , as well as the fact that the ancestor of the 11q/j isolate diverged before the ancestor of TEN Bosnia A and Iraq B strains , it is more plausible that the recombination event occurred in the common ancestor of Bosnia A and Iraq B rather than in the 11q/j isolate . The divergence of the ancestor of the 11q/j isolate before the ancestor of TEN Bosnia A and Iraq B strains is supported by greater genetic distances between the 11q/j isolate and both TEN strains ( Bosnia A , Iraq B ) compared to distances between Bosnia A and Iraq B ( Fig 3 ) . According to this scenario , the recombination occurred in a TEN strain that was ancestral to both the Bosnia A and Iraq B , which incorporated the TPA sequence into this locus ( Fig 3C ) . The sequence of the 11q/j isolate thus represents the original TEN sequence that is similar to TPE strains . The TP0548 was predicted to encode for a rare outer membrane protein [40] and , as shown by molecular typing studies , is highly variable among syphilis isolates [18 , 22 , 25 , 27] . The tendency of this locus to recombine , although shown only in TEN , should be considered during interpretation of data from both enhanced CDC and sequencing-based typing of syphilis-causing strains . The calculated probability that the observed SNVs within TP0488 and TP0548 were caused by random mutations was extremely low , i . e . 1 . 01987 x 10−83 and 1 . 12245 x 10−65 , respectively , suggesting that recombination occurred in these loci . Since mutation frequency per 1 bp within the TEN subspecies was calculated based on the sequences obtained for the 11q/j isolate , there was a potential bias in preferential sequencing of TEN variable regions . However , inclusion of additional chromosomal loci would likely lower the final probability even more . Moreover , the calculated SNV density in TEN strains ( 0 . 74 nt per 1000 bp ) differed only slightly compared to densities within other treponemal subspecies ( 0 . 36 nt per 1000 bp in TPA; 0 . 14 nt per 1000 bp in TPE; [29] ) . Both intra-genomic and inter-genomic recombination events have been identified in uncultivable pathogenic treponemes and are summarized in Table 1 . While intra-genomic homologous recombination have been found in tpr genes [7 , 10 , 41 , 42] , several inter-genomic recombination events have already been described in the literature [16 , 36] . The fact that the infection caused by the TEN 11q/j isolate resembled early syphilis with lesions located on the genitals supports previous findings that both TPA and TEN strains form similar , clinically undiscernible primary lesions . In a similar case , TEN Bosnia A was isolated from genital lesions of a 35-year old male , although in this case , lesions were also found in the oral cavity and pharynx and the patient showed secondary lesions on the face , trunk and extremities . The patient which was the source of the 11q/j isolate in this study , reported that he returned to France from Islamabad , Pakistan , where he admitted having had sexual contact with commercial sex worker . Since Pakistan is located close to countries that have recently reported cases of endemic syphilis , including Saudi Arabia and Iran [1] , such an infection should not be surprising . Given the known number of repetitions in the arp gene , the restriction pattern of the amplified tprEGJ genes , and the TP0548 sequence , the enhanced CDC genotype [18] can be deduced for TEN strains . For TEN Bosnia A and Iraq B , the 10q/c and 8q/c genotypes can be predicted based on published data , respectively [8 , 16 , 20] . Interestingly , similar subtypes have already been identified among tested clinical isolates from China including 9h/c , 10h/c , and 9o/c [43] . In fact , the electrophoretic tpr pattern “h” differs from the pattern “q” by one fragment ( i . e . , 804 bp in pattern “h” vs . 726 bp in “q” ) and “o” differs from pattern “q” by the absence of a 315 bp fragment [44] . Close similarity of identified subtypes of human syphilis isolates to predicted subtypes of TEN Bosnia A and TEN Iraq B suggests that TEN strains could and should be sporadically detected among human samples from patients suspected of having syphilis . In such situations , suspicious samples should be further analyzed to obtain an unequivocal classification of either a TPA strain or a TEN strain . Taken together , analysis of the 11q/j isolate revealed a TEN genome seemingly containing two recombination events and highlights the fact that TEN strains could cause syphilis-like lesions in humans . A more detailed analysis revealed that the 11q/j isolate had just one recombinant locus , TP0488 . The recombination of TP0548 took place in a treponeme that was the ancestor of both TEN Bosnia A and TEN Iraq B .
|
Treponema pallidum subsp . endemicum ( TEN ) is an uncultivable pathogenic treponeme that causes bejel ( endemic syphilis ) , a chronic human infection mostly affecting children under 15 years of age , occurring mainly in several African and Middle East countries . In this work , we characterized a TEN 11q/j isolate from France that was obtained from an adult male with genital lesions , who was suspected of having syphilis and who received benzathine penicillin G . DNA sequencing of the isolate revealed two loci that were , rather than to TEN , related either to T . pallidum subsp . pertenue or to T . pallidum subsp . pallidum and likely resulted from recombination events . The recombination event in TP0488 as well as the recombination in TP0548 , of the 11q/j , helped clarify the phylogeny of the TEN strains indicating that the recombination in TP0548 took place in a treponeme that was ancestral of Bosnia A and Iraq B , but was not an ancestor of the 11q/j isolate . In contrast , a recombination event in TP0488 appeared in the ancestor of the 11q/j isolate after separation of the ancestral treponeme of Bosnia A and Iraq B . This case also points to a possible role of TEN strains in development of syphilis-like lesions in countries with endemic syphilis .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"geographical",
"locations",
"microbiology",
"sequence",
"assembly",
"tools",
"genome",
"analysis",
"molecular",
"biology",
"techniques",
"bacterial",
"pathogens",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"artificial",
"gene",
"amplification",
"and",
"extension",
"bioinformatics",
"proteins",
"medical",
"microbiology",
"microbial",
"pathogens",
"recombinant",
"proteins",
"molecular",
"biology",
"genetic",
"loci",
"people",
"and",
"places",
"biochemistry",
"biomolecular",
"isolation",
"treponema",
"pallidum",
"dna",
"sequence",
"analysis",
"asia",
"dna",
"isolation",
"database",
"and",
"informatics",
"methods",
"genetics",
"biology",
"and",
"life",
"sciences",
"genomics",
"iraq",
"computational",
"biology",
"polymerase",
"chain",
"reaction"
] |
2017
|
Human Treponema pallidum 11q/j isolate belongs to subsp. endemicum but contains two loci with a sequence in TP0548 and TP0488 similar to subsp. pertenue and subsp. pallidum, respectively
|
Vibrio parahaemolyticus is a leading cause of seafood-borne gastroenteritis in many parts of the world , but there is limited knowledge of the pathogenesis of V . parahaemolyticus-induced diarrhea . The absence of an oral infection-based small animal model to study V . parahaemolyticus intestinal colonization and disease has constrained analyses of the course of infection and the factors that mediate it . Here , we demonstrate that infant rabbits oro-gastrically inoculated with V . parahaemolyticus develop severe diarrhea and enteritis , the main clinical and pathologic manifestations of disease in infected individuals . The pathogen principally colonizes the distal small intestine , and this colonization is dependent upon type III secretion system 2 . The distal small intestine is also the major site of V . parahaemolyticus-induced tissue damage , reduced epithelial barrier function , and inflammation , suggesting that disease in this region of the gastrointestinal tract accounts for most of the diarrhea that accompanies V . parahaemolyticus infection . Infection appears to proceed through a characteristic sequence of steps that includes remarkable elongation of microvilli and the formation of V . parahaemolyticus-filled cavities within the epithelial surface , and culminates in villus disruption . Both depletion of epithelial cell cytoplasm and epithelial cell extrusion contribute to formation of the cavities in the epithelial surface . V . parahaemolyticus also induces proliferation of epithelial cells and recruitment of inflammatory cells , both of which occur before wide-spread damage to the epithelium is evident . Collectively , our findings suggest that V . parahaemolyticus damages the host intestine and elicits disease via previously undescribed processes and mechanisms .
Vibrio parahaemolyticus is a Gram-negative bacterium that resides in the marine environment , often in association with shellfish . It is a leading cause of gastroenteritis linked to consumption of raw or undercooked seafood throughout the world , and especially in Asia [1] . Infections caused by V . parahaemolyticus can occur sporadically or in outbreaks , which can be relatively large . For example , in 2005 , more than 10 , 000 people were sickened by V . parahaemolyticus-contaminated clams and mussels in Chile [2] . The most common clinical manifestation of V . parahaemolyticus infection of the gastrointestinal tract is acute , self-limited watery diarrhea that is often accompanied by abdominal pain , nausea and vomiting . However , the severity of symptoms can range from mild watery diarrhea to a severe dysentery-like illness [3] . V . parahaemolyticus has also been linked to wound infections and septicemia; however , such infections are far less common and generally not food-borne . Despite the prevalence of V . parahaemolyticus-induced gastroenteritis , there is limited understanding of the pathogenesis of V . parahaemolyticus-induced diarrhea . In part , this paucity of knowledge is explained by the fact that few analyses of intestinal tissue from V . parahaemolyticus-infected patients have been performed , few human volunteer studies have been carried out , and there is no oral-infection-based animal model that closely resembles the human disease . However , human studies suggest that V . parahaemolyticus infection disrupts the intestinal epithelium . Autopsy findings from the first reported epidemic of V . parahaemolyticus in Japan noted ‘slight erosion of the jejunum and ileum’ in several patients who died due to the infection [4] . In the most extensive study to date , rectal and duodenal biopsies from patients with acute V . parahaemolyticus infections revealed ‘epithelial degeneration and denudation’ at both sites [5] , creating some uncertainty regarding the intestinal site ( small or large bowel ) and processes that give rise to the watery diarrhea that usually accompanies V . parahaemolyticus infection . Additionally , there was evidence of an acute inflammatory response , marked by polymorphonucleocyte ( PMN ) infiltration at both sites , as well as elevated levels of TNF-α and IL-1β in stool , and TNF-α in blood [5] . Thus , studies to date suggest that V . parahaemolyticus causes disease via different mechanisms than the related pathogen V . cholerae , which does not disrupt the intestinal epithelium nor induce significant inflammation in infected individuals ( reviewed in [6] ) . Comparison of clinical and environmental isolates of V . parahaemolyticus has enabled identification of several likely virulence factors in this pathogen . Of note , most clinical isolates exhibit β-hemolytic activity on Wagatsuma agar ( aka the Kanagawa phenomenon ) [7] , [8] , while this phenotype is lacking in most environmental isolates [9] . Hemolytic activity is mediated by the thermostable direct hemolysin ( TDH ) [10] , and has also been demonstrated for the Tdh-related hemolysin ( TRH ) using human erythrocytes [11] . Recently , genome sequencing has revealed that pathogenic isolates of V . parahaemolyticus also encode two type III secretion systems ( T3SS ) [12] . T3SS have been found to contribute to the virulence of many bacterial pathogens , as they enable injection of bacterial proteins ( effectors ) directly into host cells and subsequent modulation of numerous host processes [13] . T3SS2 of V . parahaemolyticus is encoded in a pathogenicity island on the small chromosome ( chrII ) of pathogenic strains . This pathogenicity island also encodes TDH; consequently , the Kanagawa phenomenon is indirectly a marker for 2 distinct potential virulence factors . T3SS1 is encoded on the bacterium's large chromosome ( chrI ) , and unlike T3SS2 , appears to be ubiquitous within both clinical and environmental isolates of V . parahaemolyticus [12] . Several in vitro and in vivo assays have been used to characterize the effects of V . parahaemolyticus and its putative virulence loci upon host cells and tissues . V . parahaemolyticus causes a variety of changes in cultured cell lines , including release of cellular contents , cell rounding , disruption of tight junction complexes and cytoskeletal structures , and ion influx ( reviewed in [14] ) . In particular , tissue-culture-based assays have been useful for linking T3SS1 with cytotoxicity [15]–[21] , although the relationship between cytotoxicity and enteritis remains unclear . The effects of V . parahaemolyticus upon animal tissues have also been studied , primarily using rabbit ligated ileal loops , a model of intestinal fluid response . In this model , both TDH and T3SS2 contribute to fluid accumulation , with T3SS2 exerting the predominant effect [19] , [22] . Studies using ligated ileal loops also suggest that T3SS2 is linked to inflammation and epithelial denudation . T3SS2 was likewise required to induce mild transient ( ∼4 hr ) diarrhea in 2-day-old piglets orally infected with V . parahaemolyticus; however , pig intestines showed minimal histopathologic abnormalities , even when infected with wild type bacteria [23] . Limitations in each of these models indicate that studies of the pathogenesis of V . parahaemolyticus-linked enteritis would be transformed by development of a non-surgical small animal model in which bacterial colonization and host response , including diarrhea , histopathology and inflammation , could all be monitored and evaluated . Here , we report that infant rabbits oro-gastrically inoculated with V . parahaemolyticus develop severe diarrhea and enteritis . The pathogen primarily colonizes the distal small intestine , the site where inflammation and dramatic histopathologic and ultrastructural changes in the epithelium were also observed . Similar to attaching and effacing ( A/E ) pathogens such as enteropathogenic E . coli ( EPEC ) , which also cause intestinal disease , and in marked contrast to V . cholerae , V . parahaemolyticus causes effacement of the microvilli . However , V . parahaemolyticus causes more widespread disruption of villus structure than A/E pathogens . V . parahaemolyticus proliferates in epithelial cavities , initially forming large , dense microcolonies and ultimately inducing extensive extrusion and/or erosion of villous epithelial cells and a loss of epithelial barrier function in the small intestine . V . parahaemolyticus proliferates in cavities created in the epithelium forming large , dense microcolonies . T3SS2 proved to be essential for V . parahaemolyticus to colonize the small intestine and cause disease , but T3SS1 and TDH also modulate V . parahaemolyticus virulence in the intestine . When taken together , our findings suggest that V . parahaemolyticus damages the host intestine via a previously undescribed process and that infant rabbits are an outstanding model host for investigating V . parahaemolyticus pathogenicity .
We tested whether rabbits could be used as a model host to study V . parahaemolyticus-induced gastroenteritis , as we previously found that the intestinal diseases caused by V . cholerae O1 , V . cholerae non-O1 non-O139 , and enterohemorrhagic E . coli ( EHEC ) can be successfully modeled using these animals [24]–[27] . Based on our previous experience , 3-day-old rabbits were treated with cimetidine , a histamine H2 receptor antagonist that transiently alleviates gastric acidity , prior to bacterial inoculation [28] . Most rabbits oro-gastrically inoculated with V . parahaemolyticus developed severe diarrhea , but the clinical course and kinetics of disease in V . parahaemolyticus-infected rabbits differed from the diseases caused by V . cholerae O1 , V . cholerae non-O1 non-O139 serogroups , or EHEC . In almost all rabbits , inoculation with 1×109 colony forming units ( cfu ) of V . parahaemolyticus resulted in release of loose , unformed gelatinous stools , followed by liquid yellow diarrheal fluid , which typically developed between 20–40 hr after infection ( Figure 1A ) and soaked the ventral surfaces of the rabbits . If the experiments were continued , the animals subsequently lost weight , became lethargic and died by 64 hr ( Figure 1A and Table 1 ) . Thus , in most experiments described below , we used end points at or prior to 38 hr to enable a variety of analyses at times when most rabbits exhibited disease but had not yet succumbed to infection . At 38 hr post-infection ( PI ) , the intestines of most surviving rabbits were swollen and filled with fluid ( Figure 1B ) , with elevated fluid accumulation ratios versus mock-infected rabbits ( Table 1 ) . Notably , the fluid contained higher levels of total protein ( 0 . 53 vs 0 . 2 g dL−1 ) and K+ ( 35 vs 18 mmol dL−1 ) than the fluid from rabbits infected with V . cholerae [24] . The elevations in the levels of protein and K+ are consistent with the marked disruption of small intestinal villous epithelial cells observed in V . parahaemolyticus-infected animals ( see below ) and suggestive of a diarrheal mechanism that is distinct from the classic secretory diarrhea of cholera , where the intestinal epithelium remains intact [29] . The lower half of the small intestine appears to be the primary target of V . parahaemolyticus . The highest numbers of V . parahaemolyticus were recovered from homogenates of the mid and distal regions of the small intestine ( ∼109 cfu g−1 ) ; ∼10–200-fold fewer cfu were recovered from the cecum , colon and proximal small intestine ( Figure 1C ) . Additionally , histologic analyses revealed the more extensive pathologic changes to the tissue occurred within the lower third of the small intestine ( Figure 1D ) . In hematoxylin and eosin ( H&E ) -stained sections from the distal small intestine , the tips of the villi appeared ragged and irregular , and epithelial cell debris from the disrupted villi was detected in the intestinal lumen ( compare Figures 1D & E ) . Discrete and oftentimes dense clusters of bacteria adherent to the epithelium were also evident in these sections ( Figure 1F , arrowhead ) . Heterophils , the rabbit equivalent of neutrophils , were observed in the lamina propria as well as apparently migrating through the epithelial barrier toward adherent bacteria and the intestinal lumen ( Figure 1F arrow ) ; degranulated heterophils were seen as well ( Figure 1G ) . Consistent with the histologic evidence of marked inflammation , 4–20 fold elevations in the abundance of transcripts for IL-8 , IL-6 , TNF-α , and IL-1β by quantitative PCR analyses of RNA were found in homogenates of the distal small intestines of infected relative to uninfected rabbits ( Figure 1H ) . In addition , the zone of proliferating cells at the base of the villi was larger in infected animals than in controls , suggesting that V . parahaemolyticus infection induces host cell proliferation ( see below ) . Pathological abnormalities in the cecum were limited to moderate sub-mucosal edema which contained few inflammatory cells . Importantly , attached bacteria and histologic abnormalities were not observed in the colon , despite large numbers of bacteria recovered from tissue homogenates taken from this site . Finally , blood samples and homogenates of the spleen , gall bladder and liver rarely contained V . parahaemolyticus ( bacteria were recovered in 3 of 16 samples at approx . 10–100 cfu mL−1 homogenate ) , suggesting that the infection does not ordinarily extend beyond the intestine in the infant rabbits . Collectively , these observations show that V . parahaemolyticus adheres to and colonizes the distal small intestine of the infant rabbit , where it induces inflammatory enteritis accompanied by severe disruption of the epithelial lining . To gain insight into the temporal progression of V . parahaemolyticus-linked pathological changes in the small intestine , as well as their relationship to overt signs of disease , we compared tissue from the distal small intestines of infected rabbits at 12 , 18 , 28 and 38 hr PI . Bacterial numbers within tissue homogenates increased markedly between 12 and 18 hr PI ( Figure 2A . ) After 18 hr PI , the number of organisms recovered from the distal small intestine remained relatively constant , despite the onset of diarrhea ( and the loss of bacteria-laden fluid containing 109 cfu mL−1 ) at ∼28 hr; thus , bacterial proliferation clearly continues despite the absence of an increase in the measured bacterial load within the intestine . The amount of fluid accumulated in the distal small intestine increased gradually , reaching statistically greater levels than in mock-infected rabbits at 38 hr PI ( Figure 2B ) . Bacterial abundance as assessed by plating tissue homogenates only partially corresponded to the microcolonies of bacteria evident in H&E-stained tissue sections . Bacterial microcolonies were not usually detected in tissue sections at 12 hr PI; however , by ∼18 hr PI , small clusters were more frequently found on the epithelial surface ( Figure 2C ) . By 28 hr PI , the bacterial microcolonies were larger and more abundant , and appeared to be located within ‘cavities’ on the epithelial surface ( Figure 2D–F ) . The observed increase in visible bacterial microcolonies at 28 hr relative to 18 hr , which did not correspond with any further increase in the concentration of bacteria recovered in tissue homogenates , may reflect better attachment of the bacteria by 28 hr or better retention of bacteria in the cavities during staining of tissue sections . Finally , at 38 hr PI , large numbers of bacteria , both as individual cells and as large ( sometimes exceeding 50 µm in diameter ) , dense microcolonies , were associated with the villi and the luminal cellular debris ( Figure 2G ) . Bacteria within tissue sections were confirmed to be V . parahaemolyticus via infection of rabbits with a strain that constitutively expresses green fluorescent protein ( GFP ) ; confocal microscopy analyses revealed that the distribution of GFP-expressing bacteria , both within cavities on the epithelial surface ( Figure 2E ) , and in association with epithelial debris corresponded closely to that seen in H&E-stained sections . Furthermore , no bacteria were seen in mock-infected rabbits at any point . Serial histologic analyses also revealed a striking characteristic progressive disruption of the small intestine's epithelial morphology during the infection . At 12 and 18 hr PI , intestinal villi appeared intact , with a fairly smooth surface , although by 18 hr , heterophils were observed in the lamina propria of villi , where small clusters of bacteria were also often observed to be attached ( Figure 2C , data not shown ) . Even at this earlier time point , adherent bacterial clusters were associated with erosions in the epithelial surface ( Figure 2C inset ) . By 28 hr PI , these erosions were more pronounced , generating cavities in the epithelial surface and resulting in loss of the peripheral actin ring that ordinarily encircles each villus ( Figure 2D ( inset ) , 2E ) . Formation of these cavities appears to rely on extrusion of epithelial cells and/or their contents ( Figure 2F ) . V . parahaemolyticus cells were often observed at the base of extruding epithelial cells ( Figure 2F , arrowheads ) . At 28 hr , heterophils had become more abundant within the lamina propria , and a few were observed in the intestinal lumen ( Figure 2D , long arrows ) . However at this point , most of the epithelial layer remained intact and little luminal debris was observed . Extensive disruption of the epithelium and disintegration of villi was not evident until 38 hr PI; at this point , bacteria were often associated with epithelial cells that were loosely connected together , giving villi a ‘flower-like’ appearance , and with the abundant epithelial debris within the intestinal lumen ( Figure 2G , also Figure 1D ) . Luminal debris consisted of both whole cells ( with nuclei ) and membrane-bound fragments of cytoplasmic material ( Figure 2G ) . At this time point , while increasing numbers of heterophils continued to be recruited into the intestine , it was somewhat surprising that more were not present given the amount of tissue destruction that was observed . However , immunostaining revealed that macrophages , which were present in low numbers at earlier time points , were more abundant at 38 hr PI , both in the tissue and in the intestinal lumen ( Figure S1 ) . Collectively , these analyses suggest that the initial attachment of V . parahaemolyticus to discrete areas in the distal small intestine initiates a cascade of changes in the host . As the pathogen adheres and proliferates , small surface erosions give rise to bacterial-filled cavities in the epithelium , followed by disintegration of the villi . An acute ( heterophil-based ) inflammatory response occurs at the same time as early epithelial surface erosion , when a relatively small number of bacteria are attached to the surface . Thus , the inflammatory response to V . parahaemolyticus infection precedes , rather than occurs as a result of , the extensive tissue disruption observed at later ( e . g . , 38 hr ) stages in the infection . In normal tissue , the loss of dying or damaged epithelial cells from the villus is balanced by the generation of new progenitor cells by proliferation in the crypts . Thus , to explore the relationship between intestinal cell proliferation and epithelial cell loss , we also assessed cell proliferation at 18 and 28 hr PI . Tissue sections were labeled with anti-Ki67 , an intrinsic marker for actively dividing cells [30] , and the zone of proliferating cells relative to villus length was calculated ( Figure S2 ) . Strikingly , even at 18 h PI , the zone of proliferating cells was significantly greater in infected compared to uninfected rabbits ( mean ± SEM: 0 . 52±0 . 06 and 0 . 21±0 . 02; P<0 . 01 ) . This difference was also apparent at 28 hr PI ( 0 . 65±0 . 04 and 0 . 36±0 . 06; P<0 . 05 ) . ( We did not perform Ki67 staining at 38 hr PI due to the massive tissue disruption that had occurred ) . These data strongly suggest that epithelial cell proliferation in response to V . parahaemolyticus , like the inflammatory response , does not occur as a result of tissue disruption , but is instead an early step in the host response to infection . To further characterize how V . parahaemolyticus interacts with epithelial cells , we used scanning and transmission electron microscopy ( EM ) to visualize the epithelium in the small intestine of infected rabbits . Numerous clusters of attached bacteria were observed with scanning EM by 28 hr PI , particularly near the villus tips ( Figure 3A ) ; bacteria were not observed within the crypts . Epithelial cells with attached V . parahaemolyticus exhibited dramatic ultrastructural changes , whereas epithelial cells without bacteria appeared normal ( Figure 3B , data not shown ) . Attached bacteria were typically found in clusters , with individual cells often oriented perpendicular to the epithelial surface ( Figure 3B ) . Notably , the appearance of the epithelial surface surrounding bacterial clusters was grossly distorted by the presence of numerous elongated ( 5–10 µm long ) hair-like cellular projections ( Figure 3B , arrowheads; Figure S3A ) . In transmission EM , depending on the angle of sectioning , these protrusions appear as a disorganized mix of cross-sections or short fragments that extend well beyond the range of the normal brush border ( Figure 3C ) . Higher magnification images revealed that these structures were surrounded by a clearly defined membrane and contained internal filaments ( Figure 3D ) , two features that strongly suggest that these protrusions represent elongated microvilli . Similar but shorter projections , also thought to represent elongated microvilli , were observed around clusters of EPEC attached to human duodenal tissue explants [31] , [32] . Although V . parahaemolyticus induces elongation of microvilli from the surface of epithelial cells near attached bacteria , in virtually all EM images , V . parahaemolyticus was only closely apposed to host cells without detectable microvilli ( Figure 3E ) . Microvilli loss ( effacement ) appears to occur by the process of membrane vesiculation [33] , as vesicles with poorly defined membranes and lacking core proteins were observed near the attached bacteria ( Figure S3B ) . Some V . parahaemolyticus cells closely apposed to the epithelial cell membrane appeared to be held in ‘cup-like’ structures that are reminiscent of the lesions caused by A/E pathogens ( Figure 3E , arrowheads ) . However , with V . parahaemolyticus there was limited evidence of actin accumulation beneath the attached bacteria and pedestals were never observed , suggesting that the adherence mechanisms of these two enteric pathogens are not equivalent . In addition , the villi destruction caused by V . parahaemolyticus ( which contrasts dramatically to the intact epithelium of tissue infected with V . cholerae O1 ( [34]; Figure S3C ) ) is far more severe and widespread than is typically seen with A/E pathogens . Consistent with our observations with H&E-stained sections , in transmission EM images , bacterial clusters were typically seen beneath the level of the surrounding ( intact ) epithelium ( Figure 3F ) . EM evidence for both extrusion of epithelial cells ( Figure 3G ) and their contents ( see Figures 3B ( arrow ) and 3H ) was observed , suggesting that both processes contribute to the development of cavities in the epithelial surface . Cell shedding is a normal process in which apoptotic epithelial cells are routinely extruded into the lumen [35] . Re-distribution of cytoskeletal proteins ( e . g . , actin ) as well as proteins that form apically located tight junction complexes ( e . g . , ZO-1 , claudins and occludin-1 ) is thought to contribute to cell shedding while maintaining the integrity of the epithelial barrier [36] , [37] . Therefore , to begin to investigate the molecular processes that underlie V . parahaemolyticus-induced cell extrusion from the small bowel epithelium , we compared the localization of host cytoskeletal components in tissues from infected and mock-infected rabbits . Re-distribution of F-actin , ZO-1 , and occludin-1 from their normal locations at the apical cell periphery and the apical boundaries between cells to the lateral membrane was apparent for cells undergoing extrusion . The redistributed proteins formed an intense focal point or funnel-like structure at the base of the extruding cell ( Figure 4A , B ) , similar to structures that have been observed during both physiologic and pathologic epithelial cell shedding [36] , [37] . These observations suggest that V . parahaemolyticus may usurp at least certain components of the normal shedding process to promote cell extrusion during the course of infection . However , in contrast to ZO-1 and occludin-1 , claudin-1 , which is located along the lateral membranes in uninfected tissue , did not appear to co-localize with F-actin at the lateral membrane as occurs in physiologic and pathologic cell shedding ( Figure 4C; [37] ) . The claudin family of proteins is thought to control the paracellular permeability of tight junctions [38] , [39] and so the aberrant localization of claudin-1 in extruding cells of infected tissues may contribute to fluid loss . Furthermore , while shedding cells normally exhibit signs of apoptosis [35] , [40] , most cells undergoing V . parahaemolyticus-induced extrusion did not , as no TUNEL staining was detected in extruding cells ( Figure 4D ) . Thus , V . parahaemolyticus appears to elicit extrusion of cells that would not ordinarily undergo shedding . TUNEL-positive cells were present in the luminal debris ( Figure 4D ) especially at later points in the infection , suggesting that cell death occurs after extrusion during V . parahaemolyticus infection . Our observations of increased epithelial cell shedding and the formation of bacteria-filled cavities in the epithelium of V . parahaemolyticus-infected rabbits raised the possibility that infection might compromise epithelial barrier function , even when the intestinal tissue appears largely intact . Consequently , we tested whether a small molecular tracer , biotin , could penetrate beyond the luminal surface of the epithelium after injection into the dissected distal small intestine or colon of infected and control rabbits at 25 hr PI . In the mock-infected rabbits , biotin remained localized to the luminal side of the epithelial barrier in the small intestine and colon ( Figure 5A; data not shown ) . In contrast , biotin was detected in the lamina propria as well as on the luminal surface of tissue from the small intestine of V . parahaemolyticus-infected rabbits ( Figure 5B ) , even though at the time point assayed the villi structure was not extensively disrupted . The route by which biotin reaches the lamina propria cannot definitively be discerned from this experiment; however , V . parahaemolyticus-induced disruptions in the proteins that form the tight junction complex , and a resulting increase in paracellular permeability , seem likely to play an important role . No penetration of biotin was observed in the colon of infected rabbits , consistent with the previously noted lack of bacterial attachment and pathology at this site ( Figure S4 ) . This finding provides additional evidence that the small intestine is the pathologically relevant site during V . parahaemolyticus infection . To begin to address which bacterial factors are important in V . parahaemolyticus-induced enteritis , we assessed colonization by and disease associated with a previously described set of isogenic V . parahaemolyticus mutants , each lacking one or more of the 3 principal described virulence factors: TDH ( Δtdh ) , T3SS1 ( ΔvscN1 , an essential component of T3SS1 ) and T3SS2 ( ΔvscN2 , an essential component of T3SS2 ) [22] . A triple Δtdh ΔvcrD1 ΔvcrD2 mutant ( vcrD1 and vcrD2 are also essential components of T3SS1 and T3SS2 , respectively ) did not cause disease or elicit inflammation in rabbits ( Table 1 , Figure 6 ) . Of the single deletions , inactivation of T3SS2 had the most dramatic attenuating effect , which closely resembled that of the triple mutation . Of 12 rabbits infected with the ΔvscN2 mutant , none developed diarrhea , exhibited intestinal fluid accumulation , or induced histological changes in intestinal tissue ( Table 1 , Figure 6 and Figure S5 ) . Notably , there was a more than 4-log reduction in the number of ΔvscN2 mutant bacteria recovered from the small intestine relative to the wild type strain ( Figure 6A–B ) . Thus , the T3SS2 is essential for V . parahaemolyticus intestinal colonization . As such , it is likely that the profound attenuation of virulence of the ΔvscN2 mutant at least in part reflects its reduced capacity for intestinal colonization . However , since previous studies showed that T3SS2 accounted for most of the epithelial disruption and inflammation in ileal loops [19] , a closed system in which many colonization factors may not be essential , it is likely that T3SS2 plays a critical role both in colonization and subsequent events in pathogenesis . Inactivation of T3SS1 had a more subtle effect; however , our analyses suggest that this apparatus also contributes to V . parahaemolyticus-induced enteritis . Intestinal colonization by the ΔvscN1 mutant appeared to be slightly lower than that of the wild type strain , although this difference was not statistically significant ( Figure 6A–C ) . As visualized by scanning EM , the ΔvscN1 mutant induced microvillous elongation and formed microcolonies that were grossly indistinguishable to those observed during wild type infection . However , there was a significant reduction in the percentage of rabbits with diarrhea ( 27% vs 70% for the wild type strain ) . Additionally , rabbit intestines contained significantly less fluid when infected with the ΔvscN1 mutant , as reflected by the fluid accumulation ratio within the distal small intestine , although unlike the case with the ΔvscN2 mutant , some fluid was still observed ( Table 1 ) . Inactivation of T3SS1 had no apparent effect on heterophil infiltration , epithelial cell sloughing , or cell proliferation ( Figure 6D–F; Figure S5 ) . Furthermore , a mutant lacking T3SS1 ( and also TDH , known as POR-2 ) showed no reduction in cytokine production relative to the wild type ( Figure 1G ) , suggesting that the secretion system is not required for V . parahaemolyticus to induce an inflammatory response . In contrast to T3SS2 and T3SS1 , our data suggests that TDH dampens at least some aspects of the host response to V . parahaemolyticus infection . Rabbits infected with V . parahaemolyticus lacking TDH developed diarrhea and had a fluid accumulation ratio significantly greater than that of wild type bacteria ( Table 1 ) . Compared to the wild type , similar numbers of the tdh mutant were recovered in most intestinal sections; however , significantly higher numbers were recovered in the proximal small intestine suggesting that this mutant is better able to colonize this region of the intestine ( Figure 6A ) . By scanning EM , TDH does not appear to be necessary for microvillous elongation or the formation of microcolonies on the epithelial surface as these features were grossly similar in rabbits infected with the wild type or the tdh mutant . Heterophil recruitment and epithelial cell sloughing were also not altered in response to the tdh mutant relative to the wild type strain ( Figure 6D , F; Figure S5 ) ; however , in places , the superficial mucosa appeared necrotic , suggesting a more severe pathological reaction was occurring ( Figure S6 ) . Additionally , the zone of proliferating cells was significantly enlarged in response to the tdh mutant ( Figure 6E; Figure S5 ) . The presence of augmented proliferation in the absence of increased heterophil infiltration observed with the tdh mutant suggests that these two responses to V . parahaemolyticus infection occur independently .
We have developed a simple non-surgical oral infection model of V . parahaemolyticus-induced intestinal pathology and diarrhea . This experimental model enabled us to define several previously unknown but key features of the pathogenesis of the disease elicited by this common agent of seafood-borne gastroenteritis . First , we discovered that this organism chiefly colonizes the distal small intestine , the region of the intestine that is also the major site of V . parahaemolyticus-induced damage , increased permeability of the epithelial barrier and inflammation . Together , these observations strongly suggest that disease in this region of the gastrointestinal tract accounts for most , if not all , of the diarrhea that accompanies V . parahaemolyticus infection . Second , we found that the V . parahaemolyticus T3SS2 is essential for the pathogen to colonize the small intestine; prior to our work no V . parahaemolyticus intestinal colonization factors had been definitively identified because of the absence of a robust animal model . Third , we observed that V . parahaemolyticus causes marked disruption of the villous epithelial surface in the small intestine . Effacement of microvilli , re-distribution of cytoskeletal and tight junction proteins , and extrusion of epithelial cells in the small intestine all appear to contribute to villus disruption and the breakdown of epithelial barrier function . Furthermore , the pathogen induces remarkable elongation of microvilli in epithelial cells adjacent to attached V . parahaemolyticus . Finally , early in the infection , before widespread damage to the epithelium becomes evident , V . parahaemolyticus induces both proliferation of intestinal epithelial cells and recruitment of inflammatory cells . Thus , our observations suggest that V . parahaemolyticus elicits disease via a previously undescribed sequence of events that , to our knowledge , differs from those outlined for other enteric pathogens . V . parahaemolyticus , like V . cholerae O1 [24] , preferentially colonizes the distal small intestine . However , the manner in which these two pathogenic vibrios associate with the host epithelial surface differs . Prototypical V . cholerae O1 colonizes as layers of cells embedded in mucin-rich material that covers much of the epithelial surface of the villi as well as the crypts [24] . In contrast , V . parahaemolyticus colonizes as more discrete clusters of bacteria ( i . e . microcolonies ) that predominantly localize to the upper half of the villi; furthermore , unlike V . cholerae O1 [24] , V . parahaemolyticus does not induce goblet cell degranulation . The mechanism ( s ) that hold the V . parahaemolyticus microcolonies together are not known . However , since the microcolonies appear to form during contact with the epithelial surface , it is tempting to speculate that V . parahaemolyticus lateral flagella , which can be induced by surface contact [41] , [42] , may promote microcolony formation . These lateral cell appendages have been reported to form linkages between neighboring bacteria and the surface [43] . Induction of peritrichous flagella is associated with conversion of V . parahaemolyticus from small ( 2–3 µm ) , polarly flagellated swimmer cells to swarmer cells , which are elongated ( 5–20 µm ) as well as peritrichously flagellated [42] , [44] . Interestingly , we observed V . parahaemolyticus cells up to 10 µm in length in electron micrographs of infected tissues ( Figure S3D , asterisk ) , consistent with the idea that these cells have switched from the swimmer to the swarmer cell state . The consequences of V . cholerae O1 and V . parahaemolyticus colonization of the small intestine are also fundamentally different . V . cholerae O1 does not damage the surface of the intestinal epithelium; instead , the cholera pathogen grows on the luminal side of the microvilli which remain largely intact ( [34] and see Figure S3C ) . In marked contrast , V . parahaemolyticus damages the epithelial surface , leading to villus disintegration . Our histologic analyses of samples from 12 to 38 hr PI suggest that there is a characteristic sequence of steps through which V . parahaemolyticus proceeds to cause villus disruption in the small intestine ( Figures 2 and 7 ) . Initial attachment of V . parahaemolyticus to the epithelial surface is associated with effacement of microvilli and apparent depletion of cytoplasmic contents; thus , even at 18 hr PI , attached V . parahaemolyticus was observed situated below the epithelial surface ( Figure 2C ) . Both marked depletion of epithelial cell cytoplasmic contents as well as epithelial cell extrusion contribute to the formation of these V . parahaemolyticus-filled cavities in the epithelial surface . It is not clear if V . parahaemolyticus penetration of the paracellular space to reach the base of the extruding cell is a required step for extrusion , though this was often observed ( Figures 2F and 3G ) . The benefit of the epithelial cavities for the pathogen is not known , but it seems plausible that the cavities may provide the pathogen increased access to nutrients , or serve as a niche , offering protection from peristaltic flow . The molecular mechanisms by which V . parahaemolyticus elicits epithelial cell extrusion are not known . The process bears some similarity to the normal process that leads to shedding of apoptotic epithelial cells [35] . V . parahaemolyticus-induced cell extrusion , as in physiologic cell shedding , is accompanied by redistribution of tight junction-associated proteins , including ZO-1 and occludin-1 , towards the basolateral membrane where they co-localized with actin to form a funnel-like structure . However , claudin-1 did not follow this pattern in extruding cells of infected tissues , as occurs in physiologic and pathologic shedding [37] . Furthermore , while paracellular barrier integrity is maintained during physiological cell shedding , there is an increase in paracellular permeability observed in infected rabbits . V . parahaemolyticus-induced cell shedding could be a direct effect of the pathogen ( e . g . , perhaps a consequence of the activities of translocated T3SS effectors on tight junction complexes ) , or an indirect consequence of infection . For example , certain pro-inflammatory cytokines including TNF ( whose production is stimulated by V . parahaemolyticus ) have been shown to elicit cell shedding [37] , [45] . Regardless of the mechanisms leading to epithelial cell extrusion , V . parahaemolyticus-induced break down in epithelial barrier function , which has also been observed in a polarized epithelial monolayer [46] , likely contributes to the loss of intestinal fluid ( diarrhea ) caused by V . parahaemolyticus . Pathogen-induced epithelial cell extrusion has also recently been detected for other enteric pathogens . For example , EHEC induces extrusion of cells from polarized monolayers and during infection of calves [47] , [48] . Tissue-culture-based studies have revealed that the EHEC T3SS effector EspM , which interferes with the RhoA-signaling pathways that regulate actin cytoskeleton dynamics in eukaryotic cells , is sufficient to cause extrusion [47] . The purpose of pathogen-induced cell extrusion is difficult to ascertain , particularly since it can have benefits for both the pathogen and the host . For Salmonella , intestinal epithelial cell extrusion has been shown to promote the pathogen's spread within , and escape from , the intestinal tract [49] . Cell extrusion could also promote the egress of V . parahaemolyticus from the intestine since extruded cells and debris often had adherent V . parahaemolyticus . However , it is also possible that extrusion aids the host , by enabling shedding of adherent bacteria . In support of this idea , it has been found that several bacterial pathogens ( e . g . , Shigella , Neisseria ) produce factors that appear to counteract epithelial shedding [50] . Some of the ultrastructural changes V . parahaemolyticus elicits in intestinal epithelial cells are reminiscent of phenotypes previously described for EPEC , a member of the A/E family of pathogens [31] , [32] . Similar to EPEC-induced changes in small intestine explants , we observed long spaghetti-like protrusions from epithelial cells surrounding the edges of the V . parahaemolyticus clusters , effacement of microvilli , and the close apposition of individual V . parahaemolyticus cells to the effaced epithelial surface within cup-like structures ( see Figure 3E ) . The mechanism ( s ) that mediate these dramatic alterations in host epithelial cell morphology remain to be determined . However , it seems likely that , similar to EPEC , the activities of some of the effectors translocated by one or both of the V . parahaemolyticus T3SS manipulate the host cytoskeleton and thereby alter cell morphology . Indeed , several type III translocated proteins of V . parahaemolyticus and the related pathogen V . cholerae , including the recently described VopV , have been shown to alter actin dynamics in cultured cells [51]–[55] . Furthermore , AM-19226 , a non-O1 , non-O139 V . cholerae strain that encodes a T3SS similar to the V . parahaemolyticus T3SS2 causes villus destruction in the small intestine of infant rabbits , suggesting that common effectors translocated by these systems may contribute to the pathology [27] . However , the steps leading to the destruction of intestinal villi by AM-19226 have not been elucidated , and the pathologic features of AM-19226-induced disease differ from those caused by V . parahaemolyticus . Besides damaging the villous epithelium in the small intestine , V . parahaemolyticus also causes elevated proliferation of cells in the crypts . Several other enteric pathogens , as well as the microbiota , have been reported to alter intestinal epithelial dynamics [50] . In some cases , pathogen-induced changes in epithelial homeostasis are thought to promote bacterial colonization . Similarly , since V . parahaemolyticus damages the epithelium , promoting epithelial renewal could enhance colonization . However , the increased cell proliferation occurs early during V . parahaemolyticus infection , suggesting that proliferation is not a direct response to epithelial damage . Identification of the V . parahaemolyticus factor ( s ) that lead to elevated proliferation , and the manner by which they are antagonized by TDH , may yield insight into mechanisms that normally govern turnover of intestinal stem cells . Our findings suggest that each of the three previously proposed V . parahaemolyticus virulence-linked loci – TDH and the two T3SSs – modulate the organism's pathogenicity . Our results confirm and extend earlier ileal-loop-based studies [19] , [22] indicating that T3SS2 is the major virulence factor contributing to V . parahaemolyticus' enterotoxicity . We observed that T3SS2 is not only required for intestinal fluid accumulation , but it is also essential for colonization of the small intestine . It will be interesting to explore how T3SS2 promotes colonization and investigate if one ( or more ) translocated effector ( s ) act in a similar fashion as the EPEC/EHEC translocated intimin receptor , Tir [56] , to enable V . parahaemolyticus adherence . Alternatively , do the T3SS2 effectors modulate host cell processes , including those of the immune response , to generate a niche permissive for V . parahaemolyticus proliferation ? For example , increased access to nutrients may occur as a consequence of epithelial disruption . Unexpectedly , we found that TDH negatively impacts colonization of the upper regions of the small intestine and appears to dampen some aspects of V . parahaemolyticus-induced disease . This result contrasts with findings from studies using ligated ileal loops [22] , [57] , where TDH was found to contribute to fluid accumulation . Differences between the experimental systems may explain these contradictory results; ligated ileal loops are closed systems where intestinal peristalsis is reduced and infections are of limited duration ( typically 18 hr ) . Collectively , our findings suggest that infant rabbits will be a very useful experimental model to shed light on the pathogen and host factors , and mechanisms that explain the pathogenesis of V . parahaemolyticus-induced intestinal disease . It is important to note however , that while many of the features of V . parahaemolyticus-induced disease resemble those reported in humans , rabbits do not exhibit all the signs of V . parahaemolyticus infection that have been reported . For example , infected individuals can have occult blood in their stool and occasionally present with grossly bloody stools [5] , [58] . The presence of blood in the stool appears to correlate with epithelial damage consisting of ‘superficial ulcerations’ in the lower intestine of patients [5] , [58]; no pathology was detected within the colons of infected rabbits and neither gross nor occult blood was observed in fecal material obtained from the rabbits . Nevertheless , infant rabbits reproduce the inflammatory enteritis and watery diarrhea that are the chief signs of disease in most infected individuals . Thus , studies using this model host should enable dissection of the complex interplay of pathogen and host factors that result in disease as well as testing of new therapeutics to combat and/or prevent infection .
The pandemic V . parahaemolyticus isolate RMID2210633 ( serotype O3:K6 ) was used as the wild type in this study . Derivatives of RMID2210633 containing deletions in TDH ( ΔtdhAS ) , T3SS1 ( ΔvscN1 ) , T3SS2 ( ΔvscN2 ) , TDH and T3SS1 ( aka POR-2 ( ΔtdhAS ΔvcrD1 ) ) , or all 3 virulence factors ( aka the triple mutant or POR-3 ( ΔtdhAS ΔvcrD1 ΔvcrD2 ) have been previously described [22] . A high copy number plasmid which is stably maintained in vitro and in vivo without selection [26] , [59] , was used to introduce green fluorescent protein ( GFP ) constitutively expressed from the lac promoter , into wild type V . parahaemolyticus . All strains were routinely grown in LB medium or on LB agar plates containing the appropriate antibiotics at the following concentrations: 50 µg/mL carbenicillin and 50 µg/mL spectinomycin . All animal studies were carried out in accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health ( 8th Edition ) and the Animal Welfare Act of the United States Department of Agriculture . All protocols were reviewed and approved by the Harvard Medical Area Standing Committee on Animals ( Animal Welfare Assurance of Compliance #A3431-01 ) . Litters of two-day old New Zealand White infant rabbits with the lactating doe were acquired from a commercial breeder ( Milbrook Farm , Amherst , MA ) . The following day , the infant rabbits were administered cimetidine ( 50 mg kg−1 via intraperitoneal injection; Hospira , IL ) 3 hr prior to oro-gastric inoculation with either 1×109 cfu wild type V . parahaemolyticus , or one of the isogenic mutants , or sodium bicarbonate solution ( 2 . 5 g in 100 mL; pH 9 ) using a size 4 French catheter ( Arrow International , Reading , PA ) . To prepare the inocula , cultures of bacteria grown for ∼18 hr at 30°C were harvested by centrifugation ( 5 mins 6000 g ) , and the cell pellet resuspended in sodium bicarbonate solution ( pH 9 ) to a final concentration of 2×109 cfu mL−1 . Following inoculation , the infant rabbits were monitored frequently for clinical signs of illness . Disease was scored at euthanasia as follows: no gross disease ( no adherent fecal material on fur and intestines appear normal ) , intestinal fluid ( no adherent fecal material on fur but intestines appeared red , swollen and contained fluid ) , diarrhea ( liquid fecal material stains or adheres to fur , and intestines appeared red , swollen and contained fluid ) . In most experiments , rabbits were euthanized at fixed times after infection ( i . e . 12 , 18 , 28 or 38 hr PI ) , but rabbits were euthanized prior to these time points if they appeared moribund ( categorized as ‘dead’ in Table 1 and Figure 1A ) . At necropsy , the intestinal tract from the duodenum to the rectum was removed and processed for microbiological , microscopic and histologic analyses . For some rabbits , the internal organs including the gall bladder , spleen and liver were also collected , homogenized and plated on selective media to check for systemic spread of V . parahaemolyticus . To determine fluid accumulation ratios ( FARs ) , an approx . 5 cm length of the distal small intestine was isolated from the rest of the intestine using silk ligatures . The intestinal section was weighed and then cut every 0 . 5 cm to release any luminal fluid , and the tissue pieces reweighed . The FAR was calculated as the weight of fluid divided by the weight of the drained tissue . The electrolyte and protein concentrations in serum and diarrheal fluid collected from the ceca of infected rabbits were measured on an Olympus Analyzer ( AU-2700 ) at the Brigham and Woman's Hospital clinical laboratory . The number of V . parahaemolyticus cfu in tissue samples taken from the small and large intestine , cecum and stool were determined after homogenization , serial dilution and plating on LB media containing 50 µg mL−1 carbenicillin as described previously [24] . For unknown reasons , rabbits were occasionally not colonized by the pathogen i . e . , no V . parahaemolyticus cfu were detected in any tissue sample . These rabbits ( less than 10% , regardless of strain tested ) were excluded from all further analyses . However , any rabbits that contained detectable numbers of V . parahaemolyticus cfu in at least one tissue sample were included; for these rabbits , the lower limit of detection was reported for sections where no colonies were detected at the lowest dilution plated , and this value was used in calculation of mean cfu . For routine histological analyses , tissue segments were fixed in 10% neutral buffered formalin , processed for paraffin embedding and stained with hematoxylin and eosin ( H&E ) . The slides were semi-quantitatively assessed for infiltration of inflammatory cells ( heterophils ) , cell proliferation , and tissue damage by a pathologist blinded to the origin of the tissue . Each histological parameter was evaluated on a 0–4 scale as follows: 0 ( normal ) , 1 ( mild ) , 2 ( moderate ) , 3 ( severe ) and 4 ( severe and extensive ) . Intestinal samples for quantitative real-time PCR assays were prepared as described previously [26] . The sequences for the primers used to detect rabbit IL-6 , IL-8 , TNF-α , IL-β and GADPH are available upon request . GADPH was used as a control , and all cytokine transcripts were normalized to GADPH using the ΔΔCT method as previously described [60] . Tissue samples used in immunofluorescence studies were briefly fixed in 4% paraformaldehyde and processed as described previously [24] , [26] . All sections were first blocked with 5% bovine serum albumin ( BSA ) in phosphate buffered saline ( PBS ) for 30–60 min prior to incubation with the appropriate primary antibody ( in PBS containing 0 . 5% BSA ) . The following antibodies/reagents were used: mouse anti-ZO-1 monoclonal antibody ( 1/200; #339100 , Invitrogen , CA ) , mouse anti-occludin-1 monoclonal antibody ( 1/200; #33–1500 , Invitrogen , CA ) , mouse anti-claudin-1 monoclonal antibody ( 1/500; #37–4900 , Invitrogen , CA ) , mouse anti-macrophage monoclonal antibody ( 1/200 , #MCA874GA , AbD Serotec , UK ) , or Alexa Fluor 633 phalloidin ( 1/100; A22284; Invitrogen , OR ) , usually incubated overnight at 4°C in the dark . After washing in PBS containing 0 . 5% Tween20 ( PBST ) , the slides were incubated for 1 hr with goat anti-mouse Alexa fluor 546 ( 1/200; A11030; Invitrogen , CA ) as the detection antibody . After further washing , all slides were counterstained with DAPI ( 1 µg mL−1 ) for 5 mins , rinsed in PBST and covered with Prolong Gold Antifade mounting media ( P36930 , Invitrogen , CA ) . All slides were examined for fluorescence using a Zeiss LSM510 Meta upright confocal microscope , and images were taken with the LSM510 software . To determine the zone of proliferating cells in the intestine of infected and mock infected rabbits , tissue sections were incubated with mouse anti-Ki67 monoclonal antibody ( 1/200 , #AB8191 , Abcam , MA ) overnight at 4°C , washed in PBST and incubated with goat anti-mouse Alexa fluor 546 , phalloidin Alexa fluor 633 and DAPI , and imaged as described above . Using these images , the distance from the base of the zone of actively dividing cells to the top of the zone ( the Ki-67-positive region ) , as well as to the top of the villus ( defined by DAPI staining ) , was measured . Relative proliferation was calculated as the length of the zone of proliferating cells relative to villus length . TUNEL staining was performed on sections obtained from infected and mock infected rabbits at various times after inoculation to evaluate the extent of apoptosis . Staining was performed according to the manufacturer's instructions ( In situ cell death detection kit; Roche IN ) , except that the cryopreserved tissue sections were permeabilized for 5 min and the enzyme mixture was applied for 90 min at 37°C . Biotin was used as a tracer molecule to determine the integrity of the epithelial barrier as described previously [61] , [62] . Briefly , immediately after removal of the entire intestinal tract from mock or infected rabbits , the small and large intestine were separated while maintaining their correct orientation ( ie proximal vs distal ends ) . EZ-Link Sulfo-NHS-Biotin ( Thermo Scientific , IL ) was injected slowly ( 1–2 min ) into the lumen via the open ( cut ) end of the most distal part of the small intestine or colon . After 3 min , the tissue section just proximal to the site of injection was removed , fixed in 4% paraformaldehyde and processed for immunofluorescence staining as described above . Tissue sections were incubated with streptavidin linked to Alexa 546 ( 1/500; S11225; Invitrogen , CA ) for 1 hr at room temperature , before being counterstained with phalloidin Alexa fluor 633 and DAPI as described above . Ileal and colonic tissue from the ileum and colon of control and infected rabbits that were not treated with biotin , exhibited no endogenous biotin activity when incubated with streptavidin only ( data not shown ) . Intestinal samples were prepared for scanning or transmission electron microscopy as described previously [24] , [25] . Samples were examined using a Hitashi S-4800 FESEM 2 kV scanning electron microscope and a JOEL 1200EX -80 kV transmission electron microscope . The proportion of rabbits with or without diarrhea was compared to wild type using Fisher's exact test . Fluid accumulation ratios and bacterial counts ( after log transformation ) were statistically analyzed using one way analysis of variance ( ANOVA ) and Bonferroni's test for multiple comparison . Ratios obtained from Ki67-stained samples were compared using the Student t test assuming unequal variances . Histological scores for infiltration of heterophils , cell proliferation and tissue damage were treated as non-parametric data and statistically analyzed using the Kruskal-Wallis statistic with Dunn's post-test for multiple comparisons . All statistical analyses were performed using GraphPad Prism , San Diego , CA .
|
The marine bacterium Vibrio parahaemolyticus is a leading cause worldwide of gastroenteritis linked to the consumption of contaminated seafood . Despite the prevalence of V . parahaemolyticus-induced gastroenteritis , there is limited understanding of how this pathogen causes disease in the intestine . In part , the paucity of knowledge results from the absence of an oral infection-based animal model of the human disease . We developed a simple oral infection-based infant rabbit model of V . parahaemolyticus-induced intestinal pathology and diarrhea . This experimental model enabled us to define several previously unknown but key features of the pathology elicited by this organism . We found that V . parahaemolyticus chiefly colonizes the distal small intestine and that the organism's second type III secretion system is essential for colonization . The epithelial surface of the distal small intestine is also the major site of V . parahaemolyticus-induced damage , which arises via a characteristic sequence of events culminating in the formation of V . parahaemolyticus-filled cavities in the epithelial surface . This experimental model will transform future studies aimed at deciphering the bacterial and host factors/processes that contribute to disease , as well as enable testing of new therapeutics to prevent and/or combat infection .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"infectious",
"diseases",
"gastroenterology",
"and",
"hepatology",
"biology",
"microbiology",
"bacterial",
"pathogens"
] |
2012
|
Inflammation and Disintegration of Intestinal Villi in an Experimental Model for Vibrio parahaemolyticus-Induced Diarrhea
|
Riboswitches are non-coding RNA molecules that regulate gene expression by binding to specific ligands . They are primarily found in bacteria . However , one riboswitch type , the thiamin pyrophosphate ( TPP ) riboswitch , has also been described in some plants , marine protists and fungi . We find that riboswitches are widespread in the budding yeasts ( Saccharomycotina ) , and they are most common in homologs of DUR31 , originally described as a spermidine transporter . We show that DUR31 ( an ortholog of N . crassa gene NCU01977 ) encodes a thiamin transporter in Candida species . Using an RFP/riboswitch expression system , we show that the functional elements of the riboswitch are contained within the native intron of DUR31 from Candida parapsilosis , and that the riboswitch regulates splicing in a thiamin-dependent manner when RFP is constitutively expressed . The DUR31 gene has been lost from Saccharomyces , and may have been displaced by an alternative thiamin transporter . TPP riboswitches are also present in other putative transporters in yeasts and filamentous fungi . However , they are rare in thiamin biosynthesis genes THI4 and THI5 in the Saccharomycotina , and have been lost from all genes in the sequenced species in the family Saccharomycetaceae , including S . cerevisiae .
Riboswitches are RNA regulatory elements that are located within messenger RNA , and control gene expression [1] . The most common classes bind to small molecule ligands , ranging from coenzymes , S-adenosylmethionine ( SAM ) and amino acids to metal ions [2] . Other classes respond to temperature [3] , pH [4] , and tRNA binding [5] . Binding of the ligand or changing the temperature or pH disrupts the secondary structure of the riboswitch . Riboswitches are best described in bacteria , where over 20 ligands have been identified [2 , 6] . Bacterial riboswitches are usually located in 5’ UTR regions , and control initiation of translation or premature termination of transcription . Ligand binding introduces a structural change that prevents access to the ribosome binding site , or promotes the formation of an intrinsic transcriptional terminator . In eukaryotes only one type of riboswitch has been identified , which binds thiamin pyrophosphate ( TPP , a derivative of thiamin ) . TPP riboswitches regulate expression of thiamin synthesis genes in algae and marine phytoplankton [7] , plants [8 , 9] and filamentous fungi [10 , 11] , and probably in oomycetes [12] . Eukaryotic riboswitches are often located within introns , and they function by regulating splicing . Thiamin is an enzyme cofactor that can be imported into the cell , or can be synthesized from 5- ( 2-hydroxyethyl ) -4-methylthiazole phosphate ( HET-P ) and 4-amino-5-hydroxymethyl-2-methylpyrimidine diphosphate ( HMP-PP ) . Thiamin is converted to TPP through the activity of thiamin pyrophosphokinase ( THI80 in yeast [13] ) . The filamentous fungus Neurospora crassa has three riboswitches , two of which are in introns in the 5’ UTRs of the thiamin synthesis genes THI4 ( NCU06110 ) and THI5 ( also known as NMT1 or NCU09345 ) [10 , 11] . Splicing of these introns uses at least two 5’ donor splice sites , depending on the environmental conditions . In THI5 , in the absence of TPP , a small region of the riboswitch base pairs with a complementary sequence surrounding the second donor splice site , preventing access to the splicing machinery [10] . Splicing therefore occurs at the first donor splice site , removing the entire intron , and facilitating translation of a functional protein . In the presence of TPP , the riboswitch adopts a different structure , which disrupts its interaction with the second donor splice site . Splicing at the second site is now favored , and part of the intron is retained in the mRNA . Small upstream open reading frames ( uORFs ) in the retained intron compete with the main ORF for translation . A similar process occurs in THI4 , which has a slightly more complex intron structure . Expression of N . crassa gene NCU01977 is regulated by a third TPP riboswitch , but somewhat differently to THI4 and THI5 [11] . The riboswitch in NCU01977 is in an intron within the coding sequence , not in the 5’ UTR . The intron has several potential 5’ splice sites , and only splicing at the first site produces a fully functional protein . Splicing at other sites results in the introduction of premature stop codons . In the absence of TPP , a long-range interaction between the riboswitch and a region adjacent to the first splice site increases the use of this site , possibly by looping out the intermediate donor splice sites [11] . When thiamin is present , splicing occurs at the downstream donor sites and translation stops prematurely . The logic of the switch remains the same—thiamin or TPP represses translation of NCU01977—but the mechanism is different to the thiamin biosynthesis genes . NCU01977 encodes a putative transporter . Genes with similar domains in other filamentous fungi , and in phytoplankton , also contain TPP riboswitches [7] . In marine algae , it has been hypothesized that NCU01977 orthologs may transport thiamin or thiamin intermediates such as HMP or AmMP , but experimental evidence is lacking [7 , 14] . In some algae , it has been predicted that riboswitches regulate expression of thiamin biosynthesis genes by base pairing to the branch point of the intron in the presence of TPP , preventing splicing [15] . The resulting messenger RNA contains premature stop codons . In plants , a 3’ processing site in the mRNA is removed by riboswitch controlled alternative splicing in the presence of TPP , resulting in transcripts with reduced stability [9] . Thiamin or TPP therefore represses translation of the target genes in filamentous fungi , algae and plants by controlling alternative splicing , but through many different mechanisms . Until recently , TPP riboswitches have generally been assumed to be absent from budding yeasts ( Saccharomycotina ) , although a few have been predicted bioinformatically [12 , 16–18] . We find that TPP riboswitches are common in DUR31 genes in the Saccharomycotina , and that splicing of this gene in Candida parapsilosis is regulated by thiamin . We show for the first time that DUR31 , which is an ortholog of Neurospora crassa NCU01977 , encodes a thiamin transporter . A small number of yeast species retain riboswitches in thiamin biosynthesis genes , but all TPP riboswitches have been lost from all sequenced species in the family Saccharomycetaceae , including S . cerevisiae . Riboswitch-mediated regulation of thiamin transport is therefore strongly conserved throughout fungi , including yeasts , as well as in algae and plants .
While characterizing the small RNAs transcribed in the pathogenic yeast Candida parapsilosis and its relatives [19] we noted that the RNA structure predictor software Infernal [20] identified a potential TPP riboswitch in an intron of a poorly characterized gene CPAR2_502100 , which we named DUR31 after its ortholog in Candida albicans [21 , 22] ( Fig 1A ) . Prediction of a riboswitch was surprising , because until very recently , most reports suggest that riboswitches are absent from budding yeasts [10 , 12 , 23] . We therefore tested the effect of thiamin on splicing of DUR31 in C . parapsilosis . The riboswitch in DUR31 in C . parapsilosis is in an intron near the 5’ end of the gene ( Fig 1A ) . The intron contains two potential 5’ donor splice sites , followed by the riboswitch , and a single 3’ acceptor site . The first 5’ donor site matches the C . parapsilosis 5’ donor consensus sequence ( GTATGT ) , whereas the second has a non-consensus sequence ( GTTGGA ) . In the absence of thiamin , most of the RNA is spliced at the first site , generating an mRNA ( S ) that encodes a full-length protein of 525 amino acids . In the presence of thiamin , the amount of unspliced mRNA ( U ) is increased . Some of the RNA is spliced at the second donor site ( PS ) in both the presence and absence of thiamin . The unspliced and partially spliced products do not encode a full-length protein , because of a premature stop codon after 58 amino acids , between the two 5’ splice sites . The difference in the total amount of spliced and unspliced mRNA in the presence and absence of thiamin could result from regulation of expression of DUR31 , or from regulation of splicing by the riboswitch . To explore these possibilities , we introduced the intron from C . parapsilosis into the coding sequence of a purple fluorescent protein ( yEmRFP ) gene [24] , so that it interrupts the ORF ( +intron +riboswitch ) . This modified RFP was constitutively expressed from a GAPDH promoter on a plasmid in S . cerevisiae , thus removing any effect of thiamin on the endogenous DUR31 promoter ( Fig 2A ) . When cells containing the intron plus riboswitch construct are grown in the presence of thiamin , fluorescence levels remain at a low level . In cells transferred to medium without thiamin , fluorescence levels increase with respect to growth ( Fig 2B ) . Pink/purple color is clearly higher in transformed S . cerevisiae cells grown overnight in the absence of thiamin , compared to cells grown in the presence of thiamin ( Fig 2C ) . In the presence of thiamin , most of the yEmRFP product from the intact intron is unspliced after 5 h ( U1 , Fig 2D ) , though some splicing does occur at the first donor site ( S ) . There is no evidence of splicing at the second donor site , which is seen when the intron is in its native position in C . parapsilosis ( Fig 1A ) . In the absence of thiamin , the amount of unspliced ( U1 ) mRNA is greatly reduced ( Fig 2D ) . The ratio of spliced/unspliced product is substantially different in the absence and presence of thiamin , suggesting that thiamin is regulating splicing , or regulating the stability of the unspliced product . The increase in fluorescence observed in Fig 2B in the absence of thiamin suggests that splicing , rather than stability , is regulated . The experiment also shows that the C . parapsilosis TPP riboswitch is fully functional in S . cerevisiae , even though there are no riboswitches in this species . These results suggest that all the regulatory components are present in the intron . We next made a version of the intron that maintains the donor and acceptor splice sites but does not include the riboswitch , and introduced it into yEmRFP at the same position ( +intron -riboswitch construct , ( Fig 2 ) ) . We predicted that this construct would be spliced both in the presence and absence of thiamin . However , in S . cerevisiae cells transformed with the construct , fluorescence levels approached zero , and cells are completely white ( Fig 2C ) . We used RT-PCR to show that the intron lacking a riboswitch ( +intron–riboswitch ) is not spliced from yEmRFP , even when there is no thiamin present ( U2 , Fig 2D ) . When the riboswitch is present ( +intron +riboswitch ) it acts as an “off” switch ( increased unspliced RNA and reduced expression in the presence of thiamin/TPP ) . Our results suggest that the riboswitch is also required for splicing to occur . It is possible that shortening the intron changed the accessibility of the donor and/or acceptor splice sites . However , all intron-associated features are still present in the two constructs , including the probable branch site ( TACTAAC ) . Because splicing of the DUR31 intron is regulated by thiamin , we predicted that the protein is likely to be involved in thiamin biosynthesis or transport . Dur31 is a homolog of N . crassa NCU01977 , in which splicing is also regulated by a TPP riboswitch [11] . Dur31/NCU01977 belong to the solute carrier 5 transporter family . These proteins contain SSF domains , which indicate that they act as sodium-solute symporters [25] . Dur31 is related to Dur3 in S . cerevisiae ( average identity is 12 . 6% in C . parapsilosis ) , and Dur3 homologs in both S . cerevisiae and C . albicans transport urea and polyamines [21 , 22 , 26 , 27] . In order to determine the relationship between Dur31 and the large family of related proteins in the Ascomycota , we searched for homologs using a Hidden Markov Model ( HMM ) generated using Dur31 from 14 Saccharomycotina species . The model identified both Dur3 and Dur31 homologs . By phylogenetic analysis ( Fig 3A ) , we identified several clades , including at least four that are relatively closely related to Dur3 ( Fig 3A ) . Clade I includes proteins from S . cerevisiae and C . albicans that are known to transport spermidine and other polyamines , such as Dur3 itself [21] . Clade II contains homologs of Dur4 , an uncharacterized protein in C . albicans that is assumed to be a urea transporter [28] . Clade III , which may contain two sub-clades , contains uncharacterized proteins which we have named here as Dur3-2 and Dur3-3 . Clade IV contains homologs of Dur7 , Dur32 , and Dur35 from C . albicans , again assumed to be urea transporters [29] . In contrast to clades I-IV , clade V is only distantly related to the others . Clade V contains all Dur31 orthologs , including C . parapsilosis Dur31 and N . crassa NCU01977 ( NcDur31 ) . The gene duplication that formed clade V is old , and predates the divergence between the Saccharomycotina and the Pezizomycotina lineages ( Fig 3A ) . For example , there are N . crassa ( Pezizomycotina ) genes in clades I and V , and A . nidulans genes in clades , I , II and IV ( Fig 3A ) . Mukherjee et al [12] identified orthologs of DUR31/NCU01977 in basidiomycetes and in oomycetes , some of which contain TPP riboswitches , suggesting that it is an ancient gene . However , Dur31 is completely absent from S . cerevisiae and its close relatives ( see below ) . To characterize the roles of DUR3 and DUR31 , we deleted the genes in C . parapsilosis , and we acquired equivalent deletion strains of C . albicans [21 , 22] . Deleting DUR31 , but not DUR3 , allows growth of both C . parapsilosis and C . albicans in the presence of pyrithiamine , a toxic analog of thiamin ( Fig 3B ) [30] . Only C . parapsilosis and not C . albicans DUR31 contains a riboswitch . The lack of toxicity is therefore not due to an interaction of pyrithiamine pyrophosphate ( PTPP ) with the riboswitch . The simplest interpretation is that deleting DUR31 prevents uptake of both thiamin and pyrithiamine , allowing growth in the presence of the toxic compound . Active thiamin transport has been extensively studied in S . cerevisiae , where the transporter is Thi7 [31] . THI7 has undergone a specific gene amplification in S . cerevisiae , resulting in three members; THI7 ( also called THI10 ) , THI72 , and NRT1 [31 , 32] . In S . cerevisiae Thi7 is a high affinity transporter of thiamin , and Thi72 and Nrt1 are low affinity transporters . THI72 and NRT1 are not present in other yeast species . Thi7 belongs to the Major Facilitator Superfamily ( MFS ) , which is structurally unrelated to the SSF domain family represented by Dur31 . Specifically , Thi7 belongs to a subset of MFS transporters that includes Dal4 ( allantoin permease ) , Fur4 ( uracil permease ) , and Fui1 ( uridine permease ) . Because MFS is a large family , we first used phylogenetic analysis to separate the Thi7 orthologs from related proteins ( S1 Fig ) . We then constructed a tree of these Thi7 orthologs ( Fig 4 ) . Thi7 is entirely absent from the Pezizomycotina . In the Saccharomycotina , there are Thi7 orthologs in species within the Saccharomycetaceae , the Saccharomycodaceae , and the Phaffomycetaceae , but not in other lineages such as Yarrowia and the Debaryomycetaceae /Metschnikowiaceae clades ( Fig 4 ) . Two THI7 genes were also identified in the Pichiaceae species , Brettanomyces bruxellensis and Brettanomyces anomalus ( Fig 4 ) . The Brettanomyces orthologs are more closely related to Thi7 proteins from the Hanseniaspora ( Saccharomycodaceae ) species . Riboswitches were generally assumed to be missing from budding yeasts ( Saccharomycotina ) [23] , although a very recent study has identified some in a small number of species [12] . To examine the distribution of riboswitches in yeasts , we examined 86 genomes of species from the Saccharomycotina and 10 outgroup species [33 , 34] . We used the software Infernal to predict riboswitches , together with a detailed manual analysis of associated open reading frames ( Fig 5 , S1 Data Set , see Methods ) . As expected , many of the riboswitches we found are in genes known to be involved in thiamin metabolism . Riboswitches were most commonly found in DUR31 homologs ( S2 Fig ) . TPP riboswitches are present in introns of most of the DUR31 homologs in species outside the family Saccharomycetaceae , including in many species in the Debaryomycetaceae/Metschnikowiaceae ( CUG-Ser ) clade , the Pichiaceae , and the Yarrowia clade ( Fig 5 ) . Riboswitches are also present in DUR31 in many Candida species ( as well as Candida parapsilosis ) , but are missing from the well-studied Candida albicans ( Fig 5 ) . Some of the other DUR31 orthologs in this clade contain introns near the 5’ end of the gene , but have no riboswitch ( e . g . Lodderomyces elongisporus ) . Not all of the DUR31 orthologs in the Saccharomycotina have obvious alternative donor splice sites like CPAR2_502100 . However , where a riboswitch is present , splicing , or expression , is probably controlled by thiamin . We characterized expression of DUR31 from Ogataea polymorpha ( a Pichiaceae species ) , where only one donor and one acceptor site was identified ( Fig 1B ) . A fully spliced product is present only in the absence of thiamin , and only the spliced product encodes a full-length protein ( Fig 1B ) . Thiamin therefore represses the production of a functional Dur31 protein in both C . parapsilosis and O . polymorpha by repressing production of a functional spliced isoform . Some regulation may be exerted at the level of transcriptional regulation , similar to the repression of thiamin synthesis genes in S . cerevisiae [35] . TPP riboswitches were rarely found in thiamin biosynthesis genes ( THI4 , THI5 ) in budding yeast species , unlike in Pezizomycotina [10–12 , 23] . Only two were identified in THI5 , both in basal lineages ( Geotrichum candidum and Lipomyces starkeyi ) ( Fig 5 ) . There are riboswitches in introns in THI4 in family Pichiaceae , some of the Yarrowia clade , and Ascoidea rubescens . However , riboswitches appear to have been lost from thiamin biosynthesis genes in the Debaryomycetaceae /Metschnikowiaceae . TPP riboswitches were also identified in a small number of genes that encode neither known thiamin biosynthesis enzymes nor Dur31 homologs ( Fig 5 ) . These include genes in Stagonospora nodorum , Xylona heveae , Aspergillus nidulans ( AN4526 . 2 ) , and a filamentous fungus ( incorrectly identified as Geotrichum candidum 3C [36] ) , that are predicted to encode nucleoside transporters , and share some similarities with S . cerevisiae Tpn1 , a transporter of pyridoxine ( Vitamin B6 ) , and Fcy21 , a member of the purine-cytosine permease family whose function is unknown [37] . More fungal homologs belonging to this class were identified by Moldovan et al [23] and were categorized as “putative transporters” . They are predicted to play some role in thiamin metabolism . One riboswitch-containing gene in Blastobotrys ( Arxula ) adeninivorans is a homolog of Thi9 , the thiamin transporter in Schizosaccharomyces pombe [17] . TPP riboswitches are also present in other genes in Candida apicola and Starmerella bombicola that are related to the monocarboxylate porter ( MCP ) family , part of the MFS superfamily . Finally , riboswitches were predicted in small transcripts with little obvious protein coding potential in Brettanomyces anomalus and Brettanomyces bruxellensis ( S3 Fig ) . No TPP riboswitches were predicted in any genes in Saccharomycetaceae species ( Fig 5 ) .
We found that riboswitches are common in orthologs of DUR31/ NCU01977 in budding yeasts ( Fig 5 ) and we showed that Dur31 transports thiamin in C . parapsilosis and C . albicans , a function that is likely conserved among many fungal species . Dur31 is an ancient gene , that predates the separation of the Pezizomycotina and the Saccharomycotina ( Fig 5 ) , and was probably present in the ancestor of fungi and oomycetes [12] . It has been lost from Schizosaccharomyces pombe , a member of the Taphromycotina , which lies at the base of the Saccharomycotina ( Fig 5 ) . In S . pombe , thiamin is transported by Thi9 , which is more closely related to amino acid transporters [38] . Thi9 orthologs in other Taphrinomycotina species , and in Pezizomycotina ( filamentous fungi ) and more distantly related Basidiomycetes , also contain riboswitches [23] . B . adeninivorans retains both DUR31 and THI9 , and riboswitches are present in both ( Fig 5 ) . Another thiamin transporter , Thi7 from the MFS family , is present in species within the Saccharomycetaceae , the Saccharomycodaceae and the Phaffomycetaceae , and in two Pichiaceae species ( B . bruxellensis and B . anomalus ) ( Fig 5 , Fig 4 ) . Analysis of the phylogenetic relationship of the THI7 homologs suggests that it may have arisen recently in the ancestor of the Phaffomycetaceae/Saccharomycodaceae/Saccharomycetaceae and its presence in Brettanomyces may result from Horizontal Gene Transfer ( HGT ) , most likely from a Saccharomycodaceae species ( Fig 4 ) . Dur31 has been lost from S . cerevisiae and its close relatives , and independently from other lineages including the Saccharomycodaceae , and the Zygosaccharomyces/Torulaspora branch . In S . cerevisiae Thi7 is the main transporter of thiamin [32] , and it is likely that Thi7 orthologs transport thiamin in the other species also . There have been some independent losses of Thi7 ( for example in the Eremothecium lineage , and in Cyberlindnera jadinii ) . All of the budding yeast species that lack Thi7 contain Dur31 ( Fig 5 ) . We predict that Dur31 is the major thiamin transporter in these species ( Fig 6 ) . It is not known what selective pressure drove the displacement of Dur31 by Thi7 . Many species retain both Dur31 and Thi7 ( e . g . Lachancea kluyveri ) , but only one ( Wickerhamomyces anomalus ) has both a riboswitch-containing intron in DUR31 , and THI7 . The shift from DUR31 to THI7 may therefore coincide with a move from riboswitch-mediated thiamin-dependent expression to thiamin regulation at the promoter [32] ( Fig 5 , Fig 6 ) . The exact mechanism of action of the thiamin riboswitch in C . parapsilosis Dur31 is currently unknown . All of the required regulatory sequences are contained within the intron , because thiamin-dependent splicing occurs when this region is introduced into a yEmRFP coding sequence in S . cerevisiae ( Fig 2 ) . The C . parapsilosis intron is 351 bp , whereas the intron in N . crassa is 602 bp; it is therefore unlikely that the same long-range interactions proposed for N . crassa NCU01977 occur in C . parapsilosis DUR31 [11] . However , the DUR31 riboswitch appears to be required for splicing , because when it is removed from the intron splicing does not occur . Mukherjee et al [12] suggest that in introns like this ( which they call Type III ) , access of the splicing machinery to the first and second donor sites is regulated by the riboswitch , without involving a long range interaction . We see little evidence that access to the second donor site is regulated by thiamin in C . parapsilosis DUR31 ( Fig 1A ) , and some genes have only one obvious donor site ( Fig 1B ) . However , unspliced products contain stop codons ( Fig 1 ) , and so are likely to be subject to nonsense-mediated decay [39] . The riboswitch may therefore control access to the first donor site . Moldovan et al [23] identified 5 groups of fungal genes that contain riboswitches ( although they failed to identify riboswitches in Saccharomycotina species ) . We identified the same 5 genes–THI4 , THI5 , DUR31 ( NCU01977 ) , THI9 , and a putative transporter family including A . nidulans AN4526 . 2 . We also identified a sixth group , represented by MCP in S . bombicola and C . apicola . Four of the ortholog groups have transporter domains ( DUR31 , THI9 , AN4526 . 2 and MCP ) , and the first two include members that have now been shown to transport thiamin . It is therefore likely that the second two also transport thiamin or thiamin metabolites . The AN4526 . 2 family may be restricted to species outside the Saccharomycotina , and the distribution of the MCP family is unknown . The function of the additional riboswitches in Brettanomyces species is not clear . They are not adjacent to any obvious open reading frame , though they do lie within 1 . 5 kb of a putative thiamin biosynthesis gene . In the Saccharomycotina , riboswitches are rarely found in genes encoding thiamin biosynthesis enzymes; they are present in THI5 in only two species , and they are completely absent from THI4 in the Debaryomycetaceae/Metschnikowiaceae . In addition , all TPP riboswitches have been lost from the sequenced isolates in the family Saccharomycetaceae , including S . cerevisiae . The loss of riboswitches in the thiamin synthesis genes may be associated with a switch to thiamin-dependent regulation at the level of transcriptional initiation . In S . cerevisiae , expression of thiamin synthesis genes is strongly regulated in response to thiamin levels , and requires the activity of the transcription factors THI2 , THI3 and PDC2 [35 , 40] . THI2 and THI3 are not conserved outside the Saccharomycetaceae , and the role of PDC2 orthologs in regulating thiamin-dependent expression has not been investigated in species outside this clade . The relative contribution of riboswitches versus transcription factor activity in Candida species therefore remains to be determined . Our analysis allowed us to identify loss of thiamin biosynthesis genes , as well as gain and loss of transporters and riboswitches . The alternative routes that yeast use to obtain thiamin are shown in Fig 6 . The biosynthesis of thiamin is well studied in S . cerevisiae , and involves the convergence of two separate pathways; synthesis of HET-P , which requires Thi4 , and synthesis of HMP-PP , which requires Thi5 ( Fig 6 ) . THI4 and THI5 are present in most fungi , but several species have lost both genes , including three Hanseniaspora species , and Kazachstania africana ( Fig 5 ) . These species probably cannot synthesize thiamin or thiamin precursors . This idea is supported by reports that Hanseniaspora valbyensis cannot synthesize the B vitamins biotin and pantothenate , and can only partially synthesize thiamin , niacin and pyridoxine [41] . Other Hanseniaspora and Kazachstania species also require added B vitamins , including thiamin , for growth [42 , 43] . An additional 27 species have lost either THI4 or THI5 ( Fig 5 ) . Some that are missing only THI5 ( e . g . Candida glabrata , [44] ) exhibit thiamin auxotrophy that can be complemented by supplementing with HMP [45] . These species most likely import HMP using the same transport mechanisms as thiamin ( using Thi7 or Dur31 ) . Loss of THI4 is rarer , and the only species that has lost THI4 but not THI5 is Candida sojae . It is not clear if this reflects a true gene loss , or an error in the genome assembly . In summary , we find that Dur31 is one of several types of thiamin transporters , that has been present since the divergence of oomycetes and fungi [12] . There appears to be a high turnover of thiamin transporters in fungi . There has also been a gradual loss of riboswitches in yeasts , where they are most common in DUR31 . It is likely that riboswitches also regulate expression of other thiamin transporters ( such as Thi9 , and an additional putative transporter that may be restricted to species outside the Saccharomycotina ) , but these remain to be experimentally characterized .
All strains are listed in S1 Table . S . cerevisiae BY4741 , C . albicans , C . parapsilosis , and O . polymorpha , were maintained on YPD ( 2% glucose , 2% peptone , 1% yeast extract ) and E . coli DH5α on LB ( 1% tryptone , 1% NaCl , 0 . 5% yeast extract ) or LB supplemented with 100 μg/mL ampicillin . SC-uracil agar ( 2% glucose , 2% Bacto agar , 0 . 5% ammonium sulfate , 0 . 19% YNB without amino acids or ammonium sulfate ( Formedium ) , 0 . 1926% Synthetic Complete -Uracil dropout mix ( Formedium ) ) was used to select for transformants . To investigate alternative splicing , strains were grown in SD-thiamin ( 2% glucose , 0 . 5% ammonium sulfate , 0 . 171% YNB minus thiamin ( Sunrise Science Products ) ) supplemented with 1 or 30 μM thiamin where indicated . Media for S . cerevisiae BY4741 was also supplemented with leucine ( 380 mg/L ) , methionine ( 76 mg/L ) , and histidine ( 76 mg/L ) . Pyrithiamine hydrobromide was added at a final concentration of 10 μM , and agar at 2% where indicated . 200 μL SD-Thiamin and 200 μL SD-Thiamin supplemented with 10 μM thiamin were added to triplicate wells in a round-bottom 96 well plate . 10 μL cells ( A600 of 2 ) was added to each and the plate was covered with a Breathe-Easy gas-permeable membrane ( Sigma-Aldrich ) . This was placed in a Synergy plate reader and maintained at 30°C . The A600 and fluorescence were measured at time zero , then every 20 min for 24 h with vigorous shaking . For fluorescence , excitation was measured using 590/20 nm filters , where 590 nm denotes the arithmetic mean of the wavelength at 50% of peak transmission , and 20 nm denotes the full-width at half the maximum ( FWHM ) transmission , which is the bandwidth at 50% of peak transmission . Emission was measured with 645/40 nm filters . The mean of the media-only wells was subtracted from each replicate . The error bars show the standard deviation calculated from the error of propagation ( s ( x/y ) = ( x/y ) ( sqrt ( sumsq ( s ( i ) /m ( i ) ) ) ) ) . PFAM hidden Markov models ( HMMs ) for Thi4 and Thi5/Nmt1 were obtained from the Pfam database [46] . HMMs for Dur31 or Thi7 were constructed using HMMER [47] from the relevant orthologs from a number of the species in Fig 5 ( Dur31 proteins: C . parapsilosis , C . orthopsilosis , L . elongisporus , D . hansenii , M . guilliermondii , S . passalidarum , S . stipitis , C . tenuis , C . albicans SC5314 , C . albicans WO-1 , C . dubliniensis , C . tropicalis , C . lusitaniae , O . polymorpha , O . parapolymorpha . Thi7: V . polyspora , N . dairenensis ( 2 ) , N . castellii , K . naganishii ( 2 ) , K . africana , C . glabrata , S . cerevisiae , Z . rouxii , T . delbrueckii , L . kluyveri , L . thermotolerans , L . waltii , T . blattae , T . phaffii ) . These were aligned with Muscle ( v3 . 8 . 31 ) [48] . HMMER was used to screen the proteomes of all species in Fig 5 , followed by manual inspection of phylogenetic gene trees created by RAxML [49] . ORFs were predicted using the Pico_Galaxy tool get_orfs_or_cdss . py where no proteome was available . Gene locations are listed in S1 Data Set . Genomes of 96 species were obtained from GenBank , NCBI Genome , the Candida Gene Order Browser [50] , and the Joint Genomes Institute . Riboswitches were identified using cmsearch in Infernal with the RFAM TPP Riboswitch covariance model ( RF00059 ) [20] . In species where riboswitches were predicted in regions that were not adjacent to DUR31 , THI4 , and THI5 orthologs ( S . nodorum , A . nidulans , X . heveae , G . candidum 3C , B . adeninivorans , C . apicola , S . bombicola , B . bruxellensis , and B . anomalus ) , the associated genes ( e . g . AN4526 . 2 , THI9 ) were identified by examination of the surrounding sequences . Only two of the three previously predicted N . crassa riboswitches [10 , 11] were above the threshold predicted by Infernal . We identified the third riboswitch by selecting putative riboswitch predictions that were adjacent to HMMER predictions for thiamin biosynthesis genes . Applying the same strategy to all genomes did not identify any additional riboswitches near DUR31 , THI4 , THI5 , THI9 or AN4526 . 2 homologs in any other species . The full list of riboswitches is provided in S1 Data Set . DUR3 ( CPAR2_105530 ) and DUR31 ( CPAR2_502100 ) were deleted in C . parapsilosis CPL2H1 ( leu2−/his1− ) by replacing one allele of each with HIS1 from C . dubliniensis and the second with LEU2 from C . maltosa as described in Holland et al . [51] . Approximately 500 bases were amplified upstream and downstream of DUR3 using primers CpDUR3_1/ CpDUR3_3 , and CpDUR3_4/CpDUR3_6 , and from DUR31 using primers CpDUR31_1/ CpDUR31_3 , and CpDUR31_4/CpDUR31_6 and joined to CdHIS1 and CmLEU2 by fusion PCR ( S2 Table ) . The backbone of the pRS316-GAP-yEmRFP plasmid [24] was amplified using primers RFP_Lin-1_F and RFP_Lin-1_R ( S2 Table ) . This plasmid contains a URA3 selectable marker , 2-micron origin of replication , and yEmRFP . The intron from C . parapsilosis DUR31 and a version without the riboswitch were synthesized using Gblocks ( Integrated DNA Technologies , S4 Fig ) . These were amplified using primers Cpi-1_F and Cpi-1_R ( S2 Table ) , which contain sequences that overlap with RFP_Lin-1_F and RFP_Lin-1_R . The native intron , and the intron without a riboswitch were joined to linearized pRS316-GAP-yEmRFP by gap repair in S . cerevisiae BY4741 [52] , inserting the intron after amino acid 20 of RFP , and generating pPD-yRFPcpi and pPD-yRFPcpiNR . The plasmid were rescued from S . cerevisiae by transforming into E . coli and the sequence of the intron and surrounding regions was confirmed using Sanger sequencing . Cells were first grown in YPD ( or SC-uracil for S . cerevisiae BY4741 ) overnight , diluted to an A600 of 0 . 2 in 20 mL SD-Thiamin +/- additional thiamin , and for C . parapsilosis RNA was isolated after 5 h using an Isolate II Mini kit protocol ( Bioline ) , following the manufacturers’ instructions except that 1 μg RNA in a total volume of 20 μL was treated with 1 μL DNase and 1 μL DNase buffer ( Invitrogen ) for 5 min at room temperature , followed by 1 μL DNase inactivation reagent and incubation at 65°C for 10 min . To generate cDNA , 4 μL of DNase-treated RNA and Oligo dT ( Promega , final concentration of 20 μg/mL ) was made up to 5 μL using water and incubated at 70°C for 10 min . 20 μL nuclease-free water with final concentrations of 1X MMLV-RT Buffer , 2 units/μL of RNasin , and 500 μM dNTPs were added to duplicate tubes . MMLV-RT was added to one set of duplicate tubes at a final concentration of 20 units/μL , and the same volume of water to the other set of duplicate tubes as a control to determine DNase-treatment success . The tubes were incubated at 37 C for 1 hr , followed by 2 min at 95 °C . RNA was extracted from O . polymorpha using hot-acid phenol , and SuperScript III Reverse Transcriptase ( Invitrogen ) was used for cDNA synthesis . Primers CP_TPP_F2/CP_TPP_R3 were used to characterize splicing in C . parapsilosis , HpDUR31f2/HpDUR31r1 in O . polymorpha ( Fig 1 ) , and RFPcheck_F/RFP_R in S . cerevisiae BY4741 + pPD-yRFPcpi ( S2 Table ) .
|
Thiamin , or Vitamin B1 , is an essential requirement in all living organisms because it is a co-factor for many enzymes in metabolism . Unlike animals , many yeasts can synthesize thiamin , or they can import it from the environment . Expression of thiamin biosynthesis genes and of thiamin transporters is strictly regulated in response to the presence of thiamin . In many filamentous fungi , expression of thiamin biosynthesis genes is regulated by TPP riboswitches , RNA regulatory elements that are located within messenger RNA . TPP riboswitches are rare in yeasts . However , we find that TPP riboswitches are conserved in an ancient thiamin transporter , found in filamentous fungi , yeasts and other related organisms . There appears to be a high turnover of thiamin transporters in fungi , and there has been a gradual loss of TPP riboswitches in yeasts .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"fungal",
"genetics",
"gene",
"regulation",
"pathogens",
"microbiology",
"fungi",
"model",
"organisms",
"experimental",
"organism",
"systems",
"fungal",
"pathogens",
"saccharomyces",
"research",
"and",
"analysis",
"methods",
"genome",
"complexity",
"mycology",
"medical",
"microbiology",
"gene",
"expression",
"microbial",
"pathogens",
"yeast",
"biochemistry",
"candida",
"eukaryota",
"genetics",
"biology",
"and",
"life",
"sciences",
"yeast",
"and",
"fungal",
"models",
"saccharomyces",
"cerevisiae",
"genomics",
"biosynthesis",
"computational",
"biology",
"introns",
"organisms",
"candida",
"albicans"
] |
2018
|
TPP riboswitch-dependent regulation of an ancient thiamin transporter in Candida
|
Temperature affects both the timing and outcome of animal development , but the detailed effects of temperature on the progress of early development have been poorly characterized . To determine the impact of temperature on the order and timing of events during Drosophila melanogaster embryogenesis , we used time-lapse imaging to track the progress of embryos from shortly after egg laying through hatching at seven precisely maintained temperatures between 17 . 5°C and 32 . 5°C . We employed a combination of automated and manual annotation to determine when 36 milestones occurred in each embryo . D . melanogaster embryogenesis takes 33 hours at 17 . 5°C , and accelerates with increasing temperature to a low of 16 hours at 27 . 5°C , above which embryogenesis slows slightly . Remarkably , while the total time of embryogenesis varies over two fold , the relative timing of events from cellularization through hatching is constant across temperatures . To further explore the relationship between temperature and embryogenesis , we expanded our analysis to cover ten additional Drosophila species of varying climatic origins . Six of these species , like D . melanogaster , are of tropical origin , and embryogenesis time at different temperatures was similar for them all . D . mojavensis , a sub-tropical fly , develops slower than the tropical species at lower temperatures , while D . virilis , a temperate fly , exhibits slower development at all temperatures . The alpine sister species D . persimilis and D . pseudoobscura develop as rapidly as tropical flies at cooler temperatures , but exhibit diminished acceleration above 22 . 5°C and have drastically slowed development by 30°C . Despite ranging from 13 hours for D . erecta at 30°C to 46 hours for D . virilis at 17 . 5°C , the relative timing of events from cellularization through hatching is constant across all species and temperatures examined here , suggesting the existence of a previously unrecognized timer controlling the progress of embryogenesis that has been tuned by natural selection as each species diverges .
It has long been known that Drosophila , like most poikilotherms , develops faster at higher temperatures , with embryonic [1] , larval [1] , [2] , and pupal stages [3] , [4] , as well as total lifespan [5] , [6] showing similar logarithmic trends . While genetics , ecology , and evolution of this trait have been investigated for over a century [2] , [7]–[17] , the effects of temperature on the order and relative timing of developmental events , especially within embryogenesis , are poorly understood . We became interested in the relationship between species , temperature , and the cadence of embryogenesis for practical reasons . Several years ago , we initiated experiments looking at the genome-wide binding of transcription factors in the embryos of divergent Drosophila species: D . melanogaster , D . pseudoobscura , and D . virilis . With transcription factor binding a highly dynamic process , we tried to match both the conditions ( especially temperature , which we believed would affect transcription factor binding ) in which embryos were collected and the developmental stages we analyzed . However , our initial attempts to collect D . pseudoobscura embryos at 25°C — the temperature at which we collect D . melanogaster — were unsuccessful , with large numbers of embryos failing to develop , likely a consequence of D . pseudoobscura's alpine origin . While D . virilis lays readily at 25°C , we found that their embryos develop more slowly than D . melanogaster , complicating the collection of developmental stage-matched samples . Having encountered such challenges with just three species , and planning to expand to many more , we were faced with several important questions . Given that embryogenesis occurs at different rates in different species [8] , [18] , how should we time collections to get the same mix of stages we get from our standard 2 . 5–3 . 5 hour collections in D . melanogaster , or any other stage we study in the future ? Is it better to compare embryos collected at the same temperature even if it is not optimal for , or even excludes , some species; or , should we collect embryos from each species at their optimal temperature , if such a thing exists ? Should we select a temperature for each species so that they all develop with a similar velocity ? Or should we find a set of species that develop at the same speed at a common temperature ? And even if we could match the overall rate of development , would heterochronic effects mean that we could not get an identical mix of stages ? We found a woeful lack in the kind of data needed to answer these questions . Powsner precisely measured the effect of temperature on the total duration of embryogenesis in D . melanogaster [1] , and Markow made similar measurements for other Drosophila species at a fixed temperature ( 24°C ) [18] , but the precise timing of events within embryogenesis had been described only for D . melanogaster at 25°C [19] , [20] . The work described here was born to address this deficiency . We used a combination of precise temperature control , time-lapse imaging , and careful annotation to catalog the effects of a wide range of temperatures on embryonic development in 11 Drosophila species from diverse climates . We focused on species with published genome sequences [21] ( Table 1 ) , as these are now preferentially used for comparative and evolutionary studies . Of the species we studied D . melanogaster , D . ananassae , D . erecta , D . sechellia , D . simulans , D . willistoni , and D . yakuba are all native to the tropics , though D . melanogaster , D . ananassae , and D . simulans have spread recently to become increasingly cosmopolitan [17] . D . mojavensis is a sub-tropical species , while D . virilis is a temperate species that has become holarctic and D . persimilis and D . pseudoobscura are alpine species ( Figure 1A ) .
We used automated , time-lapse imaging to track the development of embryos held at a constant and precise temperature from early embryogenesis ( pre-cellularization ) to hatching . We maintained the temperature at 0 . 1°C using thermoelectric Peltier heat pumps . Different sets of embryos were analyzed at temperatures ranging from 17 . 5°C to 32 . 5°C , in 2 . 5°C increments . Images were taken every one to five minutes , depending on the total time of development . A minimum of four embryos from each species were imaged for each temperature , for a total of 77 conditions . In total , time-lapse image series were collected and analyzed from over 1000 individual embryos . We encountered , and solved , several challenges in designing the experimental setup , including providing the embryos with sufficient oxygen [22] , [23] and humidity . We found that glass slides were problematic due to a lack of oxygenation and led to a 28% increase in developmental time , so we instead employed an oxygen-permeable tissue culture membrane , mounted on a copper plate to maintain thermal conduction . At higher temperatures , we found that the embryos dehydrated , so humidifiers were used to increase ambient humidity . Detailed photos of the apparatus and descriptions can be found in Figure S1 . We used a series of simple computational transformations ( implemented in Matlab ) to orient each embryo , correct for shifting focus , and adjust the brightness and contrast of the images , creating a time-lapse movie for each embryo . We manually examined images from 60 time-lapse series in D . melanogaster and identified 36 distinct developmental stages [19] , [20] that could be recognized in our movies ( Table 2 , http://www . youtube . com/watch ? v=dYSrXK3o86I and http://www . youtube . com/watch ? v=QKVmRy3dDR0 or “D . melanogaster with labelled stages” and “D . melanogaster with labelled stages at reduced framerate” in DOI:10 . 5061/dryad . s0p50” ) . Due to the volume of images collected , we implemented a semi-automated system to annotate our entire movie collection . Briefly , images from matching stages in manually annotated D . melanogaster movies were averaged to generate composite reference images for each stage ( Figure 2 ) . We then used a Matlab script to find the image-matrix correlation between each of these composite reference images to the images in each time-lapse to estimate the timing of each morphological stage via the local correlation maximum ( Figure S2A ) . Of the 36 events , the eight most unambiguous events ( Figure S3 ) , identifiable regardless of embryo orientation , were selected for refinement and further analysis ( pole bud appears , membrane reaches yolk , pole cell invagination , amnioproctodeal invagination , amnioserosa exposed , clypeolabrum retracts , heart-shaped midgut , and trachea fill ) ( Figure S2B , C ) . Using a Python-scripted graphical user interface , each of the eight events in every movie was manually examined and the algorithm prediction adjusted when necessary . Timing of hatching was excluded from these nine primary events because it was highly variable , likely due to the assay conditions following dechorionation , and suitable only as an indication of successful development , not as a reliable and reproducible time point . The “membrane reaches yolk stage” was used throughout as a zero point due to the precision with which the stage could be identified in all species and from all orientations . Links to representative time-lapse videos are provided in Table 3 . As expected , the total time of embryogenesis of D . melanogaster had a very strong dependence on temperature ( Figure 3 , http://www . youtube . com/watch ? v=-yrs4DcFFF0 or “D . melanogaster at 7 temperatures” in DOI:10 . 5061/dryad . s0p50 ) . From 17 . 5°C to 27 . 5°C , there was a two-fold acceleration in developmental rate , matching the previously observed doubling of total lifespan with a 10°C change in temperature [6] . The velocity of embryogenesis at 30°C is roughly the same as at 27 . 5°C , and is appreciably slower at 32 . 5°C , likely due to heat stress . At 35°C , successful development becomes extremely rare . To examine how these temperature-induced shifts in the total time of embryogenesis were reflected in the relative timing of individual events , we rescaled the time series data for each embryo so that the time from our most reliable early landmark ( the end of cellularization ) to our most reliable late landmark ( trachea filling ) was identical , and examined where each of the remaining landmarks fell ( Figure 3C ) . We were surprised to find that D . melanogaster exhibited no major changes in its proportional developmental time under any of the non-stressful temperature conditions tested . Therefore , at least as far as most visually evident morphological features go , embryogenesis scales uniformly across a two-fold range of total time . When the embryos were under heat stress ( 30°C ) , we observed a very slight contraction in the proportion of time between early development ( pole bud appears ) to the end of cellularization ( membrane reaches yolk ) , and a slight contraction between the end of cellularization and mid-germ band retraction ( amnioserosa exposure ) . In each of the ten additional Drosophila species we examined we observed all of the 36 developmental landmarks we identified in D . melanogaster in the same temporal order ( Figure 4A ) . However , there was marked interspecies variation in both the total time of embryogenesis at a given temperature ( Figure 4B–E , Table 3 ) and the way embryogenesis time varied with temperature ( Figure 5 ) . When we examined the 10 remaining species , we found not only that the relative timing of events was constant across temperature within a species , as observed in D . melanogaster , but that landmarks occurred at the same relative time between species at all non-stressful temperatures ( Figures 6 , Table 4 ) . Between 17 . 5°C and 27 . 5°C the total developmental time for all species can be approximated relatively accurately by an exponential regression ( ) . For all species we find that temperature T can be related to developmental time , agreeing with a long history of temperature-dependent rate modeling [24]:and developmental rate v:The parameters of these relations for each species , which includes two independent coefficients , are included in Table 5 . Also included in Table 5 is the , an empirical description of biological rate change from a 10°C temperature change , for the 17 . 5°C to 27 . 5°C interval . At higher temperatures , heat stress appears to counter the logarithmic trend and lengthens developmental time . Since the temperature responses are highly reproducible , the developmental time for each species can be modeled and predictions made for future experiments ( Figure S4 ) . Seven of the eleven species we examined were of tropical origin , with only two alpine , one subtropical and one temperature species . At mid-range temperatures ( 22 . 5°C–27 . 5°C ) , the tropical species developed the fastest , followed by the subtropical D . mojavensis , the alpine D . pseudoobscura and D . persimilis , and the temperate D . virilis ( Figure 5 ) , in accord with [18] . Some tropical species have expanded into temperature zones and a variety of wild strains have been collected from a variety of climates . We examined nine additional strains of D . melanogaster collected along the eastern United States [25] , [26] . Though collected along a tropical to temperate cline and there was some variation between strains , no trends were seen ( Figure S5A , B ) . The tropical species all showed highly similar responses to temperature , even though they originate from different continents ( Africa , Asia and South America ) and are not closely related ( five of the species are in the melanogaster subgroup , but D . ananassae and D . willistoni are highly diverged from both D . melanogaster and each other ) . Though they possess similar temperature-responses , these species possess significantly different and independent temperature response curves ( ) and the differences are large enough to be relevant for precise developmental experiments . These cross-species differences tend to be , but are not necessarily , larger than those seen between D . melanogaster strains ( Figure S5C ) . The embryogenesis rate for these species increases rapidly with temperature ( ) before slowing down at and above 30°C ( Figure S6A–F , http://www . youtube . com/watch ? v=vy6L4fmWkso or “D . ananassae at 7 temperatures” in DOI:10 . 5061/dryad . s0p50 ) . The two closely related alpine species ( D . pseudoobscura and D . persimilis ) match the embryogenesis rate of the tropical species at 17 . 5°C , but accelerate far less rapidly with increasing temperature ( ) , especially at 25°C and above ( Figure S6I , J , http://www . youtube . com/watch ? v=sYi-FUXpv4Q or “D . pseudoobscura at 6 temperatures” in DOI:10 . 5061/dryad . s0p50 ) . These species also show a sharp increase in embryogenesis rate and low viability above 27 . 5°C , consistent with their cooler habitat . The subtropical D . mojavensis ( Figure S6H , http://www . youtube . com/watch ? v=XWMs4oUx_mU or “D . mojavensis at 6 temperatures” in DOI:10 . 5061/dryad . s0p50 ) and temperate D . virilis ( Figure S6G , http://www . youtube . com/watch ? v=eyr4ckDb0kM or “D . virilis at 6 temperatures” in DOI:10 . 5061/dryad . s0p50 ) both develop very slowly at low temperature , but accelerate rapidly as temperature increases ( of and respectively ) . D . virilis remains the slowest species up to 30°C , while D . mojavensis is as fast as the tropical species at high temperatures . These species are both members of the virilis-repleta radiation and it remains to be seen if this growth response is characteristic of the group as a whole , independent of climate . Under heat-stress , the proportionality of development is disrupted in some embryos ( Figure S7A ) . The effect is not uniform , as some embryos developed proportionally under heat-stress and others exhibited significant aberrations , largely focused in post-germband shortening stages . This can be most clearly seen in individuals of D . ananassae , D . mojavensis , D . persimilis , and D . pseudoobscura . We did not identify any particular stage as causing this delay , but rather it appears to reflect a uniform slowing of development . Early heat shock significantly disrupts development enough to noticeably affect morphology in yolk contraction , cellularization , and gastrulation ( Figure S7B ) . Syncytial animals are the most sensitive to heat-shock ( Figure S7C ) . In D . melanogaster and several other species we observed a slight contraction of proportional developmental time between early development ( pole bud appears ) and the end of cellularization ( membrane reaches yolk ) under heat-stress ( 30°C , Figure S7D ) . While all later stages following cellularization maintain their proportionality even at very high temperatures , the pre-cellularization stages take proportionally less and less time . This indicates that at higher temperatures , some pre-cellularization kinetics scale independently of later stages , possibly leading to mortality as the temperature becomes more extreme .
In carrying out this survey , we were surprised to find that the relative timing of landmark events in Drosophila embryogenesis is constant across greater than three-fold changes in total time , spanning 15°C and over 100 million years of independent evolution . And the fact that the same holds true for 34 developmental landmarks at two temperatures in the zebrafish Danio rerio [27] , ( the only other species for which we were able to locate similar data ) , suggests that this phenomenon may have some generality . But why is this so ? Drosophila development involves a diverse set of cellular processes including proliferation , growth , apoptosis , migration , polarization , differentiation , and tissue formation . One might expect ( we certainly did ) these different processes to scale independently with temperature , much as different chemical reactions do , and as a result , different stages of embryogenesis or parts of the developing embryo would scale differentially with temperature . But this is not the case . The simplest explanation for this observation is that a single shared mechanism controls timing across embryogenesis throughout the genus Drosophila . But what could such a mechanism be ? One possibility is that there is an actual clock — some molecule or set of molecules whose abundance or activity progresses in a clocklike manner across embryogenesis and is read out to trigger the myriad different processes that occur in the transition from a fertilized egg to a larvae . However there is no direct evidence that such a clock exists ( although we note that there is a pulse of ecdysone during embryogenesis with possible morphological functions [28] , [29] ) . A more likely explanation is that there is a common rate limiting process throughout embryogenesis . Our data are largely silent on what this could be , but we know from other experiments that it is cell , or at least locally , autonomous [30]–[32] and would have to limit processes like migration that do not require cell division ( we also note that cell division has been excluded as a possibility in zebrafish [32] ) . However , energy production , yolk utilization , transcription or protein synthesis are reasonable possibilities . Although there are very few comparisons of the relative timing of events during development , it has long been noted that various measurements of developmental timing scale exponentially with [1] , [5] , [6] , [24] , [33] , but no good explanation for this phenomenon has been uncovered . Perhaps development is more generally limited by something that scales exponentially with , like metabolic rate , which , we note , has been implicated numerous times in lifespan , which is , in some ways , a measure of developmental timing . Gillooly and co-workers , noting the there was a relationship between metabolic rate , temperature and animal size , have proposed a model that incorporates mass into the Arrhenius equation to explain the relationship between these factors in species from across the tree of life [34] , [35] . We , however , do not find that mass can explain the differences in temperature-dependence between species . Even closely-related species , with nearly 2-fold differences in their mass ( e . g . D . melanogaster , D . simulans , D . sechellia , D . yakuba , and D . erecta ) , have significant divergence in their proportionality coefficients that do not converge at all when correcting for differences in mass through the one quarter power scaling proposed by Gillooly , et al . This suggests that some other factor is responsible for the differences , as has been argued by other groups [18] , [36] , [37] . The relationship between climate and temperature response raises the possibility that whatever this factor is has been subject to selection to tune the temperature response to each species' climate . However , without additional data this is purely a hypothesis . Although a common rate-limiting step is simplest explanation for uniform scaling , it is certainly not the only one . It is possible that different rate limiting steps or other processes control developmental velocity at different times or in different parts of the embryo , and that they scale identically with temperature either coincidentally , or as the result of selection ( it is important to remember that , as per Arrhenius , one does not expect different reactions to scale identically with temperature ) . If this is the result of selection , what is the selection pressure ? Evolutionary developmental biologists , perhaps most notably Stephen J . Gould , have long written about how changes in either the absolute or relative timing of different events during development have had significant effects on morphology throughout animal evolution [38]–[41] . Perhaps this is also true for fly embryogenesis , but that any such changes in morphology are selectively disadvantageous and have been strongly selected against . It is also likely that many developing fly embryos experience significant changes in temperature while developing , so there may be strong selection to maintain uniform development across temperature to ensure normal progression while the temperature is changing . Finally , we note that there are limits to this uniformity . At extreme temperatures , especially high ones , things no longer scale uniformly , likely reflecting the differential negative effects of high temperature at different stages of embryogenesis as well as the differential ability of the embryo to compensate for them . There are also clearly checkpoints in place that , while not triggered during normal embryogenesis , are important in extreme or unusual circumstances . Most strikingly , when Lucchetta et al . and Niemuth et al . examined embryos developing in chambers that allowed for independent temperature control of the anterior and posterior portions of the embryo , the two parts of the embryo developed at different velocities for much of embryogenesis [30] , [31] . They found that embryos are robust to asynchrony in timing across the embryo , though there are critical periods that , once passed , do not permit re-synchronization of development [30] , hinting at some specific checkpoints or feedback . The clustering of developmental timing and its temperature response with climate — especially amongst tropical species from different continents and parts of the Drosophila tree — suggests that this is an adaptive , or in some cases permissive , phenotype , although with only 11 species and poor coverage of non-tropical species this has to remain highly speculative . There are necessarily additional components to the temperature response , as significant variation exists within the tropical species and between D . melanogaster strains . The virilis-repleta radiation , which includes both D . virilis and D . mojavensis may have a climate-independent adaptation that leads to slowed development at cooler temperatures , a feature that is hard to rationalize . The poor response of the alpine D . pseudoobscura and D . persimilis to high temperature is consistent with their cool climate . Nevertheless , little is known about when and where most of these species lay their eggs and their natural microclimates . The clustering of developmental responses in species by their native climates rather than their climates of collection suggests that if climate adaptation is a contributing factor , the response arises slowly or rarely . The tested strains of D . melanogaster were collected in temperate , subtropical , and tropical climates and the D . simulans strain was collected in a sub-tropical climate . Nevertheless , both species performed qualitatively like other tropical species and unlike native species collected nearby . This suggests that temperature responses are neither rapidly evolving ( with D . melanogaster being present in the temperate United States for over 130 years [42] ) nor primed for change in tropical species .
Drosophila strains were reared and maintained on standard fly media at 25°C , except for D . persimilis and D . pseudoobscura which were reared and maintained at 22°C . D . melanogaster lines were raised at 18°C and 22°C for several years and their temperature response profiles were observed , verifying that transferring embryos from the ambient growth temperature for a line to the experimental temperature did not lead to heat-shock responses and had relatively little impact on the temperature response ( Figure S8A , B ) . Egg-lays were performed in medium cages on 10 cm molasses plates for 1 hour at 25°C after pre-clearing for all species except D . persimilis , which layed at 22°C . Comparisons to D . melanogaster raised and laying at 22°C confirmed that growth at lower temperatures does not account for all of the differences between the tropical and alpine species ( Figure S8C ) . To encourage egg-lay , cornmeal food media was added to plates for D . sechellia and pickled cactus was added to plates for D . mojavensis . Embryos were collected and dechorionated with fresh 50% bleach solution ( 3% hypochlorite final ) for 45 to 90 seconds ( based on the species ) in preparation for imaging . Dechorionation timing was selected as the time it took for 90% of the eggs to be successfully dechorionated . This prevented excess bleaching , as many species , such as D . mojavensis , are more sensitive than D . melanogaster . Strains used were D . melanogaster , OreR , DGRP R303 , DGRP R324 , DGRP R379 , DGRP R380 , DGRP R437 , DGRP R705 , Schmidt Ln6-3 , Schmidt 12BME10-24 , and Schmidt 13FSP11-5; D . pseudoobscura , 14011-0121 . 94 , MV2-25; D . virilis , 15010-1051 . 87 , McAllister V46; D . yakuba , 14021-0261 . 01 , Begun Tai18E2; D . persimilis , 14011-0111 . 49 , ( Machado ) MSH3; D . simulans , 14021-0251 . 195 , ( Begun ) simw501; D . erecta , 14021-0224 . 01 , ( TSC ) ; D . mojavensis wrigleyi , 15081-1352 . 22 , ( Reed ) CI 12 IB-4 g8; D . sechellia , 14021-0248 . 25 , ( Jones ) Robertson 3C; D . willistoni , 14030-0811 . 24 , Powell Gd-H4-1; D . ananassae , 14024-0371 . 13 , Matsuda ( AABBg1 ) . Embryos were placed on oxygen-permeable film ( lumox , Greiner Bio-one ) , affixed with dried heptane glue and then covered with Halocarbon 700 oil ( Sigma ) [43] . The lumox film was suspended on a copper plate that was temperature-regulated with two peltier plates controlled by an H-bridge temperature controller ( McShane Inc . , 5R7-570 ) with a thermistor feedback , accurate to 0 . 1°C . Time-lapse imaging with bright field transmitted light was performed on a Leica M205 FA dissecting microscope with a Leica DFC310 FX camera using the Leica Advanced Imaging Software ( LAS AF ) platform . Greyscale images were saved from pre-cellularization to hatch . Images were saved every one to five minutes , depending on the temperature . A humidifier was used to mitigate fluctuations in ambient humidity , though fluctuations did not affect developmental rate . Due to fluctuations in ambient temperature and humidity , the focal plane through the halocarbon oil varied significantly . Therefore , z-stacks were generated for each time-lapse and the most in-focus plane at each time was computationally determined for each image using an algorithm ( implemented in Matlab ) through image autocorrelation [44] , [45] . Time-lapse videos available from Dryad Digital Repository: doi:10 . 5061/dryad . s0p50 . A subset of time-lapses in D . melanogaster were analyzed to obtain a series of representative images for each of the 36 morphological events , selected as all events defined by [19] , [46] that were reproducibly identifiable under our conditions , described . These images were sorted based on embryo orientation and superimposed to generate composite reference images . Images from each time-lapse to be analyzed were manually screened to determine the time when the membrane reaches the yolk , the time of trachea filling , and the orientation of the embryo ( Figure S3 . This information was fed into a Matlab script , along with the time-lapse images and the set of 34 composite reference images , to estimate the time of 34 morphological events during embryogenesis via image correlation . The same D . melanogaster reference images were used for all species for consistency . A correlation score was generated for each frame of the time-lapse . The running score was then smoothed ( Savitzky-Golay smoothing filter ) and the expected time window was analyzed for local maxima . The error in event calling for the computer is very large ( greater than what we see for the overall spread across individuals of a single species at a given temperature ) , necessitating manual verification or correction of events . Many of these errors are due to aberrations in the image that confuse the computer but would not confuse a person . This results in a few bad images having a very negative effect of the overall accuracy of the computer analysis , but permits a significant improvement with just a little user input . The error in manual calls is very small compared to the variation between individuals . Computer-aided estimates were individually verified or corrected using a python GUI for all included data . Statistical significance of event timing was determined by t-test with Bonferonni multiple testing corrections . Median correction to remove outliers was used in determining the mean and standard deviation of each developmental event . Least-squares fitting was used to determine the linear approximation of log-corrected developmental time for each species . Python and Matlab scripts used in the data analysis are available at github . com/sgkuntz/TimeLapseCode . git .
|
Temperature profoundly impacts the rate of development of “cold-blooded” animals , which proceeds far faster when it is warm . There is , however , no universal relationship . Closely related species can develop at markedly different speeds at the same temperature . This creates a major challenge when comparing development among species , as it is unclear whether they should be compared at the same temperature or under different conditions to maintain the same developmental rate . Facing this challenge while working with flies ( Drosophila species ) , we found there was little data to inform this decision . So , using time-lapse imaging , precise temperature-control , and computational and manual video-analysis , we tracked the complex process of embryogenesis in 11 species at seven different temperatures . There was over a three-fold difference in developmental rate between the fastest species at its fastest temperature and the slowest species at its slowest temperature . However , our finding that the timing of events within development all scaled uniformly across species and temperatures astonished us . This is good news for developmental biologists , since we can induce species to develop nearly identically by growing them at different temperatures . But it also means flies must possess some unknown clock-like molecular mechanism driving embryogenesis forward .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"invertebrates",
"growth",
"control",
"evolutionary",
"ecology",
"ecology",
"and",
"environmental",
"sciences",
"animals",
"animal",
"models",
"developmental",
"biology",
"drosophila",
"melanogaster",
"model",
"organisms",
"organism",
"development",
"evolutionary",
"adaptation",
"morphogenesis",
"drosophila",
"research",
"and",
"analysis",
"methods",
"organogenesis",
"insects",
"arthropoda",
"ecology",
"biology",
"and",
"life",
"sciences",
"evolutionary",
"biology",
"evolutionary",
"processes",
"organisms",
"evolutionary",
"developmental",
"biology"
] |
2014
|
Drosophila Embryogenesis Scales Uniformly across Temperature in Developmentally Diverse Species
|
Hemorrhagic fever viruses , including the filoviruses ( Ebola and Marburg ) and arenaviruses ( Lassa and Junín viruses ) , are serious human pathogens for which there are currently no FDA approved therapeutics or vaccines . Importantly , transmission of these viruses , and specifically late steps of budding , critically depend upon host cell machinery . Consequently , strategies which target these mechanisms represent potential targets for broad spectrum host oriented therapeutics . An important cellular signal implicated previously in EBOV budding is calcium . Indeed , host cell calcium signals are increasingly being recognized to play a role in steps of entry , replication , and transmission for a range of viruses , but if and how filoviruses and arenaviruses mobilize calcium and the precise stage of virus transmission regulated by calcium have not been defined . Here we demonstrate that expression of matrix proteins from both filoviruses and arenaviruses triggers an increase in host cytoplasmic Ca2+ concentration by a mechanism that requires host Orai1 channels . Furthermore , we demonstrate that Orai1 regulates both VLP and infectious filovirus and arenavirus production and spread . Notably , suppression of the protein that triggers Orai activation ( Stromal Interaction Molecule 1 , STIM1 ) and genetic inactivation or pharmacological blockade of Orai1 channels inhibits VLP and infectious virus egress . These findings are highly significant as they expand our understanding of host mechanisms that may broadly control enveloped RNA virus budding , and they establish Orai and STIM1 as novel targets for broad-spectrum host-oriented therapeutics to combat these emerging BSL-4 pathogens and potentially other enveloped RNA viruses that bud via similar mechanisms .
There is an urgent and unmet need for safe and effective therapeutics against high priority pathogens , including filoviruses ( Ebola and Marburg ) and arenaviruses ( Lassa fever and Junín ) , which can cause fatal infections in humans . We and others have established that enveloped RNA viruses , including hemorrhagic fever viruses , exhibit a common requirement for host pathways , most notably ESCRT pathway functions , for efficient budding [1–7] . Indeed as host dependent budding mechanisms are highly conserved within and sometimes across virus families , they represent innovative and immutable antiviral targets for inhibiting virus transmission and disease progression [8–11] . Importantly , high mutation rates of RNA viruses in general are a factor in their ability to develop resistance to therapeutics that target specific viral proteins or functions [3 , 12–23] . Consequently , strategies that target specific host mechanisms required by viruses should reduce the development of resistance . As a number of these host mechanisms , including steps in ESCRT protein function , are targets of calcium regulation , the focus of this study was to determine whether and how hemorrhagic fever viruses mobilize calcium in host cells and whether calcium so mobilized regulates virus budding . Here we reveal a novel and fundamental requirement for host STIM1- and Orai-mediated Ca2+ entry that regulates late steps of filovirus and arenavirus egress from mammalian cells . Orai activation is typically linked to either tyrosine kinase or G-protein coupled receptors that activate phospholipase C ( PLC ) and generate diacylglycerol and inositol 1 , 4 , 5-triphoshate ( IP3 ) from membrane phospholipids . IP3 activates receptor/channels on the endoplasmic reticulum ( ER ) to allow Ca2+ to exit from the ER . The subsequent drop in ER Ca2+ below the KD ( 400–600μM , [24] ) for the N-terminal EF hands of the ER membrane-resident protein STIM1 initiates a conformational change that promotes STIM1 oligomerization and localization to ER regions adjacent to the plasma membrane . At the plasma membrane , STIM1 interacts with and activates Calcium-Release Activated Calcium ( CRAC ) channels through which extracellular Ca2+ enters the cell ( reviewed in [25] ) . CRAC channels are encoded by the Orai family of proteins ( Orai1 , 2 , & 3; [26–28] ) that provide a pathway for sustained extracellular Ca2+ entry to regulate a range of cell functions including gene expression , subcellular trafficking , and the regulation of cell shape and motility [29–31] . Herein , we demonstrate that both filovirus ( VP40 ) and arenavirus ( Z ) matrix proteins trigger Orai dependent Ca2+ entry in mammalian cells . In addition , suppression of STIM1 expression and genetic inactivation or pharmacological blockade of Orai inhibits Ebolavirus ( EBOV ) , Marburgvirus ( MARV ) , Lassa Virus ( LASV ) , and Junín Virus ( JUNV ) VLP and infectious virion production and transmission in cell culture . Together , these data establish a novel and critical role for STIM1- and Orai-mediated Ca2+ entry in late steps of hemorrhagic fever virus egress and establish STIM1 and Orai inhibitors as potential broad-spectrum anti-viral targets for regulation of these and possibly other enveloped RNA viruses that bud by similar mechanisms .
While we previously implicated Ca2+ in EBOV VP40-dependent VLP generation [32] our initial objective here was to understand if and how hemorrhagic fever virus matrix proteins trigger a change in cytosolic calcium in host cells . To do this we measured intracellular calcium in cells during an extended time course of EBOV and MARV VP40 ( eVP40 and mVP40 , respectively ) and JUNV Z matrix protein-mediated VLP production . Calcium levels ( R-GECO-1 fluorescence , [33] ) measured in HEK293T cells under physiological conditions for 18–24 hours revealed that eVP40 , mVP40 , and JUNV Z protein expression each induced a time-dependent increase in Ca2+ concentration ( Fig 1 , blue ) , while the GFP-vector backbone induced a negligible Ca2+ increase that plateaued at a low amplitude or declined to baseline levels ( Fig 1 , magenta ) . To probe the role of Orai1 in these responses we performed identical measurements in an HEK293T line that stably expresses a dominant negative mutant Orai1 having a glutamic acid ( E ) to alanine ( A ) substitution in its ion selectivity filter ( E106A ) . Incorporation of even a single Orai1 E106A subunit into endogenous WT Orai channels exerts a dominant negative block of its Ca2+ permeation [34] . Importantly , both WT and E106A Orai HEK293T cells exhibited a similar transient Ca2+ elevation following treatment with the membrane permeant SERCA pump inhibitor thapsigargin in Ca2+ free bath solution , indicating that ER stores were intact in E106A Orai1 expressing HEK293T ( S1 Fig ) . The absence of a secondary increase in cytoplasmic Ca2+ ( S1 Fig , left panel ) following reperfusion of HEK293T Orai1 E106A cells with Ca2+-containing Ringers solution ( S1 Fig , right panel ) verified the Orai permeation defect of this line . Significantly , in permeation defective Orai1 E106A cells neither eVP40 , mVP40 , JUNV Z protein ( Fig 1 , yellow ) , nor GFP vector ( Fig 1 , orange ) triggered any change in cytoplasmic Ca2+ levels indicating that Ca2+ elevations initiated by expression of EBOV , MARV , and JUNV matrix proteins required and resulted from Ca2+ entry though Orai channels . Consistent with these results from Orai E106A HEK293T cells and specifically the role of Orai , the Orai inhibitor Synta66 similarly blocked the eVP40-mediated increase in cytoplasmic Ca2+ ( Fig 1 , lower right panel ) in WT HEK293T cells . Given the ability of EBOV , MARV , and JUNV matrix proteins to initiate an Orai-dependent Ca2+ signal in HEK293T cells , we assessed the role of Orai1-mediated calcium signals in eVP40 , mVP40 , LASV Z or JUNV Z mediated VLP production in WT and Orai1 E106A HEK293T cells . Consistent with a role for Ca2+ entry via Orai1 in VLP production , we found that Orai1 E106A cells did not support efficient filovirus or arenavirus VLP production ( Fig 2 ) . Indeed , levels of eVP40 VLPs from Orai1 E106A cells were ~50 fold lower than that from WT cells ( Fig 2A , VLPs ) . Similarly , production of mVP40 VLPs exhibited an even greater dependence on Orai1-mediated calcium entry ( Fig 2B , VLPs ) , as mVP40 VLPs from Orai1 E106A HEK293T cells were ~100 fold lower than that from WT cells ( Fig 2B ) . Orai similarly regulated JUNV Z ( Fig 2C ) and LASV Z ( Fig 2D ) VLP production as both JUNV Z and LASV Z protein-mediated VLP production from E106A cells was ~100 fold lower than that from WT HEK293T cells . In all instances , cellular levels of VP40 or Z were similar in WT and E106A cells , indicating no general requirement for Orai1-mediated Ca2+ entry in viral protein expression ( Fig 2A–2D; Cells ) . Together , these data point to a conserved and selective role for Orai-mediated Ca2+ entry in hemorrhagic fever virus budding . Implicit in this common critical role for Orai1-mediated Ca2+ entry in EBOV , MARV , JUNV , and LASV VLP production is an upstream requirement for STIM1 , the only known trigger for Orai activation in mammalian cells . STIM1 is a single pass ER membrane protein whose activity is regulated by ER Ca2+ binding . Ca2+ dissociation from STIM1 following a decrease in ER concentration triggers a N-terminal conformational change that initiates its multimerization and relocalization within the ER membrane to domains juxtaposed to the plasma membrane [35–37] . The resulting subplasmalemmal STIM1 clusters physically activate Orai channels to allow extracellular Ca2+ to enter the cell [25] . Using eVP40 VLP budding as our model , we probed the role of STIM1 in VLP formation by assessing VLP production from STIM1 suppressed HEK293T cells . eVP40 VLP budding from STIM1 suppressed cells was reduced by approximately 10 fold relative to that from cells receiving random siRNAs or no siRNA ( Fig 3A ) , and the loss of STIM1 had no effect on cellular expression of eVP40 or actin ( Fig 3A; Cells ) . To further confirm this requirement for STIM1 in VLP formation , we utilized a bicistronic vector to suppress endogenous STIM1 ( by targeting the 5’ UTR ) and rescued its expression with exogenous human STIM1 translated from a shRNA resistant cDNA ( shSTIM1-STIM1 plasmid ) ( Fig 3B ) . HEK293T WT cells expressing a fixed amount of eVP40 were transfected with increasing amounts of the shSTIM1-suppression vector or empty vector ( Fig 3B ) . While cellular eVP40 expression levels were equivalent under all conditions ( Fig 3B , Cells ) , progressive suppression of STIM1 expression led to a dose-dependent decrease in eVP40 VLP production ( Fig 3B , VLPs , middle panel ) . Importantly , STIM1 re-expression in suppressed cells fully rescued eVP40 VLP production across all levels of STIM1 suppression ( Fig 3B , VLPs , bottom panel ) . Similar to results with STIM1 shRNA , budding of eVP40 VLPs was significantly reduced ( Fig 3C , VLP ) following siRNA mediated STIM1 suppression ( >90% , Fig 3C , middle panel ) ; and over-expression of exogenous STIM1 restored eVP40 VLP production ( Fig 3C ) . Taken together with results from experiments performed on E106A HEK293T cells , these data definitively establish a role for STIM1/Orai dependent Ca2+ signals in regulation of VLP egress . Genetic approaches outlined above to modulate STIM1 expression and Orai1 permeation establish a novel and common critical role for STIM1 and Orai1 in filovirus and arenavirus budding . Given the broad utility of targeting ion channels to regulate a range of cell physiological functions , we asked whether pharmacological blockade of Orai might represent an effective strategy for regulating filovirus and arenavirus budding . Although high affinity Orai1 blockers for in vivo applications are not available at present , we tested several commercially available inhibitors including Synta66 and 2-APB , both of which inhibit Orai-mediated Ca2+ entry in HEK293T cells at micromolar levels ( 10–50 μM ) [38 , 39] without impacting calcium release from the ER ( S2 Fig ) . Both 2-APB and Synta66 inhibited eVP40- ( Fig 4A and 4B ) and mVP40-induced ( Fig 4C and 4D ) VLP production with identical potency as inhibition of Orai-mediated calcium entry , and neither drug affected cellular expression of VP40 or actin ( Fig 4A–4D ) . A concentration of 2-APB that fully blocks Orai1 channels ( 50 μM ) [38] inhibited eVP40 VLP production ( Fig 4A , right panel ) by ~5 fold and mVP40 VLP production ( Fig 4C , right panel ) by ~50 fold . Likewise , Synta66 ( 50μM ) substantially inhibited eVP40 ( ~5 fold ) and mVP40 VLP ( ~10 fold ) production ( Fig 4B and 4D ) with no effect on steady state levels of cellular VP40 or actin and without altering membrane localization of viral proteins ( S3 Fig ) . Importantly , as neither 2-APB ( Fig 4E ) nor Synta66 ( Fig 4F ) exerted cytotoxic effects on cells under conditions of these measurements ( cell viability , cellular production of VP40 , or VP40 membrane localization , Fig 4 and S3 Fig ) , their anti-budding activity can be attributed to inhibition of Orai-mediated Ca2+ entry . Finally , an additional Orai selective inhibitor ( RO2959 [40] ) , which recently became commercially available , also blocks eVP40 VLP formation ( ~10-fold ) with a potency that parallels its inhibition of calcium permeation of the channel ( S4 Fig , IC50 = ~2 . 5μM ) . Thus , the sensitivity of budding to three chemically distinct Orai inhibitors , at the same concentration that blocks calcium permeation of Orai , further substantiates the critical role of Orai-mediated Ca2+ entry in VLP production . We next sought to validate VLP findings by examining the effect of the Orai1 inhibitors Synta66 and 2-APB on budding of the live attenuated Candid-1 JUNV vaccine strain [41 , 42] . Briefly , VeroE6 cells infected with live attenuated Candid-1 JUNV were cultured in the absence or presence of Orai inhibitors , and infectious virions produced from these cells were quantified in a focus forming assay ( Fig 5 ) . Enumeration of JUNV foci revealed a statistically significant , dose-dependent reduction in JUNV virus production following treatment with Synta66 ( Fig 5A ) or 2-APB ( Fig 5B ) . Moreover , neither compound affected the viability of cells cultured under conditions mimicking those used for JUNV infection experiments ( Fig 5A and 5B , right panels ) , nor affected the synthesis of JUNV GP in infected VeroE6 cells at any concentration tested ( Fig 5A and 5B , Western blots ) . Together , these findings corroborate results of VLP budding assays ( Fig 2 ) and demonstrate that Orai1-mediated calcium entry is required for efficient budding of infectious JUNV . Based on the general requirement we identify for Orai channels in filovirus and arenavirus VLP production and JUNV ( Candid-1 ) budding , we next sought to determine whether Orai channels regulate spread of infectious pathogenic strains of EBOV , MARV , LASV , and JUNV . We first examined viral spread , an indicator of efficient viral budding , in HEK293T cells that constitutively express the dominant negative permeation defective variant of Orai1 ( E106A ) used in VLP assays described above ( Fig 2 ) . These cells were infected at a low multiplicity of infection ( MOI ) , which resulted in the infection of approximately 2–5% of the cells . Cells were then incubated for a period of time that equates to several rounds of viral replication , allowing us to assess viral spread . We observed that the percent of Orai1 E106A expressing cells infected with live BSL-4 variants of EBOV , MARV , JUNV , or LASV was significantly lower than Orai WT cells infected with the same viruses ( Fig 6 ) . These results are consistent with a role for Orai in the spread of infectious filoviruses and arenaviruses . We next assessed the effect of the Orai blocker Synta66 on the spread of these BSL-4 pathogens , because it is a more consistent Orai blocker than 2-APB . Viral spread was assessed by infecting HeLa cells with LASV , JUNV , MARV , or EBOV at a low MOI and then treating with vehicle or Synta66 at the indicated concentrations beginning 1 hour post infection and for the duration of experiments . Seventy two ( LASV , JUNV ) or 96 ( MARV , EBOV ) hours post infection we quantified the percentage of cells infected with virus . For each virus , we observed a significant Synta66 dose-dependent decrease in the percentage of cells infected ( Fig 7A ) . Consistent with inhibition of viral spread , we also observed a general decrease in the number and size of infected cell clusters with increasing Synta66 concentration ( Fig 7B ) . Similar to the more potent inhibition by Synta66 of mVP40- versus eVP40-mediated VLP production ( Fig 4 ) , Synta66 also exerted more potent inhibition of live MARV than EBOV . Interestingly , the spread of both arenaviruses was more sensitive to Orai inhibition than either filovirus ( Fig 7A ) . In general , cultures treated with higher concentrations of Synta66 and for a prolonged period of time ( 72–96 hours ) contained fewer cells than vehicle control treated cultures as measured by the number of nuclei ( Fig 7B ) . For this reason , we normalized the results as the percent of cells infected at the time of analysis for each condition . The decrease in cell numbers , however , does not appear to reflect toxicity ( see Figs 4F and 5A ) . Indeed , as HeLa cells autonomously divide , fewer cells more likely reflects an effect of prolonged Synta66 treatment on cell proliferation . Nonetheless , we evaluated Synta66 induced toxicity by two separate methodologies . Cell-titer Glo “viability” measurements revealed that prolonged Synta66 produced a dose-dependent decrease in ATP ( S5 Fig ) that is attributed to a decrease in the overall number of cells in cultures and not an effect of Synta66 on cell viability ( see Fig 7B ) . We then utilized an Alamar Blue assay to assess the metabolic health of Synta66 treated cells . Indeed , cellular oxidation-reduction potential of Synta66 treated and vehicle control treated cells were equivalent , confirming comparable metabolic activity ( S5 Fig ) . Thus , while prolonged Synta66 treatment resulted in lower overall cell numbers , those cells that are present are metabolically healthy and are fully capable of producing virus . We next sought to definitively establish that the effect of Synta66 on virus spread is due to inhibition of virus egress and not entry . We first pretreated HeLa cells with Synta66 and then infected with a high MOI of LASV , JUNV , MARV , or EBOV . Cells were then fixed after only 2–3 viral replication cycles . Infecting with a high MOI and fixing soon after infection ensured that the extent of infection minimally involves spread between cells and rather reflects the extent of primary infection . Under these high MOI conditions , we observed relatively little effect of Synta66 on infection levels with only modest inhibition of infection evident at high Synta66 concentrations ( S6 Fig ) . Further confirmation that Synta66 blocks egress of live filoviruses and arenaviruses was obtained by assessing the amount of virus released into culture supernatants . Supernatants were collected between 48 and 96 hours post-infection with JUNV , LASV , MARV , or EBOV from Synta66 or vehicle treated HeLa cells . Consistent with all of our VLP ( Figs 2–4 ) and live virus data ( Figs 5 , 6 , 7A and 7B ) we found that Synta66 ( 30μM ) significantly reduced the titer of infectious Lassa , Junín , Marburg , and Ebola virion particles in culture supernatants ( Fig 8 ) . Taken together , this data provide a clear and comprehensive demonstration that Synta66 treatment significantly impairs authentic filovirus and arenavirus budding and release from infected cells . In summary , our results clearly establish that 1 ) Orai1-mediated Ca2+ entry is a critical virus-triggered host signal that regulates filovirus and arenavirus budding , and 2 ) STIM1 and Orai1 represent novel targets for broad-spectrum control of these emerging and often fatal viruses . Indeed , the conserved role for Orai mediated calcium entry among these four hemorrhagic fever viruses raises the interesting possibility that Orai inhibitors may have general utility for broad spectrum control of these and other enveloped RNA viruses that bud by similar Ca2+ dependent mechanisms .
The recent catastrophic outbreak of EBOV in West Africa highlights the need to develop therapeutics for EBOV and other hemorrhagic fever viruses . Indeed , much progress has been made toward the development of candidate vaccines and therapies against EBOV that are currently in clinical trials . Nevertheless , it is critically important that we improve our understanding of the mechanisms of hemorrhagic fever virus pathogenesis not only to identify novel viral targets , but also to identify host targets and common mechanisms that these viruses require for completion of their life cycles as these could lead to the development of broad spectrum host oriented therapeutics . A key advantage of therapeutics that target conserved host pathways required broadly by families of viruses for transmission is the potential for broad spectrum efficacy compared with drugs that target strain specific viral targets . Moreover , host targets should be essentially immutable and thereby insensitive to selective pressures that normally allow pathogens to develop drug resistance [3 , 12–23] . Here we focused on the second messenger Ca2+ and the host proteins responsible for its mobilization and asked whether calcium signals within host cells orchestrate virus assembly and budding . While calcium has been implicated generally in EBOV and HIV-1 budding [32 , 43 , 44] , previous efforts have not addressed if and how matrix proteins encoded by filoviruses or arenaviruses might trigger changes in Ca2+ concentration in host cells . Herein , we demonstrate for the first time that the filovirus matrix protein VP40 and JUNV Z protein trigger STIM1/Orai activation and that the resulting influx of extracellular Ca2+ controls both VLP formation and production of infectious filovirus and arenavirus progeny . Moreover , using Orai channel inhibitors , Orai permeation defective lines , and by suppressing STIM1 expression , we establish STIM1 and Orai as effective host targets for pharmacological regulation of virus egress . It should be noted; however , that we cannot rule out a role for other Orai isoforms ( Orai2 and Orai3 ) in to the residual budding observed for live virus from E106A or Synta66 treated cells . While we have established a critical role for Orai-mediated calcium entry in budding of hemorrhagic fever viruses , the mechanism by which Ca2+ does so remains an important question and the focus of ongoing efforts . Indeed , a number of critical steps implicated in efficient budding of enveloped RNA viruses have been linked to cellular Ca2+ signals , including the activation and localization of specific ESCRT components . Although not absolutely required , the ESCRT pathway has been shown to play a key role in efficient budding of a plethora of RNA viruses including filoviruses , arenaviruses , rhabdoviruses , and retroviruses [5] . It is tempting to speculate that the observed calcium regulation of budding described here may be linked mechanistically to the role of ESCRT during virus egress . For example , the structure , activation , and interactions of ESCRT-related proteins such as Tsg101 , Nedd4 , and Alix have been shown to be regulated in part by calcium [44–47] . Additionally , given that Ca2+ control of membrane repair reflects ESCRT induced shedding of damaged membrane [48] , one might also speculate that Ca2+ dependent mechanisms are similarly triggered by insertion of viral proteins in the plasma membrane . Studies underway are thus focused on determining whether Ca2+ controls budding through regulation of ESCRT pathway function . STIM1 and Orai1 mediated Ca2+ signals have been implicated in distinct steps of the life cycle of other viruses including the replication of Rotaviruses , which are non-enveloped RNA viruses that do not bud from the plasma membrane . Constitutive STIM1 ( and Orai1 ) activation observed in rotavirus-infected cells reflects an effect of its nonstructural protein 4 ( NSP4 ) on endoplasmic reticulum Ca2+ permeability [49] . Indeed , ongoing efforts within our group to understand the mechanisms by which hemorrhagic fever virus matrix proteins trigger STIM1/Orai activation include testing whether VP40 might likewise trigger Ca2+ leak from the ER by inhibiting SERCA pump activity . Furthermore , Ca2+ influx also seems to regulate entry of West Nile virus [50 , 51] , Coxsackievirus [52 , 53] , Hepatitis B virus [54] , and Epstein Barr virus [55 , 56] . Recently it was shown that subunits of a functionally distinct family of voltage-gated calcium channels ( VDCCs ) also play a role in JUNV and Mouse Mammary Tumor pseudovirus entry and infection [57] and that the VDCC blockers nifedipine and verapamil suppressed host cell entry by these viruses . Surprisingly; however , in this instance the involvement of VDCC subunits seemed to be distinct from any role in regulating Ca2+ levels . How VDCC inhibitors might operate independently of any action on VDCC Ca2+ permeation is unclear , but could reflect the promiscuous affinity of VDCC inhibitors for channels including voltage-gated potassium ( Kv ) channels . Indeed , verapamil and nifedipine also block voltage-gated potassium channels that set the membrane potential of non-excitable cells [58 , 59] . Depolarization of the plasma membrane as a result of Kv channel blockade could indirectly block calcium entry by dissipating the electrical driving force ( membrane potential ) required for calcium permeation of Orai [60–62] . While these studies cumulatively point to additional roles for Orai1-mediated and independent Ca2+ influx in steps of infection and replication used by a range of disparate viruses , these roles are distinct from the selective requirement we identify for Orai-dependent calcium entry in budding of filoviruses and arenaviruses . However , Orai might represent a conserved target for regulating budding of additional enveloped RNA viruses , including retroviruses such as HIV-1 , which buds by similar mechanisms . Indeed , similar to hemorrhagic fever viruses , the HIV-1 matrix protein Gag directs HIV-1 budding in part , via well-characterized L-domain interactions with ESCRT proteins , and Gag mediated VLP formation also exhibits dependence on Ca2+ regulation [43] . Further study is needed to fully assess the role for calcium in the HIV-1 lifecycle because , unlike filoviruses and arenaviruses , Gag-mediated VLP production was found to be insensitive to high concentrations of 2-APB ( up to 200uM ) that fully block Ca2+ permeation of Orai channels [44] . In conclusion , we provide the first direct evidence that host Ca2+ signals , triggered by virus activation of STIM1 and Orai , are among key host mechanisms that orchestrate late steps of filovirus and arenavirus assembly and budding . Importantly , from a therapeutic perspective , Orai channels are ubiquitously expressed and like ion channels in general , they represent pharmacologically accessible ( cell surface ) therapeutic targets . While Orai1 inhibitors by themselves appear to have broad spectrum efficacy , an exciting possibility raised by our results is that drug cocktails formulated to target sequential steps in the virus life cycle , including entry , L-domain/host interactions , and other steps involved in budding , could produce enhanced potency , coverage and efficacy over approaches targeting any one host dependent step in the virus life cycle . Thus , while other calcium channel modulators identified may have distinct targets and even calcium independent effects , they may synergize with Orai1 , and also L-domain inhibitors we’ve described previously that block VP40 and Z protein L-domain interactions with host Nedd4 and Tsg101 [42 , 63] . Finally , the ability of certain individuals to survive hemorrhagic fever virus infection seems to reflect their capacity to mount a robust anti-viral immune response . In the context of the severity and the acute nature of these viral diseases , the impact of side effects and even minor effects on cell proliferation that might be associated with long term administration of Orai inhibitors that would be required for immune suppression and immune modulation , may be tolerable in the context of infection with these highly pathogenic and often fatal viruses . Indeed , there is no evidence from murine models that the loss of STIM or Orai activity or function would affect antigen induced lymphocyte activation required for an antiviral immune response [64 , 65] . Thus our prediction is that administration of Orai1 or STIM1 inhibitors , or cocktails that could also include L-domain inhibitors , would slow or dampen virus transmission within and between individuals , and thereby could provide infected individuals additional time needed to mount a protective adaptive immune response . Although Synta66 and the more potent compound RO2959 are no longer being developed as therapeutics , several smaller pharmaceutical companies and academic groups persist in efforts to develop potent Orai1 inhibitors to suppress the pathogenesis of chronic immune-mediated and inflammatory diseases . If and when these become available , direct inhibition of enveloped RNA virus budding from host cells and transmission between individuals may represent an entirely novel use for these channel blockers .
HEK293T , HeLa , and VeroE6 cells were maintained in Dulbecco’s modified Eagle’s medium ( DMEM ) supplemented with 10% fetal calf serum ( FCS ) , penicillin ( 100 U/ml ) /streptomycin ( 100μg/ml ) at 37°C in a humidified 5% CO2 incubator . The stable HEK293T Orai1 E106A mutant-expressing cell line was maintained in DMEM with 10% FCS , penicillin ( 100 U/ml ) /streptomycin ( 100μg/ml ) in the presence of 500 μg/ml of G418 . HeLa cells were maintained in Minimal Essential Medium ( MEM ) with 5% fetal bovine serum , penicillin ( 100 U/ml ) /streptomycin ( 100μg/ml ) at 37°C in a humidified 5% CO2 incubator . The pCAGGS based plasmids expressing EBOV VP40 , MARV VP40 , JUNV Z , LASV Z , and GFP-eVP40 have been described previously [42 , 63 , 66] . mVP40 and JUNZ Z protein are flag tagged while eVP40 was detected using an anti-eVP40 polyclonal antibody previously described [67] . mVP40 and JUNV Z protein were detected with an anti-flag monoclonal antibody ( Sigma-Aldrich ) , and STIM1 was detected with a rabbit anti-STIM1 specific polyclonal antibody ( gift of Dan Billadeau , Mayo Clinic ) . 2-aminoethoxy diphenyl borate ( 2-APB , Sigma Aldrich ) , Synta66 , and RO2959 ( Glixx Labs , Southborough , MA ) were freshly prepared from stock solutions in DMSO . Cell viability in VLP budding and live virus infection assays was examined using an MTT assay ( Amresco ) . 5×103 HEK293T or VeroE6 cells were plated in collagen-coated 96-well tissue culture plates in triplicate . Cells were transfected with empty vector using Lipofectamine for 6 hours , and incubated in serum-free or 2% FCS in phenol-red-free OPTI-MEM in the presence of Synta66 or 2-APB at the indicated concentrations for 20 hours , which mimics the transfection and treatment conditions for VP40 VLP budding . 20μl of MTT solution ( 5mg/ml in PBS ) was added into each well and cells were incubated for 3 . 5 hours . Media was discarded and 150 μl DMSO was added . Absorbance was determined by spectrophotometry using a wavelength of 590 nm . For experiments with BSL-4 variants of filoviruses and arenaviruses , HeLa cells were seeded in 96 well plates ~24 hours prior to addition of Synta66 or vehicle control at indicated concentrations . Cells were incubated for 72 or 96 hour at 37°C in a humidified 5% CO2 incubator and viability was assessed using either CellTiter-Glo assay ( Promega ) , or AlamarBlue assay ( Life Technologies ) , in accordance with manufacturer’s instructions . HEK 293T Orai1-wild type and Orai1-E106A mutant cells ( kind gift from Dr . Jonathan Soboloff , Temple University ) were plated at 5x105 cells/well in a 2-chambered Lab-Tek II Chambered #1 . 5 slide ( Nunc , Rochester , NY ) and grown overnight at 37°C , 5% CO2 in Dulbecco’s modified Eagle’s Medium supplemented with 4 . 5g/L glucose , L-glutamine , sodium pyruvate ( Mediatech , Inc . , Manassas , VA ) , 10% FBS ( Gibco ) , and 1% Penicillin/Streptomycin ( Gibco ) . Cells were transfected with 1μg of R-GECO-1 plasmid ( Addgene , Cambridge , MA ) using Lipofectamine 2000 reagent ( Invitrogen ) according to manufacturer’s instructions in phenol-red-free OPTI-MEM . The next day , cells were transfected with GFP-eVP40 fusion plasmid ( GFP-eVP40 ) using Lipofectamine 2000 reagent ( Invitrogen ) according to manufacturer’s instructions in phenol-red-free OPTI-MEM . Six hours post transfection , fresh phenol-red-free OPTI-MEM was added to the wells , and cells were imaged at 37°C and 5% CO2 in a custom environmental chamber for the duration of the imaging on a Leica DMI4000 with Yokagawa CSU-X1 Spinning Disk Microscope with a 20X dry objective . Cellular R-GECO-1 fluorescence was imaged every 4 seconds for 1 minute periods , repeated every 10 minutes over 18 hours with a Hamamatsu 16-bit cooled EMCCD camera . Imaging and data analysis were performed using the Metamorph 7 . 6 imaging suite . Normalized fluorescence intensity ( F/F0 , where F0 is calculated as the average fluorescence intensity for the initial 1 minute interval ) was calculated for each region of interest ( ROI ) in the time-series . WT HEK293T or HEK293T E106A cells were seeded in collagen-coated six-well plates and transfected with 0 . 5μg of the indicated expression plasmids using Lipofectamine ( Invitrogen ) and the protocol of the supplier . At 6 hours post-transfection , cells were incubated in serum-free OPTI-MEM media or 2% FCS DMEM for 20–24 hours . Cells were incubated with vehicle ( DMSO ) alone , Synta66 , or 2-APB at the indicated concentrations . Culture medium was harvested and centrifuged at 2500 rpm for 10 minutes to remove cellular debris , layered over a 20% sucrose cushion in STE buffer and centrifuged at 220 , 000xG for 2 hours at 4°C . The VP40 VLP-containing pellet was suspended in 50 μl of STE buffer at 4°C overnight . Cells were lysed in RIPA buffer as described above . Viral proteins in VLPs and cell lysates was detected by SDS-PAGE and Western-blot using primary rabbit anti-VP40 antibody for Ebola VP40 , mouse anti-flag antibody for Marburg VP40 , or mouse anti-HA antibody for LASV and JUNV Z followed by an appropriate HRP-conjugated secondary antibody . Human STIM1 siRNAs were purchased from Dharmacon SMARTpools . ON-TARGETplus STIM1 siRNA ( Thermo SCIENTIFIC ) is a mixture of 4 siRNAs to specifically silence the target gene . HEK293T cells in OPTI-MEM media in collagen-coated six-well plate were transfected twice with either control siRNA or STIM1 siRNA at a final concentration of 200nM using Lipofectamine ( Invitrogen ) at 2-day intervals . The final transfection included both siRNAs and 0 . 5μg of Ebola VP40 expression plasmid . VLPs and cell lysates were harvested at 48 hours following the last transfection as described above . VP40 protein levels in VLPs and VP40 and STIM1 levels in cell extracts were analyzed by Western-blot with rabbit anti-VP40 antibody , or rabbit anti-STIM1 antibody , followed by anti-rabbit IgG HRP-conjugated secondary antibody . STIM1 suppression and rescue was accomplished using bicistronic vectors developed in house . STIM1 shRNA generated against the 5’ untranslated region of human STIM1 was expressed using the H1 promoter and human STIM1 cDNA expressed from a CMV promoter in the same construct . VLP samples and cell lysates were harvested at 48 hours after last transfection and analyzed by Western-blot . The Candid-1 vaccine strain of JUNV was kindly provided by Robert B . Tesh ( U . T . M . B . , Galveston , TX ) via Susan R . Ross ( Univ . of Penn . , Philadelphia , PA ) , and was propagated in VeroE6 cells as described previously [41] . For JUNV infection , VeroE6 cells were infected with JUNV ( Candid-1 ) at an MOI of 0 . 02 for 42 hours at 37°C . Supernatants were removed and the cells were washed 3X with 1X PBS . Cells were then treated with DMSO alone , or the indicated concentrations of Synta66 for an additional 30 hours . Virions were harvested from the supernatant samples as described above for VLPs , and then used to infect fresh monolayers of VeroE6 cells for 48 hours for quantification of all foci detected in 6-well plates using fluorescence microscopy as described previously [42] . For all experiments using authentic viruses , Ebola virus ( Kikwit isolate ) , Marburg virus ( Ci67 isolate ) , Lassa virus ( Josiah isolate ) , and Junin virus ( Espindola isolate ) were used . Wild-type HEK293T cells and Orai1 E106A mutant-expressing HEK293T cells , seeded in 96 well black plates ( Corning BioCoat Cellware , Collagen Type I ) , were incubated with EBOV ( MOI = 0 . 5 ) , MARV ( MOI = 0 . 1 ) , JUNV ( MOI = 0 . 1 ) , or LASV ( MOI = 0 . 05 ) in a Biosafety Level 4 laboratory located at USAMRIID . Following 1 hour absorption , virus inoculum was removed and growth media was added . Cells were then incubated at 37°C , 5% CO2 , 80% humidity for 72 ( LASV ) or 96 ( EBOV , MARV , JUNV ) hours , at which time the cells were washed once with PBS and submerged in 10% formalin prior to removal from the BSL4 laboratory . Formalin was removed and cells were washed 3 times with PBS . For LASV infection only , cells were treated with 300mM NaOH for 20 minutes at room temperature prior to the blocking step . Cells were blocked by adding 3% BSA/PBS to each well and incubating at 37°C for 2 hours . EBOV GP-specific mAb KZ52 ( kind gift from Kartik Chandran , Albert Einstein College of Medicine , Bronx , NY ) , MARV GP-specific mAb 9G4 ( USAMRIID ) , LASV GP-specific mAb 52-161-6 ( USAMRIID ) , and JUNV GP-specific mAb GD01-AG02 ( BEI Resources ) , diluted in 3% BSA/PBS , were added to appropriate wells containing infected cells and incubated at room temperature for 2 hours . Cells were washed 3 times with PBS prior to addition of goat anti-mouse or goat anti-human IgG-AlexaFluor 594 ( Invitrogen , Molecular Probes ) secondary antibody . Following 1 hour incubation with secondary antibody , cells were washed 3 times prior to addition of Hoechst 33342 ( Invitrogen , Molecular Probes ) diluted in PBS . Cells were imaged and percent of virus infected cells calculated using the Operetta High Content Imaging System ( PerkinElmer ) and Harmony High Content Imaging and Analysis Software ( PerkinElmer ) . Statistical significance between wild-type cells and Orai1 E106A mutant cells was determined using student t test , two-tailed . For immunofluorescence based assays , HeLa cells , seeded in 96 well black plates ( Greiner Bio-One Cellcoat ) , were incubated with EBOV ( MOI 0 . 1 ) , MARV ( MOI 0 . 1 ) , LASV ( MOI 0 . 01 ) , or JUNV ( MOI 0 . 1 ) for 1 hour at 37°C , 5% CO2 , 80% humidity . Virus inoculum was removed and cells were washed once with PBS . Synta66 was diluted in HeLa cell culture media at indicated concentrations and added to cells . An equivalent percentage of DMSO in HeLa media served as the vehicle control . Cells were then incubated at 37°C , 5% CO2 , 80% humidity for 72 ( LASV and JUNV ) or 96 ( EBOV and MARV ) hours at which time , the cells were washed once with PBS and submerged in 10% formalin prior to removal from the BSL4 laboratory . As described above , virus specific antibodies were added to appropriate wells containing infected cells and samples processed as previously described , except that goat anti-mouse or goat anti-human IgG-AlexaFluor 488 ( Invitrogen , Molecular Probes ) was used as the secondary antibody . Statistical significance was determined by two way ANOVA with Bonferroni multiple comparisons relative to vehicle control treated cells . For viral titer analysis , HeLa cells were seeded in 6 well plates ~24 hours prior to infection with LASV ( MOI = 0 . 01 ) , JUNV ( MOI = 0 . 1 ) , MARV ( MOI = 0 . 1 ) , or EBOV ( MOI = 0 . 1 ) . One hour after infection , cells were treated with Synta66 or vehicle control at indicated concentrations . Culture supernatants were collected at 48 ( MARV ) , 72 ( LASV , JUNV ) or 96 ( EBOV ) hours and cell debris removed by centrifugation . Viral titers of clarified supernatants were determined by routine plaque assay as previously described [68–70] . All data is a graphical representation of at least two independent , replicate experiments . Statistical significance of log transformed data was determined by two way ANOVA with Bonferroni multiple comparisons relative to vehicle control treated cells .
|
Filoviruses ( Ebola and Marburg viruses ) and arenaviruses ( Lassa and Junín viruses ) are high-priority pathogens that hijack host proteins and pathways to complete their replication cycles and spread from cell to cell . Here we provide genetic and pharmacological evidence to demonstrate that the host calcium channel protein Orai1 and ER calcium sensor protein STIM1 regulate efficient budding and spread of BSL-4 pathogens Ebola , Marburg , Lassa , and Junín viruses . Our findings are of broad significance as they provide new mechanistic insight into fundamental , immutable , and conserved mechanisms of hemorrhagic fever virus pathogenesis . Moreover , this strategy of targeting highly conserved host cellular protein ( s ) and mechanisms required by these viruses to complete their life cycle should elicit minimal drug resistance .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
Calcium Regulation of Hemorrhagic Fever Virus Budding: Mechanistic Implications for Host-Oriented Therapeutic Intervention
|
There has been much excitement about the possibility that exposure to specific environments can induce an ecological memory in the form of whole-sale , genome-wide epigenetic changes that are maintained over many generations . In the model plant Arabidopsis thaliana , numerous heritable DNA methylation differences have been identified in greenhouse-grown isogenic lines , but it remains unknown how natural , highly variable environments affect the rate and spectrum of such changes . Here we present detailed methylome analyses in a geographically dispersed A . thaliana population that constitutes a collection of near-isogenic lines , diverged for at least a century from a common ancestor . Methylome variation largely reflected genetic distance , and was in many aspects similar to that of lines raised in uniform conditions . Thus , even when plants are grown in varying and diverse natural sites , genome-wide epigenetic variation accumulates mostly in a clock-like manner , and epigenetic divergence thus parallels the pattern of genome-wide DNA sequence divergence .
Differences in DNA methylation and other epigenetic marks between individuals can be due to genetic variation , stochastic events or environmental factors . Epigenetic marks such as DNA methylation are dynamic; they can be turned over during mitosis and meiosis or altered by chromatin remodeling or upon gene silencing caused by RNA-directed DNA methylation ( RdDM ) . Moreover , changes in DNA sequence or structure caused by , for instance , transposable element ( TE ) insertion , can induce secondary epigenetic effects at the concerned locus [1] , [2] , or , via processes such as RdDM , even at distant loci [3]–[5] . The high degree of sequence variation , including insertions/deletions ( indels ) , copy number variants ( CNVs ) and rearrangements among natural accessions in A . thaliana provides ample opportunities for linked epigenetic variation , and the genomes of A . thaliana accessions from around the globe are rife with differentially methylated regions ( DMRs ) [6]–[10] , but it remains unclear how many of these cannot be explained by closely linked genetic mutations and thus are pure epimutations [11] that occur in the absence of any genetic differences . The seemingly spontaneous occurrence of heritable DNA methylation differences has been documented for wild-type A . thaliana isogenic lines grown for several years in a stable greenhouse environment [12] , [13] . Truly spontaneous switches in methylation state are most likely the consequence of incorrect replication or erroneous establishment of the methylation pattern during DNA replication [14]–[16] . A potential amplifier of stochastic noise is the complex and diverse population of small RNAs that are at the core of RdDM [17] and that serve as epigenetic memory between generations . The exact composition of small RNAs at silenced loci can vary considerably between individuals [13] , and stochastic inter-individual variation has been invoked to explain differences in remethylation , either after development-dependent or induced demethylation of the genome [18] , [19] . Such epigenetic variants can contribute to phenotypic variation within species , and epigenetic variation in otherwise isogenic individuals has been shown to affect ecologically relevant phenotypes in A . thaliana [20]–[22] . In addition to these spontaneous epigenetic changes , the environment can induce demethylation or de novo methylation in plants , for example after pathogen attack [23] . Recently , it has been proposed that repeated exposure to specific environmental conditions can lead to epigenetic differences that can also be transmitted across generations , constituting a form of ecological memory [24]–[27] . The responsiveness of the epigenome to external stimuli and its putative memory effect have moved it also into the focus of attention for epidemiological and chronic disease studies in animals [28] , [29] . How the rate of trans-generational reversion among induced epivariants with phenotypic effects compares to the strength of natural selection , which in turn determines whether natural selection can affect the population frequency of epivariants , is largely unknown [30]–[33] . To assess whether a variable and fluctuating environment is likely to have long-lasting effects in the absence of large-scale genetic variation , we have analyzed a lineage of recently diverged A . thaliana accessions collected across North America . Using a new technique for the identification of differential methylation , we found that in a population of thirteen accessions originating from eight different locations and diverged for more than one hundred generations , only 3% of the genome had undergone a change in methylation state . Notably , epimutations at the DNA methylation level did not accumulate at higher rates in the wild as they did in a benign greenhouse environment . Using genetic mutations as a timer , we demonstrate that accumulation of methylation differences was non-linear , corroborating our previous hypothesis that shifts in methylation states are generally only partially stable , and that reversions to the initial state are frequent [12] , [34] . Many methylation variants that segregated in the natural North American lineage could also be detected in the greenhouse-grown population , indicating that similar forces determined spontaneous methylation variation , independently of environment and genetic background . Population structure could be inferred from differences in methylation states , and the pairwise degree of methylation polymorphism was linked to the degree of genetic distance . Together , these results suggest that the environment makes only a small contribution to durable , trans-generationally inherited epigenetic variation at the whole-genome scale .
Previous studies of isogenic mutation accumulation ( MA ) lines raised in uniform greenhouse conditions identified many apparently spontaneously occurring pure epimutations [12] , [13] . To determine whether variable and fluctuating environments in the absence of large-scale genetic variation substantially alter the genome-wide DNA methylation landscape over the long term , we analyzed a lineage of recently diverged A . thaliana accessions collected across North America . Different from the native range of the species in Eurasia , where nearly isogenic individuals are generally only found at single sites , about half of all North American individuals appear to be identical when genotyped at 139 genome-wide markers [35] . We selected 13 individuals of this lineage , called haplogroup-1 ( HPG1 ) , from locations in Michigan , Illinois and on Long Island , including pairs from four sites ( Fig . 1A , S1 Table ) . Seeds of the accessions had been originally collected between 2002 and 2006 during the spring season , from plants at the end of their life cycle . Because rapid flowering in the greenhouse was dependent on an extended cold treatment , or vernalization , we conclude that the parental plants had germinated in autumn of the previous year and overwintered as rosettes . Climate data from the nearest respective weather station confirmed that precipitation and temperature regimes had varied considerably between sites in the growing season preceding collection ( S1-S2 Fig . ) . Whole-genome sequencing of pools of eight to ten siblings from each accession identified a shared set of 670 , 979 single nucleotide polymorphisms ( SNPs ) and 170 , 998 structural variants ( SVs ) relative to the Col-0 reference genome , which were then used to build a HPG1 pseudo reference genome ( SOM: Genome analysis of HPG1 individuals; S2-S3 Table; S3 Fig . ) . Only 1 , 354 SNPs and 521 SVs segregated in this population ( S4 Table , S4-S5 Fig . ) , confirming that the 13 strains were indeed closely related . Segregating SNPs were noticeably more strongly biased towards GC→AT transitions than shared SNPs , especially in TEs , although the bias was not as extreme as in the greenhouse-grown MA lines ( Fig . 1B ) [36] . A phylogenetic network and STRUCTURE analysis based on the segregating polymorphisms reflected the geographic origin of the accessions ( Fig . 1A , C; S6 Fig . ) . Three of the pairs of accessions from the same site were closely related , and were responsible for many alleles with a frequency of 2 in the sampled population ( Fig . 1D ) . If the spontaneous genetic mutation rate is similar to that seen in the greenhouse [36] , the HPG1 accessions would be 15 to 384 generations separated from each other . With a generation time of one year , their most recent common ancestor would have lived about two centuries ago , which is consistent with A . thaliana having been introduced to North America during colonization by European settlers [37] . This is also in line with the fact that in several US herbarium collections , A . thaliana specimens from the mid-19th century can be found , among these specimens from the Eastern Seaboard and the Upper Midwest . We conclude that the HPG1 accessions constitute a near-isogenic population that should be ideal for the study of heritable epigenetic variants that arise in the absence of large-scale genetic change under natural growth conditions . Because we observed only a weak positive correlation between genetic distance and phenotypic differences in the greenhouse ( S7 Fig . ) , we also infer that life history differences on their own should have little effect on the epigenetic landscape . To assess the long-term heritable fraction of DNA methylation polymorphisms in the HPG1 lineage , we grew plants under controlled conditions for two generations after collection at the natural sites , before performing whole methylome bisulfite sequencing on two pools of 8–10 individuals per accession ( S5 Table ) . We sequenced pools to reduce inter-individual methylation variation and fluctuations in methylation rate caused by stochastic coverage or read sampling bias . After mapping reads to the HPG1 pseudo reference genome , we first investigated epigenetic variation at the single-cytosine level . There were 535 , 483 unique differentially methylated positions ( DMPs ) , with an average of 147 , 975 DMPs between any pair of accessions ( SD = 23 , 745 ) ; thus , 86% of methylated cytosines accessible to our analyses were stably methylated across all HPG1 accessions . The vast majority of variable sites ( 97% ) were detected in the CG context ( CG-DMPs ) . As we have discussed previously [12] , this can be largely attributed to the lower average CHG and CHH methylation rates at individual sites compared to CG methylation , whereby differences in methylation rates are smaller and statistical tests of differential methylation fail more often for CHG and CHH sites . . Additionally , stable silencing-associated methylation of repeats and TEs , elements rich in CHG and CHH sites , may contribute to this pattern . That only about 2% of all covered cytosines were differentially methylated in the relatively uniform HPG1 population contrasted with a previous epigenomic study , in which most cytosines in the genome were found to be differentially methylated in 140 genetically divergent accessions [10] . Fewer than 10% of all cytosines in the genome were never methylated across these 140 accessions , although most methylation events were confined to single strains ( S9 Table of ref . [10] ) . To make our data more comparable to this other study [10] , we identified DMPs of each HPG1 accession against the Col-0 reference genome . On average we found 383 , 237 DMPs per accession , affecting a total of 1 , 046 , 892 unique sites . We estimated that we would have detected 3 . 6 million DMPs , if we had sequenced 140 accessions from the HPG1 lineage ( see Materials and Methods; S8 Fig . ) . The considerably larger number of DMPs in the 140 accessions [10] is likely due both to different methodology and to the higher degree of genetic variation between the analyzed accessions . For example , Schmitz and colleagues [10] did not directly test for differential methylation at individual sites nor did they apply multiple testing correction , which might contribute to the high number of CHH-DMPs reported in that study . Using the geographic outlier LISET-036 as a reference strain , we found that 61% of CG-DMPs as well as 36% of the small number of CHG- and CHH-DMPs were present in at least two independent accessions ( S9A Fig . ) , many of them shared between accessions from the same site . As is typical for A . thaliana [38] , most methylated positions clustered around the centromere and localized to TEs and intergenic regions ( Fig . 2A; S9B Fig . ) . In contrast , differential methylation in the CG context was over-represented on chromosome arms , localizing predominantly to coding sequences ( Fig . 2A; S9B Fig . ) , similar to what we had previously observed in the greenhouse-grown MA lines [12] . We asked whether DMPs had accumulated more quickly in natural environments than in the greenhouse , using DNA mutations in the HPG1 and MA populations as a molecular clock ( SOM: Estimating DMP accumulation rates ) . Our null hypothesis was that a variable and highly fluctuating natural environment increases the rate of heritable methylation changes . In contrast to this expectation , DMPs appear to have accumulated in sub-linear fashion in both the HPG1 and MA populations [12] ( Fig . 2B ) – with similar trends for DMPs in all three contexts – and the number of DMPs did not increase more rapidly in the HPG1 than in the MA lines . The steeper initial increase relative to SNP differences as well as the broader distribution of MA line differences relative to HPG1 differences were most likely the result of having compared individual plants in the MA experiment [12] , rather than pools of siblings , as in the HPG1 experiment . The effect of pooling individuals , as shown by simulation ( S10 Fig . ) , and a potentially higher genetic mutation rate in the wild than in the greenhouse , for example because of increased stress [39] , could lead to a slight underestimation of the true HPG1 epimutation rate , but it remains unlikely that it greatly exceeds the one of the MA lines ( SOM: Estimating DMP accumulation rates ) . Because it is unclear whether variation at individual methylated cytosines has any consequences in plants , we next focused on differentially methylated regions ( DMRs ) in the HPG1 population . A limitation of previous plant methylome studies using short read sequencing has been that these relied on integration over methylated or single differentially methylated sites , or on the analysis of fixed sliding windows along the genome to identify DMRs . What appears intuitively to be more appropriate is to first identify regions that are methylated in individual strains ( SOM: Differentially methylated regions ) [40] , and to test only these for differential methylation . We therefore adapted a Hidden Markov Model ( HMM ) , which had been developed for segmentation of animal methylation data [41] , to the more complex DNA methylation patterns in plants . We identified on average 32 , 529 methylated regions ( MRs ) per strain ( median length 122 bp ) , with the unified set across all strains covering almost a quarter of the HPG1 reference genome , 22 . 6 Mb ( Fig . 2A , C; S11A Fig . ; S6 Table ) . MRs overlapping with coding regions were over-represented in genes responsible for basic cellular processes ( p-value <<0 . 001 ) , in agreement with gene body methylation being a hallmark of constitutively expressed genes [42] . Only 1% of mCHH and 2% of mCHG positions were outside of methylated regions ( Fig . 2D ) , consistent with the dense CHH and CHG methylation found in repeats and silenced TEs [38] . Compared to mCGs within methylated regions , mCGs in unmethylated space localized almost exclusively to genes ( 94% ) , were spaced much farther apart , and were separated by many more unmethylated loci ( Fig . 2E; S11B-C Fig . ) . This explains why sparsely methylated genes were under-represented in HMM-determined methylated regions , even though gene body methylation accounts for a large fraction of methylated CG sites . The accuracy of our MR detection method was well supported by independent methods ( SOM: Validation of methylated regions ) . Using the unified set of MRs , we tested all pairs of accessions for differential methylation , identifying 4 , 821 DMRs with an average length of 159 bp ( S12 Fig . ; S11A Fig . ; S7 Table ) . Of the total methylated genome space , only 3% were identified as being differentially methylated , indicating that the heritable methylation patterns had remained largely stable in this set of geographically dispersed accessions . Indeed , 91% of genic and 98% of the TE sequence space were devoid of DMRs . Of the DMRs , 3 , 199 were classified as highly differentially methylated ( hDMRs; S8 Table ) , i . e . they had a more than three-fold change in methylation rate and were longer than 50 bp . The DMR allele frequency spectrum was similar to that of variably methylated single sites ( Fig . 2F ) . Most DMRs and hDMRs showed statistically significant methylation variation in only one cytosine context , often CG ( Fig . 2G ) , even though DMRs were dominated by CHG and CHH methylation ( Fig . 2D , S13 Fig . ) . Different from individual sites ( DMPs ) , the densities for DMRs and hDMRs were highest in centromeric and pericentromeric regions , and overlapped more often with TEs than with genes ( Fig . 2A , C ) . Relative to all methylated regions , genic regions were two-fold overrepresented in the genome sequence covered by DMRs , and three-fold in hDMRs ( Fig . 2C ) . Currently , we do not know whether this simply reflects the greater power of detecting differential methylation at the typically more highly methylated CG sites compared to CHG or CHH sites , or whether this reflects actual biology . DNA methylation in gene bodies has been proposed to exclude H2A . Z deposition and thereby stabilize gene expression levels [42] . We therefore asked what impact differential methylation had on transcriptional activity . We identified 269 differentially expressed genes across all possible pairwise combinations ( S9-S10 Table ) , most of which were found in more than one comparison . When we clustered accessions by differentially expressed genes , closely related pairs were placed together ( S14 Fig . ) . We identified 28 differentially expressed genes that overlapped with an hDMR either in their coding or 1 kb upstream region , but the relationship between methylation and expression was variable ( S11 Table ) . By visual examination , we found not more than five instances of demethylation that were associated with increased expression; examples are shown in S15 Fig . With the caveat that there are uncertainties about the genetic mutation rate in the wild , and therefore how the number of SNPs relates to the number of generations since the last common ancestor , there was no evidence for faster accumulation of variably methylated sites in the HPG1 population , nor for very different epimutation rates among HPG1 lines ( Fig . 2B ) . Importantly , the overlap of differential methylation between the two populations was much greater than expected by chance: the probability of a random mC site in the MA population of being variably methylated in the HPG1 population was only 7% , but it was 41% among sites that were also variably methylated in the MA population – a six-fold enrichment ( four-fold enrichment in the reciprocal comparison; Fig . 3A ) . In other words , almost half of the DMPs in the MA lines were also polymorphic in the HPG1 lines , and almost a third of HPG1 DMPs were also variably methylated in the MA population . These shared DMPs were more heavily biased towards the chromosome arms and towards genic sequences than population-specific epimutations ( S16A-S16B Fig . ) . Conversely , DMPs unique to one population were more likely to be unmethylated throughout the other population when compared to random methylated sites ( Fig . 3A ) , as one might expect for sites that sporadically gain methylation . DMPs unique to the HPG1 lineage appeared to be less frequent in the pericentromere compared to MA- line-specific DMPs ( S16A Fig . ) , which was also reflected in an apparently higher epimutation frequency in the MA lines for these regions ( S16B Fig . ) . We therefore investigated whether the annotation spectrum differed between these two classes of differentially methylated sites . Even though MA-specific DMPs were more often found in TEs compared to HPG1-specific DMPs , this bias was also observed for all cytosines accessible to our methylome analyses ( S16C Fig . ) , and can therefore be explained by a more accurate read mapping and better TE annotation in the Col-0 reference compared to the HPG1 pseudo-reference genome . Indeed , except for chromosome 4 , the average sequencing depth in the pericentromere was higher in the MA lines ( S16B Fig . ) . DMPs distinguishing MA lines that were separated from each other by only a few generations were more frequently variably methylated in the HPG1 lineage than DMPs identified between distant MA lines ( S17 Fig . ) . We interpret this observation as an indication of privileged sites that are more labile and therefore more likely to have already changed in status after a small number of generations . We used the methods implemented for the HPG1 population to detect DMRs also in the MA strains . Similar to variable single positions , or DMPs , the overlap between 2 , 523 DMRs of the MA lines that we could map to the HPG1 methylated genome space with the 4 , 821 DMRs of the HPG1 accessions was greater than expected and highly significant ( Ζ-score = 32 . 9; 100 , 000 permutations ) . HPG1 DMRs were four-fold more likely to coincide with MA DMRs than with a random methylated region from this set ( Fig . 3B ) . We observed similar degrees of overlap independently of sequence context . Shared DMRs between both lineages were , in contrast to shared DMPs , not biased towards genic regions ( S18 Fig . ) . Differentially methylated regions in the HPG1 lineage , however , overlapped with genic sequences more often than MA DMRs ( S18 Fig . ) , which might again be explained by the different efficiencies in mapping to repetitive sequences and TEs ( S16B Fig . ) . We next wanted to know how this short-term variation compared to methylation variation across much deeper splits . To this end , we identified variably methylated regions between a randomly chosen MA line and a randomly chosen HPG1 line; these DMRs , which differentiate distantly related accessions , were also enriched in each of the two sets of within-population DMRs ( MA or HPG1 ) ( Fig . 3C ) . Finally , we compared DMRs found in the HPG1 population to DMRs that had been identified with a different method among 140 natural accessions from the global range of the species [10] ( Fig . 3D ) . Although only 9 , 994 , less than one fifth , of the variable regions from the global accessions were covered by methylated regions in the HPG1 strains , the overlap of DMRs was highly significant ( Ζ-score = 19 . 8; 100 , 000 permutations ) . Together , the high recurrence of differentially methylated sites and regions from different datasets points to the same loci being inherently biased towards undergoing changes in DNA methylation independently of genetic background and growth environment . To explore potential sources of such lability , we compared variation in the HPG1 lines to that caused by mutations in various components of epigenetic silencing pathways [43] . Almost all variable sites and regions in CG-methylated parts of the HPG1 genome were hypomethylated in mutants deficient in DNA methylation maintenance , most notably in the met1 single and the vim123 triple mutants ( S19 Fig . ) . This is consistent with polymorphic methylation arising primarily because of errors in the maintenance of symmetrical CG methylation during DNA replication . Hypermethylated sites in the rdd triple mutant , which shows impaired demethylation , were also found slightly more often within variably methylated regions of all contexts ( S19D Fig . ) . To quantify how many methylation differences were co-segregating with genome-wide genetic changes at both linked and unlinked sites , we estimated heritability for each highly differentially methylated region by applying a linear mixed model-based method . We used segregating sequence variants with complete information as genotypic data and average methylation rates of hDMRs with complete information as phenotypes . The median heritability of all hDMRs was 0 . 41 ( mean 0 . 44 ) , which means that genetic variance across the entire genome contributed less than half of methylation variance ( Fig . 4A ) . hDMRs in the HPG1 strains that were not methylated in the greenhouse-grown MA lines had a higher median heritability , 0 . 48 , than HPG1 hDMRs also found among MA DMRs ( 0 . 29 ) , which held true for all sequence contexts ( Fig . 4A; S20 Fig . ) . Regions of highly differential methylation found only in the HPG1 population , especially those in unmethylated regions of the MA lines , were thus more likely to be linked to whole-genome sequence variation than hDMRs found in both populations . For 19% of all hDMRs ( 21% CG-hDMRs , 14% CHG-hDMRs , 7% CHH-hDMRs ) , the whole-genome genotype explained more than 90% of their methylation differences ( with a standard error of at most 0 . 1 ) . Of these hDMRs , half had a heritability of greater than 0 . 99 . That 6 . 7% of the sequence space of these heritable hDMRs still overlapped with MA DMRs ( versus 9 . 4% for the less heritable hDMRs ) was in agreement with the hypothesis that there are regions that vary highly in their methylation status independently of genetic background . To identify genetic variants that potentially directly cause methylation changes in their local genomic neighborhood , we focused on variably methylated regions that were within 1 kb of segregating SNPs or indels . Of 191 such DMRs , only three showed a systematic correlation with nearby sequence polymorphisms . We noticed , however , that coding regions with structural variants larger than 20 bp that distinguished the MA and HPG1 populations were more likely to be methylated in both lineages than non-polymorphic coding regions ( Fig . 4B ) . Consequently , DMPs unique to the HPG1 lines were on average closer to insertions or deletions than DMPs shared between the HPG1 and MA populations ( Fig . 4C ) . Lastly , we asked whether the genome-wide methylation pattern reflected genetic relatedness , i . e . , population structure . Hierarchical clustering by methylation rates of variable sites and regions grouped strains by sampling location ( Fig . 4D , E ) . This result was largely independent of the sequence or the annotation context of these loci , and not seen with sites that our statistical tests had identified as stably methylated ( S21 Fig . ) . That variably methylated regions grouped the accessions similar to DMPs , albeit with less confidence ( shorter branch lengths; S21 Fig . ) , suggested that our DMR calling algorithm was conservative . Methylation data thus paralleled similarity between accessions at the genetic level , in agreement with the interpretation that methylation differences primarily reflect the number of generations since the last common ancestor .
We have tested the hypothesis that accumulation of epigenetic variation under natural conditions proceeds over the short term in a very different manner than the clock-like behavior of genetic variation [24]–[27] . To this end , we have taken advantage of a unique natural experiment , the A . thaliana HPG1 lineage , which has likely diverged for at least a century throughout North America . Our analyses have revealed little evidence for broad-scale and durable epigenetic differentiation that might have been induced by the variable and fluctuating environmental conditions experienced by the HPG1 accessions since they separated from each other . While the exact conditions these plants have been subjected to since their separation from a common ancestor remain unknown , the time scale and diversity of geographic provenance are strong indicators of the variability of the environment between the different sampling sites , supported by temperature and precipitation data from nearby weather monitoring stations . The general analytical framework enabled by the HPG1 lineage – nearly isogenic lines grown for more than a century under variable and fluctuating conditions – could not have been achieved in a controlled greenhouse experiment . Studies of epiRIL populations have shown that pure epialleles can be stably transmitted across several generations [5] , [19] , but how often this is the case for environmentally induced epigenetic changes has been heavily debated [33] , [44]–[46] . The recent excitement about the transmission of induced epigenetic variants comes from such variants having been proposed to be more often adaptive than random genetic mutations [24]–[26] . Contrary to the expectations discussed above , we found that epimutation rates under natural growth conditions at different sites did not differ substantially from those observed in a controlled greenhouse environment , with polymorphisms accumulating sub-linearly in both situations , apparently because of frequent reversions . Note that we grew the HPG1 plants under controlled conditions for two generations after sampling at the natural site , to reduce the range of epigenetic variation to the long-term heritable fraction . Given that the environment can induce acute methylation changes [23] , [47] , it is likely that we would have observed greater epigenetic variation , if we had sampled field-grown individuals directly . However , most of such variation induced during ontogeny does not appear to be heritable , as we did not find evidence for it after two extra generations in the greenhouse . Additional studies that directly compare plants grown outdoors to their progeny grown in a stable and controlled environment will help to further clarify this issue . We found that positions of differential methylation in the HPG1 population are more likely to overlap with DMPs detected between closely related MA lines than between more distantly related MA lines . This observation supports the hypothesis that there are different classes of polymorphic sites . One of these includes ‘high lability’ sites that are independent of the genetic background , that change with a high epimutation rate , and that are therefore more likely to appear in each population . Another class of DMPs comprises more stable sites that gain or lose methylation more slowly and that therefore are less likely to be shared between different populations . Differences between accessions in terms of DNA methylation recapitulated their genetic relatedness , further corroborating our hypothesis that heritable epigenetic variants arise predominantly as a function of time rather than as a consequence of rapid local adaptation . Epigenetic divergence thus does not become uncoupled from genetic divergence when plants grow in varying environments , nor does the rate of epimutation noticeably increase . A minor fraction of heritable epigenetic variants may be related to habitat , which could be responsible for LISET-036 being epigenetically a slight outlier ( Fig . 4E ) , even though it is not any more genetically diverged from the most recent common ancestor of HPG1 than other lines . Such local epigenetic footprints could also explain fluctuations in epimutation frequency between the MA and HPG1 lineages . Subtle adaptive changes at a limited number of loci would go unnoticed in the present analysis of genome-wide patterns and can therefore not be excluded . However , on a genome-wide scale there was little indication of adaptive change: neither were LISET-036 specific regions of differential methylation in and near genes enriched for GO terms with an obvious connection to environmental adaptation , nor were there overlapping differentially expressed genes ( S22 Fig . , SOM: Analysis of LISET-036 specific hDMRs ) . In combination with the general lack of correlation between differential methylation and changes in gene expression , our findings suggest that epigenetic changes in nature are mostly neutral , and thus comparable to genetic mutations . We point out that an annual species such as A . thaliana might be differently disposed to record environmental signals in its epigenome compared to more long-lived species . From an evolutionary perspective , in perennial species the advantage of epigenetically mediated local adaptation to changing conditions could be more pronounced , and future studies are warranted to address this question . Because of the near-isogenic background of the HPG1 accessions , we were also able to gauge how much of epigenetic variation is either caused by , or stably co-segregates with genetic differences . HPG1-specific highly differentially methylated regions were more often linked to genotype variation than regions that were variably methylated in both the HPG1 and MA populations . This suggests that heritable hDMRs can , to a certain extent , be considered facilitated epigenetic changes [11] . Both differentially methylated regions and positions are over-represented in genes , but TEs and intergenic regions contain many variable regions and only very few variable single sites . Altogether our data indicate that both variably and constitutively methylated positions in genes are typically separated by many unmethylated sites and that a large fraction of these is therefore not classified as being ( differentially ) methylated . Variability of DNA methylation in plant genes thus mainly affects single , sparsely distributed cytosines , the biological relevance of which remains unclear . Our comparisons between MA laboratory strains and natural HPG1 accessions have revealed that loci of variable methylation overlapped much more between the two groups than expected by chance , despite these populations having experienced very different environments that also differ greatly in their uniformity , and despite completely different genetic backgrounds . The observation that changes at many sites and loci are independent of the genetic background and geographic provenance suggests that spontaneous switches in methylation predominantly reflect intrinsic properties of the DNA methylation and gene silencing machinery , with the CG maintenance system seemingly being the most error-prone . Our most important finding is probably that DNA methylation is highly stable across dozens , if not hundreds of generations of growth in natural habitats; 97% of the total methylated genome space was not contained in a DMR . The stark contrast to published data , which describes more than 90% of cytosines in the genome as variably methylated in a set of 140 divergent natural accessions [10] , can be explained both by the low amount of genetic divergence among the HPG1 accessions and by methodological differences . For future studies , we recommend the application of non-permissive statistical tests in the analysis of differential methylation . The overall stability of methylation presented here is in accordance with the high similarity of methylation in evolutionarily conserved gene sequences [48] . It contrasts , however , with our recent report showing that over longer evolutionary distances that separate species in the same genus or closely related genera , there is very little conservation of global DNA methylation , simply because the sequences that are typically methylated are much more evolutionarily fluid than non-methylated sites [47] . In summary , we propose that the stability of DNA methylation first and foremost depends on the stability of the underlying genetic sequence and that heritable polymorphisms that arise in response to specific growth conditions appear to be much less frequent than those that arise spontaneously . These conclusions are of importance when considering epimutations as a potential evolutionary force .
Accessions [35] were collected in the field at locations indicated in S1 Table . Seeds had been bulked in the Bergelson lab at the University of Chicago before starting the experiment . Plants were then grown at the Max Planck Institute in Tübingen on soil in long-day conditions ( 23°C , 16 h light , 8 h dark ) after seeds had been stratified at 4°C for 6 days in short-day conditions ( 8 h light , 16 h dark ) . We grew one plant of each accession under these conditions; seeds of that parental plant were then used for all experiments . Eight plants of the same accession were grown per pot in a randomized setup . All accessions used in this paper have been added to the 1001 Genomes project ( http://1001genomes . org ) and have been submitted to the stock center . DNA was extracted from rosette leaves pooled from eight to ten individual adult plants . Plant material was flash-frozen in liquid nitrogen and ground in a mortar . The ground tissue was resuspended in Nuclei Extraction Buffer ( 10 mM Tris-HCl pH 9 . 5 , 10 mM EDTA , 100 mM KCl , 0 . 5 M sucrose , 0 . 1 mM spermine , 0 . 4 mM spermidine , 0 . 1% β-mercaptoethanol ) . After cell lysis in nuclei extraction buffer containing 10% Triton-X-100 , nuclei were pelleted by centrifugation at 2000 g for 120 s . Genomic DNA was extracted using the Qiagen Plant DNeasy kit ( Qiagen GmbH , Hilden , Germany ) . Total RNA was extracted from rosette leaves pooled from eight to ten individual adult plants using the Qiagen Plant RNeasy Kit ( Qiagen GmbH , Hilden , Germany ) . Residual DNA was eliminated by DNaseI ( Thermo Fisher Scientific , Waltham , MA , USA ) treatment . DNA libraries for genomic and bisulfite sequencing were generated as described previously [12] . Libraries for RNA sequencing were prepared from 1 µg of total RNA using the TruSeq RNA sample prep kit from Illumina ( Illumina ) according to the manufacturer's protocol . All sequencing was performed on an Illumina GAII instrument . Genomic and bisulfite-converted libraries were sequenced with 2×101 bp paired-end reads . For bisulfite sequencing , conventional A . thaliana DNA genomic libraries were analyzed in control lanes . Transcriptome libraries were sequenced with 101 bp single end reads . Four libraries with different indexing adapters were pooled in one lane; no control lane was used . For image analysis and base calling , we used the Illumina OLB software version 1 . 8 . The SHORE pipeline v0 . 9 . 0 [49] was used to trim and quality-filter the reads . Reads with more than 2 ( or 5 ) bases in the first 12 ( or 25 ) positions with a base quality score of less than 4 were discarded . Reads were trimmed to the right-most occurrence of two adjacent bases with quality values equal to or greater than 5 . Trimmed reads shorter than 40 bases and reads with more than 10% ( of the read length ) of ambiguous bases were discarded . Reads were aligned against the Arabidopsis thaliana genome sequence version TAIR9 in iteration 1 and against updated “Haplogroup 1-like” genomes in further iterations . The mapping tool GenomeMapper v0 . 4 . 5s [50] was used , allowing for up to 10% mismatches and 7% single-base-pair gaps along the read length to achieve high coverage . All alignments with the least amount of mismatches for each read were reported . A paired-end correction method was applied to discard repetitive reads by comparing the distance between reads and their partner to the average distance between all read pairs . Reads with abnormal distances ( differing by more than two standard deviations ) were removed if there was at least one other alignment of this read in a concordant distance to its partner . The command line arguments used for SHORE are listed in S1 File . Genetic variants were called in an iterative approach . In each step , SNPs and structural variants common to all strains were detected and incorporated into a new reference genome . The thus refined “HPG1-like” genomes served as the reference sequence in the subsequent iterations ( S3 Fig . ) . We performed three iterations to call segregating variants and built two reference genomes to retrieve common polymorphisms . The steps performed in each iteration are described in the following paragraphs . Base counts on all positions were retrieved by SHORE v0 . 9 . 0 [49] and a score was assigned to each site and variant ( SNP or small indel of up to 7% of read length ) depending on different sequence and alignment-related features . Each feature was compared to three different empirical thresholds associated with three different penalties ( 40% , 20% and 5% reduction of the score , initial score: 40 ) . They can be found in S13 Table . For comparisons across lines , positions were accepted if at most one intermediate penalty on their score was applicable to at least one strain ( score ≥32 ) . In this case , the threshold for the other strains was lowered , accepting at most one high and two intermediate penalties ( score ≥15 ) . In this way , information from other strains was used to assess sites from the focal strain under the assumption of mostly conserved variation , allowing the analysis of additional sites . Only sites sufficiently covered ( ≥5x ) and with accepted base calls in at least half of all strains ( ≥7 out of 13 ) were processed further . Variable alleles with a frequency of 100% were classified as "common" and variants with a lower frequency as "segregating" . Additional SNPs were called using the targeted de novo assembly approach described below . Although a plethora of SV detection tools have been developed , the predicted variants show little overlap between each other on the same data sets . Furthermore , the false positive rate of many methods can be drastic [51] . Hence , rather than taking the intersection of the output from different tools , which would yield only a small number of SVs , we combined different tools and applied a stringent evaluation routine to identify as many true SVs as possible . Since SVs common to all strains should be incorporated into a new reference , only methods that identify SVs on a base pair level could be used . Currently , there are four different SV detection strategies ( based on depth of coverage , paired-end mapping , split read alignments or short read assembly , respectively ) . Only tools based on split read alignments and assemblies are capable of pinpointing SV breakpoints down to the exact base pair . Programs that were used include Pindel v2 . 4t [52] , DELLY v0 . 0 . 9 [53] , SV-M v0 . 1 [54] and a custom local de novo assembly pipeline targeted towards sequencing gaps ( described below ) . We reported deletions and insertions from all tools , and additionally inversions from Pindel . DELLY combines split read alignments with the identification of discordant paired-end mappings . Thus , our SV calling made use of three out of four currently available methodologies . Reads for DELLY were mapped using BWA v0 . 6 . 2 [55] against the TAIR9 Col-0 reference genome to produce a BAM file as DELLY's input format . The arguments for the command line calls of all tools are listed in S1 File . While using a re-sequencing strategy , there are regions without read coverage ( “sequencing gaps” ) because either the underlying sequence is being deleted in the newly sequenced strain , or highly divergent to the reference sequence , or present in the focal strain , but not represented in the read set . To access sequences in the first two classes , a local de novo assembly method was developed . Insertion breakpoints or small deletions , however , can mostly not be detected by zero coverage due to reads ranging with a few base pairs into or beyond the structural variants . Therefore , we defined a “core read region” as the read sequence without the first and last 10 nucleotides . To be able to assemble the latter cases , the definition of “sequencing gaps” was extended from zero-covered regions to stretches not spanned by a single read's core region . All reads aligned to the surrounding 100 nucleotides of such newly defined sequencing gaps as well as the unmappable reads from the re-sequencing approach together with their potential mapped partners constituted the assembly read set . Two assembly tools were used to generate contigs , SOAPdenovo2 v2 . 04 [56] and Velvet v1 . 2 . 0 [57] ( command line arguments in S1 File ) . Contigs shorter than 200 bp were discarded . To map the remaining contigs of each assembler against the iteration-specific reference genome , their first and last 100 bp were aligned with GenomeMapper v0 . 4 . 5s [50] , accepting a maximal edit distance of 10 . If both contig ends mapped uniquely within 5 , 000 bp , the thus framed region on the reference was aligned to the contig using a global sequence alignment algorithm after Needleman-Wunsch ( ‘needle’ from the EMBOSS v6 . 3 . 1 package ) . In addition , non-mapping contigs were aligned with blastn ( from the BLAST v2 . 2 . 23 package ) [58] to yield even more variants . All differences between contig and reference sequences were parsed ( including SNPs , small indels and SVs ) for each assembly tool . Only identical variants retrieved from both assemblers were selected . For each strain , all variants from the SV tools and the de novo assemblies were consolidated ( S3A Fig . ) and positioned to consistent locations to be comparable using the tool Dindel v1 . 01 [59] . In the case of contradicting or overlapping variants , identical variants ( having the same coordinates and length after re-positioning ) predicted by a majority of tools were chosen and the rest discarded , or all were discarded if there was no majority . Despite sequencing errors or cross-mapping artifacts of the re-sequencing approach , genomic regions covered by reads are generally trusted . Chances of long-range variations in the inner 50% of a mapped read's sequence ( “inner core region” of a read ) are assumed to be low , since gaps would deteriorate the alignment capability towards the ends of the read . Therefore , we filtered out variants from the consolidated variant set spanning a genomic region already covered by at least one inner core region of a mapped read of the corresponding strain ( S3A Fig . ) , assuming homozygosity throughout the genome . This “core read criterion” had to be fulfilled at each genomic position spanned by the variant . Variants passing the core read filter in all strains were classified as common variants and were incorporated into the reference sequence of the previous iteration , thus replacing the reference allele . Segregating variants , which could not be detected in all strains , were additionally built into the reference in separate “haplotype regions” ( or “branches” of the reference sequence ) to eventually be able to assess whether reads preferentially mapped to the reference or the alternative haplotype sequence ( S3A Fig . ) . Linked variant haplotypes of a strain ( distance between consecutive variants ≤107 bp , the maximal possible span of a read on the reference ) as well as identical haplotype regions among strains were merged into one branch sequence . For each strain , all reads were re-mapped to this new reference sequence yielding read counts at the variant site on each branch ( rb ) and at the corresponding site on the reference haplotype sequence ( rref ) ( S3A Fig . ) . Here , the read count of a site was defined as the number of inner core regions spanning the site . To increase certainty of variant calling and to rule out heterozygosity , the read count of the major allele was tested against a binomial distribution that assumed 95% allele frequency out of a total of rb+rref observations , i . e . sole presence of either the branch or the reference haplotype ( if 100% had been assumed , it would not yield a distribution ) . The null hypothesis of homozygosity was rejected after P value correction by Storey's method [60] for q values below 0 . 05 . The same variant could be part of several different haplotypes and thus , could be included into different branch sequences . Reads supporting this variant would map at multiple locations in the reference . Therefore , we allowed all aligned rather than only unique reads to contribute to read counts and omitted the paired-end correction procedure . We followed a similar “population-aware” approach to prefer commonalities among strains as was used for the SNP calling for labeling variants as being common or segregating . Here , variable sites with accumulated coverage over both branch and reference sequence of ≤3x were marked as “missing data” . If there was at least one haplotype in a strain with a q value above 0 . 05 , it was assumed to be present in the population . If the test on the same haplotype failed in another strain , but the absolute read count of the haplotype sequence exceeded the alternative haplotype read count by ≥2-fold , then this haplotype was considered present in the corresponding strain as well . We classified variants where at least 7 out of 13 strains did not show missing data as ‘common’ if the branched haplotype was present in all strains , as ‘not present’ if the reference haplotype was present in all strains , or into ‘segregating’ if there was support for both haplotypes . To combine common variants identified by the described stepwise algorithm into potentially less evolutionary events , we aligned 200 bp around each variant of the last iteration's genome back to the TAIR9 Col-0 reference genome using a global alignment strategy ( ‘needle’ from the EMBOSS v6 . 3 . 1 package ) . In total , we found 842 , 103 common and 2 , 017 segregating polymorphisms without removing linked loci compared to Col-0 after two iterations , to which the different tools contributed to different extent depending on the variant type ( S3C Fig . ) . Genomic and bisulfite sequencing were performed as described in ref . [12] . The procedure followed one described [12] , except that we aligned reads against the HPG1-like as well as against the Col-0 reference genome sequences . Command line arguments for SHORE are listed in S1 File . We performed whole methylome bisulfite sequencing to an average depth of 18x per strand ( S5 Table ) on two pools consisting of 8-10 individuals per accession . We followed the same procedures as described [12] to retrieve statistically significantly methylated positions . Here , we restricted the set of analyzed positions to cytosine sites with a minimum coverage of 3 reads and sufficient quality score ( Q25 ) in at least half of all strains ( i . e . ≥7 ) , that is , 21 million positions in total . Out of those , we identified 3 . 8 million methylated cytosines in at least one strain by applying a false discovery rate ( FDR ) threshold at 5% , and between 2 , 120 , 310 and 2 , 927 , 447 methylated sites per strain ( S5 Table ) . False methylation rates retrieved from read mapping against the chloroplast sequence can be found in S5 Table . Using the HPG1 pseudo reference genome instead of the Col-0 reference genome increased the number of cytosines sufficiently covered for statistical analysis by 5% on average , and the number of positions called as methylated by 7% ( S5 Table ) . We performed the same methods as in ref [12] to obtain DMPs . First , cytosine positions were tested for statistical difference between both replicates of a sample using Fisher's exact test and a 5% FDR threshold . Because individual samples consisted of a pool of several plants , the number of DMPs between replicates was negligible ( between 0 and 161 ) . After excluding them , we applied Fisher's exact test on the 3 . 8 million cytosine sites methylated in at least one strain for all pairwise strain comparisons . Using the same P value correction scheme as in Becker et al . , we identified 535 , 483 DMPs across all 13 strains . Using the model developed in ref [61] , a beta prior distribution was estimated that determined the non-ancestral frequency for each variable site . We assumed the methylation state in Col-0 to be ancestral , which resulted in beta distribution parameters of a = 0 . 029 and b = 0 . 644 , corresponding to a mean non-ancestral DMP frequency of 0 . 043 and a corresponding standard deviation of 0 . 157 . These were then used to estimate the fraction of common DMPs that were expected to be found by sequencing a given number of methylomes . Based on the formula presented in supporting section 3 of ref [61] , we estimated the total number of DMPs in the population: For Nind = 13 ( the number of accessions in this study ) and Δ ( 1 ) = 1 , 046 , 892 ( the total number of DMPs versus the Col-0 reference ) , we estimated a total number of possible DMPs in the population of N = 59 , 770 , 415 , which is close to the 43 million cytosines in the A . thaliana genome . Given such an estimate for N , the Δ function can be evaluated numerically to estimate the number of DMPs we would have detected had we analysed the same number of accessions as in ref [10] ( S8 Fig . ) . The value of an approach that defines methylated regions ( MRs ) before identifying differentially methylated regions ( DMRs ) has been demonstrated before with a Hidden Markov Model ( HMM ) method developed for the analysis of methylated-DNA-immunoprecipitation followed by array hybridization ( MeDIP-chip ) [40] . An HMM based on next-generation sequencing data was also applied to segment the maize genome , which is much more highly methylated than the A . thaliana genome , into hypo- and hypermethylated regions [62] . We modified the HMM implementation from Molaro and colleagues [41] based solely on within-genome variation in methylation rate . It assumes that the number of methylation-supporting reads at each cytosine follows a beta binomial distribution and that distributions over positions within and between methylated regions will differ from each other , providing a way to distinguish them . Thus , the model learns methylation rate distributions for both an unmethylated and a methylated state for each sequence context separately ( CG , CHG and CHH ) while simultaneously estimating transition probabilities between the two states from genome-wide data . On the trained model , the most probable path of the HMM along the genome is then used to define regions of high and low methylation . The method of Molaro and colleagues was designed for calling MRs in human samples , where the vast majority of methylated cytosines are in a CG context . In plants , however , one observes considerable methylation in all three contexts ( CG , CHG and CHH ) , each with a different methylation rate distribution . Hence , we extended the HMM by learning the parameters of three different beta binomial distributions per state , one for each context . Additionally , in contrast to humans , only the minority of cytosines in the CG context is methylated , as are cytosines in the other contexts . Hence , methylation rates were inverted to find hypermethylated , rather than hypomethylated regions as in the original HMM implementation . Apart from these changes and some final filtering steps ( see below ) , we followed the same computational steps as described by Molaro and colleagues [41]: The describing parameters of the – in our case – six distributions ( determining the emission probabilities ) and the transition probabilities between states were iteratively trained ( using the Baum-Welch algorithm ) from methylation rates of all cytosines in the corresponding context throughout the genome . After each iteration , all cytosines were probabilistically classified into the most likely state via Posterior Decoding , given the trained model . After training of the HMM , i . e . after maximally 30 iterations or when convergence criteria were met , consecutive stretches of high methylation state were scored , in our case by the sum of all contained methylation rates . Next , P values were computed by testing the scores against an empirical distribution of scores obtained by random permutation of all cytosines throughout the genome . After FDR calculation , consecutive stretches in high state with an FDR <0 . 05 are defined as methylated regions ( MRs ) . The HMM was run on all genome-wide cytosines , independent of their coverage . Methylation rates were obtained using accumulated read counts from the strain replicates , resulting in one segmentation of the genome per strain . Gaps of at least 50 bp without a covered C position within a high methylation state automatically led to the end of the high methylation segment . Positions with a methylation rate below 10% at the start or end of highly methylated regions ( until the first position with a rate larger than 10% ) , were assigned to the preceding or subsequent low methylation region , respectively . The method to identify MRs yielded 13 different segmentations of the genome , one for each strain . We selected regions being in different or highly methylated states between strains and statistically tested them for differential methylation ( including FDR calculation ) . To obtain epiallele frequencies , we clustered strains into groups based on their pairwise comparisons and statistically tested the groupings against each other . Regions that showed statistically significant methylation differences between at least two sets of strains were identified as DMRs . Finally , because of the sensitivity of the statistical test , we empirically filtered DMRs for strong signals and defined highly differentially methylated regions ( hDMRs ) . All these steps are described in depth in the following . We defined a breakpoint set containing the start and end coordinates of all predicted methylated regions . Each combination of coordinates in this set defined a segment to perform the test for differential methylation in all pairwise comparisons of the strains , if at least one strain was in a high methylation state throughout this whole segment ( S12A Fig . ) . To also detect quantitative differences rather than solely presence/absence methylation , we also compared entirely methylated regions in more than one strain to each other . Because of the sheer number of such regions , we applied the following greedy filter criteria: Regions were discarded from any pairwise comparison if less than 2 strains contained at least 10 cytosines covered by at least 3 reads each ( accumulated over strain replicates ) in this region ( S12A Fig . ( a ) ) . Regions were discarded from any pairwise comparison if the reciprocal overlap of this region to at least one previously tested region was more than or equal to 70% ( S12A Fig . ( b ) ) . This was done to prevent “similar” regions to be tested twice . Pairwise tests of a region were not performed if both strains were in low methylation state throughout the whole region ( S12A Fig . ( c ) ) . Strains were excluded from pairwise comparisons in a region if the number of positions covered by at least 3 reads each was less than half of the maximum number of such positions of all strains in the same region ( S12A Fig . ( d ) ) . This prevented comparing regions with unbalanced coverage to each other , e . g . a strain with 10 data points against another one with only 2 . These filters reduced the set of regions to test from ∼2 . 5 million to ∼230 , 000 per pairwise comparison . We designed a statistical test for differential methylation between two strains for a given region . The test assumes that the number of methylated and unmethylated read counts per position along a region follows a beta binomial distribution – similar to the HMM in MR calling . More precisely , there are 3 distributions for each sequence context and for each strain . Using gradient-based numerical maximum likelihood optimization , we fitted the parameters for each beta binomial distribution on the available read count data of the region in the respective strain . This was done a ) for each of the two strains separately ( while taking strain replicates into account ) , resulting in ( two times three ) strain-specific beta binomial distributions , and b ) for the read counts of both strains including their replicates together , resulting in ( three ) common beta binomial distributions . In this way , we obtained each distribution's mean µ and standard deviation σ . We selected only regions for potential DMRs , whose intervals [µ1 – 2σ1 , µ1 + 2σ1] for strain 1 and [µ2 – 2σ2 , µ2 + 2σ2] for strain 2 did not overlap . To further corroborate statistical significance , we computed P values by calculating the ratio of the strain-specific and the common log likelihoods of the available read count data using the corresponding beta binomial distributions and by testing it against a chi-squared distribution ( with 6 degrees of freedom ) . Let sample S have NSc cytosines in context c in total and CScp reads at position p in context c , from which xScp are methylated , then we compute: After correction for multiple testing using Storey's method [60] , an FDR threshold of 0 . 01 defined statistically different methylated regions ( DMRs ) between two strains . Additionally , this method allowed calling differential methylation in a region for each context separately by computing P values as described above without summing over the contexts ( c = 1 , 2 or 3 ) . We termed resulting DMRs CG-DMRs if the methylation at only CG sites within this region was statistically significantly different , and similarly CHG-DMRs and CHH-DMRs . For 13 strains there are at maximum 78 pairwise comparisons per region . To summarize pairwise comparisons and obtain epiallele frequencies , we assigned strains into differentially methylated groups . To achieve such clustering , we constructed a graph for each region where strains were represented as vertices and connected to other strains by an edge if the region was identified as a DMR between them ( S12B Fig . ) . We assume that strains within a group are then similarly methylated . The task is to find the smallest number of groups of vertices so that no two strains within a group are connected by an edge . We set up a custom algorithm , which iteratively solves the “vertex coloring problem” for an increasing number of different colors , starting with two and quitting once all strains could be successfully assigned a color ( S12B Fig . ) . In each iteration , strains were processed in descendent order of their degree ( i . e . number of edges it is connected to ) . Each strain was assigned to all possible colors that did not invoke a collision . Subsequently , the algorithm continued recursively to assign the color of the next strain . Each strain had 3 context-dependent means of its beta binomial distributions per region ( termed strain means from now on ) . We roughly approximated each group's mean methylation values ( group means ) as the mean values of all strain means within a group . The grouping diversity describes the accumulated absolute differences between the strain means and their respective group means divided by the number of strains . As an example , consider S12B Fig . For simplicity , it only displays methylation rates for one out of three contexts . In the real data , the respective values were accumulated over all three contexts . The group mean for the blue strains in the example is ( 89+90+90+93+87 ) /5 = 89 . 8% and for the white strains 52% . The grouping diversity of the clustering shown here would be ( from strains A to K ) : ( |56–52|+|59–52|+|64–52|+|89–89 . 8|+|41–52|+|93–89 . 8|+|90–89 . 8|+|45–52|+|47–52|+|90–89 . 8|+|45–52|+|87–89 . 8| ) /11 = 2 . 84 . If there was more than one possible grouping of the strains , we chose the one with the lowest grouping diversity . A strain with no edges ( i . e . which is not statistically differentially methylated to any other strain ) was assigned into the group to which the accumulated absolute difference between its strain mean and the group mean was lowest . In the example of S12B Fig . , strain L is grouped to the blue strains because its mean methylation value ( 81% ) is closer to the blue group mean ( 90% ) than to the white one ( 52% ) . This procedure summarized the ∼221 , 000 DMRs of all pairwise strain comparisons into 11 , 323 DMRs between groups of strains . Once grouped , the same statistical test as for differential methylation between two strains was used to test groups of strains . Beta binomial distributions were approximated using the read counts of all strains in a group as if they were replicate data . This procedure identified 10 , 645 groups of regions showing significantly different methylation . Because the method used for the selection of the regions to perform the differential test can result in overlapping regions , DMRs can still overlap each other . From sets of overlapping DMRs , the non-overlapping DMR ( s ) with the lowest ‘grouping diversity’ was ( were ) retained , resulting in 4 , 821 final DMRs . For the vast majority of DMRs ( 98% ) , strains were classified into two groups , i . e . there are only two epialleles . Our sensitive statistical test classified as differential some regions with low variance and only subtle methylation difference; we therefore defined as highly differentially methylated regions ( hDMRs ) with potentially greater biological relevance all DMRs that were longer than 50 bp and that showed a more-than-three-fold difference in methylation rate in at least one sequence context , when considering at least five cytosines of that context ( S12 Fig . ) . In addition , the overall methylation rate of the DMR in the more highly methylated strain had to be greater than 20% . Of 3 , 909 size-filtered DMRs , 3 , 199 ( 80% ) were classified as hDMRs ( S8 Table ) . The grouping of hDMRs yielded similar epiallele frequencies as for the DMPs ( 54% with frequency larger than 1; Fig . 2F ) . The data from Stroud and colleagues [43] contain position-wise methylation rates for each sample . We defined a single site as methylated in wild type ( WT ) if both Col-0 samples Col_WA034L3 and Col_WB023L8 had a methylation rate of 10% or higher , and if at least one of them is more than 20% methylated . We declared a site in a mutant sample as having ‘lost’ methylation where the wild type was methylated and the mutant showed a methylation rate of less than 10% . In contrast , a ‘gained’ methylation site had less than 10% methylation in at least one of the WT samples and more than 20% methylation in the mutant . To assess if epigenetic variation in the HPG1 lines is enriched at sites affected by impaired methylation machinery , for each mutant , we constructed a set of positions , which were methylated in WT , covered in the mutant sample ( i . e . present with a rate in the mutant sample file ) , and which were covered in the HPG1 and MA populations . A site was considered covered in a population when more than half of the strains showed a high quality and a more than 3-fold covered base call ( see ‘Determination of methylated sites’ or [12] ) . For those positions and different subsets thereof , the fractions of sites with gained or lost methylation in the mutant compared to the wild type samples were plotted in S19 Fig . For each differentially methylated region , we considered a linear mixed model to estimate the proportion of variance that is attributable to genetic effects ( heritability ) and its standard error . The approach is similar to variance component models used in GWAS , e . g . refs . [63] , [64] . Briefly , we considered the log average methylation rate of DMRs as phenotype and assessed the variance explained by genotype using a Kinship model constructed from all segregating genetic variants . We considered only DMRs and genetic polymorphisms that had no missing data in all accessions . We identified non-synonymous SNPs using SHOREmap_annotate [65] and excluded them from population structure analyses . We ran STRUCTURE v . 2 . 3 . 4 [66] with K = 2 to K = 9 with a burn-in of 50 , 000 and 200 , 000 chains for 10 repetitions and determined the best K value using the ΔK method [67] . The phylogenetic network was generated using SplitsTree v . 4 . 12 . 3 [68] . We used the TAIR10 annotation for genes , exons , introns and untranslated regions; transposon annotation was done according to [69] . Positions and regions were hierarchically assigned to annotated elements in the order CDS> intron> 5′ UTR> 3′ UTR> transposon> intergenic space . We defined as intergenic positions and regions those that were not annotated as either CDS , intron , UTR or transposon . Positions were associated to the corresponding element when they were contained within the boundaries of that element . ( D ) MRs were associated to a class of element if they overlapped with that class of element; a ( D ) MR could only be associated to one class of element . When summing up basepairs of an element class covered by ( D ) MRs , the number of basepairs of a ( D ) MR overlapping with that class of element were considered . In that case the space covered by a ( D ) MR could be assigned to different classes of elements , while each basepair of the ( D ) MR could be assigned to only one class . We tested for significant overlap of DMRs using multovl version 1 . 2 ( Campus Science Support Facilities GmbH ( CSF ) , Vienna , Austria ) . We reduced the genome space to the basepair space covered by MRs identified in at least one HPG1 accession . DMRs were considered in the analysis if their start and end positions were contained within the MR space . DMRs that only partially overlapped with the MR space were trimmed to the overlapping part . Overlap between DMRs from different datasets was analyzed by running 100 , 000 permutations of both DMR sets within the MR basepair space . multovl commands are listed in S1 File . Reads were processed in the same way as genomic reads , except that trimming was performed from both read ends . Filtered reads were then mapped to the TAIR9 version of the Arabidopsis thaliana ( http://www . arabidopsis . org ) genome using Tophat version 2 . 0 . 8 with Bowtie version 2 . 1 . 0 [70] , [71] . Coverage search and microexon search were activated . The command lines for Tophat are listed in S1 File . For quantification of gene expression we used Cufflinks version 2 . 0 . 2[72] . We ran a Reference Annotation Based Transcript assembly ( RABT ) using the TAIR10 gene annotation ( ftp://ftp . arabidopsis . org/home/tair/Genes/TAIR10_genome_release/TAIR10_gff3/ ) supplied with the most recent transposable element annotation [69] Fragment bias correction , multi-read correction and upper quartile normalization were enabled; transcripts of each sample were merged using Cuffmerge version 2 . 0 . 2 , with RABT enabled . For detection of differential gene expression we ran Cuffdiff version 2 . 0 . 2 on the merged transcripts; FDR was set to <0 . 05 and the minimum number of alignments per transcripts was 10 . Fragment bias correction , multi-read correction and upper quartile normalization were enabled . The command lines for the Cufflinks pipeline are listed in S1 File . Analysis and graphical display of differential gene expression data was done using the cummeRbund package version 2 . 0 . 0 under R version 3 . 0 . 1 . When not mentioned otherwise in the corresponding paragraph , graphical displays were generated using R version 3 . 0 . 1 ( www . r-project . org ) . Circular display of genomic information in Fig . 2A was rendered using Circos version 0 . 63 [73] . Leaf area was determined using the automated IPK LemnaTec System and the IAP analysis pipeline [74] . Plants were grown in a controlled-environment growth-chamber in an alpha-lattice design with eight replicates and three blocks per replicate , taking into account the structural constraints of the LemnaTec system . Each block consisted of eight carriers , each carrying six plants of one line . Stratification for 2 days at 6°C was followed by cultivation at 20/18°C , 60/75% relative humidity in a 16/8 h day/night cycle . Plants were watered and imaged daily until 21 days after sowing ( DAS ) . Adjusted means were calculated using REML in Genstat 14th Edition , with genotype and time of germination as fixed effects , and replicate|block as random effects . Local temperature and liquid precipitation data was calculated from National Climatic Data Center ( NCDC ) Global Summary of Day ( GSOD ) data . Collection locations were matched to the closest weather station with <5% missing data for five years prior to the collection date . Cumulative liquid precipitation was calculated each year starting from January 1 . The DNA and RNA sequencing data have been deposited at the European Nucleotide Archive under accession number PRJEB5287 and PRJEB5331 . A GBrowse instance for DNA methylation and transcriptome data is available at http://gbrowse . weigelworld . org/fgb2/gbrowse/ath_methyl_haplotype1/ . DNA methylation data , MR coordinates and genetic variant information have also been uploaded to the genome browser of the EPIC consortium ( https://www . plant-epigenome . org/; https://genomevolution . org/wiki/index . php/EPIC-CoGe ) and can be accessed at http://genomevolution . org/r/939v . The software of our methylation pipeline can be downloaded at http://sourceforge . net/projects/methpipeline .
|
It continues to be hotly debated to what extent environmentally induced epigenetic change is stably inherited and thereby contributes to short-term adaptation . It has been shown before that natural Arabidopsis thaliana lines differ substantially in their methylation profiles . How much of this is independent of genetic changes remains , however , unclear , especially given that there is very little conservation of methylation between species , simply because the methylated sequences themselves , mostly repeats , are not conserved over millions of years . On the other hand , there is no doubt that artificially induced epialleles can contribute to phenotypic variation . To investigate whether epigenetic differentiation , at least in the short term , proceeds very differently from genetic variation , and whether genome-wide epigenetic fingerprints can be used to uncover local adaptation , we have taken advantage of a near-clonal North American A . thaliana population that has diverged under natural conditions for at least a century . We found that both patterns and rates of methylome variation were in many aspects similar to those of lines grown in stable environments , which suggests that environment-induced changes are only minor contributors to durable genome-wide heritable epigenetic variation .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biogeography",
"ecology",
"and",
"environmental",
"sciences",
"plant",
"science",
"model",
"organisms",
"plant",
"genomics",
"epigenetics",
"theoretical",
"ecology",
"plant",
"ecology",
"dna",
"methylation",
"arabidopsis",
"thaliana",
"research",
"and",
"analysis",
"methods",
"gene",
"expression",
"plant",
"genetics",
"phylogeography",
"dna",
"modification",
"plant",
"and",
"algal",
"models",
"ecology",
"genetics",
"biology",
"and",
"life",
"sciences"
] |
2015
|
Century-scale Methylome Stability in a Recently Diverged Arabidopsis thaliana Lineage
|
Macrophages can be niches for bacterial pathogens or antibacterial effector cells depending on the pathogen and signals from the immune system . Here we show that type I and II IFNs are master regulators of gene expression during Legionella pneumophila infection , and activators of an alveolar macrophage-intrinsic immune response that restricts bacterial growth during pneumonia . Quantitative mass spectrometry revealed that both IFNs substantially modify Legionella-containing vacuoles , and comparative analyses reveal distinct subsets of transcriptionally and spatially IFN-regulated proteins . Immune-responsive gene ( IRG ) 1 is induced by IFNs in mitochondria that closely associate with Legionella-containing vacuoles , and mediates production of itaconic acid . This metabolite is bactericidal against intravacuolar L . pneumophila as well as extracellular multidrug-resistant Gram-positive and -negative bacteria . Our study explores the overall role IFNs play in inducing substantial remodeling of bacterial vacuoles and in stimulating production of IRG1-derived itaconic acid which targets intravacuolar pathogens . IRG1 or its product itaconic acid might be therapeutically targetable to fight intracellular and drug-resistant bacteria .
Intracellular bacteria are major causes of morbidity and mortality . Upon infection , many intracellular pathogens establish intracellular membrane-bound compartments , where they resist lysosomal degradation and humoral immune responses [1] . As a result of co-evolution , host cells have in turn developed strategies to target the vacuoles or the bacteria inside in order to control infections [2 , 3] . Interferons ( IFNs ) , which are classified into type I , II and III IFNs [4] , are potent inducers of intracellular immunity in vertebrates [2] . They fulfill this function by activating transcription of partly overlapping sets of so-called IFN-stimulated genes ( ISGs ) , several of which with antiviral or antibacterial activities . However , the exact functions of many ISGs remain unknown [2] . Legionella pneumophila is a frequent cause of severe pneumonia in humans and a model for investigating immune responses to intravacuolar bacteria . Upon infection , L . pneumophila is phagocytosed by alveolar macrophages , where L . pneumophila establishes a specialized replication vacuole , named the Legionella-containing vacuole ( LCV ) . This process requires the Dot/Icm type IV secretion system ( T4SS ) which injects around 300 bacterial effector molecules into the host cytosol [5] . In replication-permissive cells , the LCV escapes fusion with lysosomes and instead recruits secretory vesicles from the endoplasmic reticulum ( ER ) as well as mitochondria [5–8] . In macrophages of C57BL/6 mice ( and most other inbred strains ) , however , wild-type ( wt ) L . pneumophila is restricted by the NAIP5 inflammasome which detects bacterial flagellin and stimulates cell death as well as phagolysosomal maturation [9–13] . In contrast to wt bacteria , L . pneumophila lacking flagellin are not recognized by NAIP5 and are thus able to replicate in mouse macrophages . We and others recently demonstrated that L . pneumophila is additionally controlled by a cell-autonomous defense pathway that is activated by auto-/paracrine type I IFN signaling [14–19] . This defense pathway restricts the bacteria in their vacuole without preventing LCV formation or triggering lysosomal fusion [15] . In the present study , we systematically examined the antibacterial innate immune response to L . pneumophila infection and demonstrate that type I and II IFNs substantially alter the composition of bacterial vacuoles , induce production of bactericidal itaconic acid via IRG1 , and restrict L . pneumophila replication in alveolar macrophages and lungs .
In order to identify master regulators of the innate immune response to intracellular bacteria , we compared gene expression in the lungs of L . pneumophila-infected and sham-treated C57BL/6 WT mice . We identified 1526 genes ( S1 Dataset ) that were induced upon infection . Upstream regulator analysis was performed with Ingenuity Pathway Analysis ( IPA ) ( Fig 1A ) and revealed that type I and II IFNs and their related transcription factors ( e . g . STAT1 , IRF3 , IRF7 ) play a predominant role in controlling gene transcription in response to L . pneumophila infection ( Fig 1B ) . This in silico prediction was confirmed by transcriptome analysis of L . pneumophila-infected Ifnar-/- , Ifngr-/- , Ifnar/Ifngr-/- mice , all of which showed a severely impaired transcriptional response compared to WT animals ( Fig 1C , S1 Dataset ) . To investigate the functional relevance of the type I and II IFNs for the antibacterial defense against L . pneumophila , we analyzed bacterial clearance following infection of WT , Ifnar-/- , Ifngr-/- and Ifnar/Ifngr-/- mice . Whereas WT , Ifnar-/- and Ifngr-/- mice were able to clear or strongly reduce bacterial burdens by day 6 post infection ( p . i . ) , bacterial loads remained high in Ifnar/Ifngr-/- mice ( Fig 1D ) . This is in line with our previously published results from infections with L . pneumophila ΔflaA [15] . Together , these data indicate that type I and type II IFNs are critical regulators of early gene expression and the antibacterial innate immune response during L . pneumophila infection . Alveolar macrophages , but not dendritic cells ( DCs ) , are the primary cell type supporting L . pneumophila infection in vivo [20–22] . Therefore , we questioned whether an IFN-mediated alveolar macrophage-intrinsic defense pathway is relevant during L . pneumophila infection in vivo . To this end , we constructed a chimeric mouse model in which IFN signaling was selectively abrogated in CD11c+ cells , whereas at least 50% of all other hematopoietic cell types express the IFN receptors ( Fig 2A ) . In the lung >90% of CD11c+ cells are alveolar macrophages and only a minority of pulmonary CD11c+ cells in steady state are DCs . CD45 . 1+ mice were irradiated and reconstituted with a 1:1 mixture of CD45 . 2+ bone-marrow cells from Ifnar/Ifngr-/- or WT and CD11c-DTR-GFP mice ( expressing the diphtheria toxin receptor ( DTR ) under the control of the CD11c promoter ) . Repopulation was assessed to be >90% after 10 weeks ( S1A Fig ) and mice were subsequently infected with L . pneumophila wt . First , we analyzed all L . pneumophila-infected mice repopulated with Ifnar/Ifngr-/- and CD11c-DTR-GFP cells including those showing a weak depletion of CD11c+ GFP+ cells by diphtheria toxin ( DTX ) in the lung ( S1B Fig ) . We observed a significant negative correlation between remaining CD11c+ GFP+ cells ( expressing IFN receptors ) and pulmonary bacterial load ( Fig 2B ) . This correlation indicates that the number of IFN-responsive CD11c+ cells has a direct positive impact on bacterial clearance . Second , we examined bacterial loads only in the bone-marrow-chimeric mice which showed a highly efficient DTX-mediated depletion of CD11c+ DTR-expressing GFP+ cells ( with <10% remaining , S1 Fig ) . Strikingly , chimeric mice lacking the IFN receptors in CD11c+ cells ( CD11c-DTR / Ifnar/Ifngr-/- + DTX ) were unable to clear L . pneumophila wt infection ( Fig 2C ) , and were thus comparable to Ifnar/Ifngr-/- mice ( Fig 1D ) . In contrast , chimeric mice without depletion of IFNAR/IFNGR-expressing CD11c+ cells ( CD11c-DTR / Ifnar/Ifngr-/- + PBS ) showed a significant reduction of bacterial burdens ( Fig 2C ) . Chimeric mice reconstituted with solely IFN-responsive cells ( CD11c-DTR / WT + DTX ) finally were able to clear the infection almost completely . Given that DCs do not support L . pneumophila growth [20 , 21] , our data strongly suggest that IFNs induce alveolar macrophage-intrinsic effects to restrict intracellular infection . In line with this conclusion , L . pneumophila ΔflaA , which is able to replicate in WT alveolar macrophages due to evasion of the NAIP5 inflammasome [9–12] , is partially inhibited by IFNβ and completely blocked by IFNγ treatment ( Fig 2D ) . Conversely , Ifnar-/- and Ifnar/Ifngr-/- alveolar macrophages supported replication of otherwise growth-restricted L . pneumophila wt ( Fig 2E ) . These data indicate that endogenously produced type I IFNs control bacterial growth , whereas type II IFN is not relevant in this ex vivo model since alveolar macrophages produce no or only negligible levels of IFNγ [23] . Collectively , our data indicate that L . pneumophila lung infection is controlled by an IFN-dependent alveolar macrophage-intrinsic mechanism . To determine the molecular basis of how macrophages restrict L . pneumophila upon activation by IFNs , we made use of bone marrow-derived macrophages ( BMMs ) , an easily available and frequently used cell model to study L . pneumophila infection [9–12 , 14 , 15] . As shown in alveolar macrophages ( Fig 2D ) , treatment of BMMs with IFNβ or IFNγ restricted the growth of L . pneumophila ΔflaA ( S2A and S2B Fig ) , which is in line with previous reports [14–16] . Importantly , treatment of BMMs with suboptimal doses of both cytokines alone or in combination resulted in comparable growth inhibition ( S2C Fig ) suggesting that type I and II IFNs might activate an identical intracellular restriction mechanism . Moreover , lack of responsiveness to endogenous IFNβ in Ifnar-/- BMMs resulted in replication of otherwise growth-restricted L . pneumophila wt , and further enhanced the growth of L . pneumophila ΔflaA ( S2D Fig ) . Type I IFNs have previously been reported to induce cell death via e . g . caspase-11-dependent pyroptosis or RIP3-dependent necroptosis [24 , 25] . In order to detect pyroptosis and necroptosis of infected BMMs , we measured cell viability by flow cytometry as a general readout for both types of cell death . The use of GFP-expessing Legionella allowed us to exclusively consider bacteria-harboring cells ( S3A Fig ) . As expected , infection with L . pneumophila wt enhanced cell death compared to L . pneumophila ΔflaA as a consequence of NAIP5/NLRC4-dependent pyroptosis [10–12] ( S3B , S3C and S3E Fig ) . However , cell death in L . pneumophila wt infected cells was not affected by the lack of IFNAR ( S3B and S3C Fig ) , and was only marginally affected by IFNs upon L . pneumophila ΔflaA infection ( S3B–S3D Fig ) . This indicates that IFNs can slightly enhance cell death in L . pneumophila-infected cells independently of the NAIP5 pathway . Moreover , cell death was completely independent of RIP3 ( S3E Fig ) , and RIP3 as well as caspase-11 deficiency did not influence bacterial growth or its restriction by IFNs ( S4A , S4B , S4D and S4E Fig ) . Another important restriction mechanism against intracellular bacteria is the production of nitric oxide ( NO ) via inducible NO synthase ( iNOS ) [26] . However , L . pneumophila wt and ΔflaA replication and IFN-mediated bacterial restriction were comparable in WT and iNOS-deficient macrophages ( S4C and S4F Fig ) . Thus , neither cell death nor production of reactive nitrogen species by iNOS appear to be of critical importance for the IFN-mediated control of L . pneumophila infection . IFNs induce the expression of hundreds of ISGs , several of which possess antimicrobial activities . Since some antimicrobial ISGs have been associated with microbial vacuoles [2] , we hypothesized that IFNs target antibacterial effector proteins to the LCV to restrict infection . In order to test this hypothesis in an unbiased and systematic fashion , we examined the proteome of Legionella-containing vacuoles ( LCVs ) in resting and IFN-activated macrophages 2 h post infection . First , resting macrophages were infected , LCVs were purified as previously described [27] , and LCVs were analyzed by quantitative mass spectrometry . We identified 2307 proteins from the host and 547 from the bacterium in 6 of 6 samples of LCVs from untreated cells ( Fig 3A , S2 Dataset ) . In order to determine the cellular origin of the identified proteins , we performed gene ontology ( GO ) enrichment analysis of the identified host proteins for cellular components ( CC ) . This analysis revealed the highest significance values for the GO terms ‘membrane-bounded organelle’ and ‘intracellular membrane-bounded organelle’ as predicted cellular source of the identified proteins ( S2 Dataset ) . Highest significance values were found for ‘mitochondrion’ , and ‘endoplasmic reticulum’ as predicted child terms of ‘intracellular membrane-bounded organelle’ ( Fig 3B , S2 Dataset ) , reflecting both the ER-derived nature of the LCV as well as the previously reported close association of LCVs with mitochondria [5–8] . Additional GO enrichment analyses of biological processes ( BP ) indicated an enrichment of proteins involved in metabolic as well as transport and localization processes ( S5 Fig , S2 Dataset ) . Macrophage activation by type I or II IFNs did not change the abundance of LCV marker proteins like ARF1 and SEC22b , or ER marker proteins , nor did it lead to an enrichment of endosomal or lysosomal proteins ( Table 1 ) , indicating that neither the LCV establishment is inhibited by IFNs nor do they trigger endo-lysosomal fusion . However , IFNβ or IFNγ treatment led to a significant ( >2-fold ) vacuolar enrichment of 260 or 321 proteins , respectively , and to a decreased vacuolar abundance of 60 or 67 proteins ( Figs 3C , 3D , S6A–S6F , S6B , S6D , S6E and S3 Dataset ) . The direct comparison of LCV proteomes from IFNβ- or IFNγ-activated cells revealed rather minor differences with only a few proteins being differentially regulated ( Figs 3E , S6C and S6F ) . Although we cannot exclude the possibility that some of the ISG product found on the vacuole could potentially only be a contaminant due to the massive up-regulation of ISGs in the cell , the data clearly show that IFNs substantially modify the LCV proteome . Computational analysis of all IFN-directed proteins at the LCV using the STRING database of known and predicted protein-protein interactions ( http://string-db . org ) generated a dense network of protein interactions , with many proteins being involved in immune response processes ( Fig 4 ) . These proteins included molecules contributing to microbial nucleic acid detection ( e . g . TMEM173 , also known as STING ) , ubiquitinylation/ISGylation ( e . g . ISG15 , TRIM25 ) , antimicrobial defense ( e . g . IRGM1 , GBPs ) , and antigen processing/presentation . The comparison of the IFN-dependently LCV-enriched proteins with our transcriptome data ( S1 Dataset ) as well as the INTERFEROME database of ISGs [28] revealed distinct subsets of IFN-regulated proteins ( Fig 4 ) . Whereas several LCV-enriched proteins are also transcriptionally induced by IFNs and thus represent bona fide ISGs , others such as kinases Syk and Lyn or proteins of the proteasomal complex are not directly transcriptionally regulated but appear spatially affected by IFNs . In order to identify new IFN-regulated proteins possessing antibacterial activity against L . pneumophila we decided to examine proteins that were most strongly targeted to the LCV by both IFNs ( Fig 3C and 3D ) for their roles in restricting L . pneumophila growth . BMMs were first transfected with a pool of two siRNAs for each of our candidate molecules as well as IFNAR as a control , and efficient gene silencing was verified ( Fig 5A ) . We found that silencing the expression of IRG1 enhanced replication of L . pneumophila to a similar extent as silencing of IFNAR , whereas knock-down of THEMIS2 , GBP3 , and GBP7 only slightly increased bacterial growth ( Fig 5B ) . We thus decided to focus on IRG1 . Each IRG1-siRNA sequence was also effective in inhibiting their target gene expression ( Fig 6A ) and in increasing bacterial replication when used individually ( Fig 6B ) . To demonstrate that the IRG1-mediated bacterial restriction is also relevant in primary alveolar macrophages we silenced IRG1 expression by siRNA ( Fig 6C ) , which led to a significantly increased L . pneumophila growth in these cells compared to control cells ( Fig 6D ) . We found that IRG1 was strongly induced in BMMs by IFNβ and IFNγ treatment ( S7 Fig ) , confirming its status as a bona fide ISG . Moreover , IRG1 expression was induced upon L . pneumophila infection at both transcriptional and protein levels , which strongly relied on the endogenous type I IFN signaling ( Fig 6E and 6F ) . Strikingly , overexpression of IRG1 in Ifnar-/- cells restricted intravacuolar growth of L . pneumophila , while the percentage of infected cells was not influenced ( Fig 6G–6I ) . Thus , IRG1 is regulated by IFNs , and restricts replication of L . pneumophila within the LCV . In agreement with previous studies in RAW264 . 7 macrophages [29] , IRG1 showed a mitochondrial localization ( Fig 7A and 7B ) . Moreover , super-resolution fluorescence microscopy demonstrated that mitochondria were distributed throughout the cell and closely associated with LCVs ( Fig 7C ) . As viewed by time-lapse fluorescence microscopy , mitochondria moved very dynamically within living cells , although single mitochondria appeared to stay in close proximity to intracellular L . pneumophila for at least 1 h , most likely attached to the LCV membrane ( S8 Fig and S1 Video ) . In order to directly evaluate whether the mitochondria-localized IRG1 associates with the LCV , we visualized homogenized Legionella-infected IRG1-GFP-overexpressing cells by fluorescence microscopy , and found that LCVs are surrounded by IRG1 ( Fig 7D ) . In summary , these data confirm that overexpressed IRG1 localizes to mitochondria and that the latter are in close contact with LCVs . In absence of a specific antibody we could not obtain sufficient staining of endogenous IRG1 . Nevertheless , our data strongly suggest that IFN signaling stimulates up-regulation of endogenous IRG1 within mitochondria , which results in its close association of IRG1 with LCV . IRG1 has recently been described as an enzyme catalyzing the production of itaconic acid , which was found to exert bacteriostatic effects on Mycobacteria and Salmonella in liquid cultures [30] , but the mechanisms that regulate this pathway and its relevance for infections remained incompletely understood . Subsequently , another study found that IRG1 mediates production of mitochondrial ROS ( mROS ) [31] . We found that mROS is produced in Legionella-infected macrophages by a largely IRG1- and IFNAR-independent mechanism ( S9A and S9B Fig ) . In contrast , metabolic analyses of BMMs via gas chromatography-mass spectrometry ( GC-MS ) revealed that IFNs as well as L . pneumophila stimulate the production of itaconic acid ( Fig 8A and 8B ) , whereas gene-silencing of IRG1 strongly reduced its production ( Fig 8C ) . In line with the robust expression of IRG1 upon L . pneumophila infection in vivo ( Fig 8D ) , GC-MS measurements revealed a strong production of itaconic acid also in L . pneumophila-infected mouse lungs ( Fig 8E ) . Notably , IRG1 expression as well as itaconic acid production in vivo were largely dependent on functional IFN signaling ( Fig 8D and 8E ) . Next , we assessed the antibacterial potential of itaconic acid on L . pneumophila . In line with previous findings for M . tuberculosis and S . enterica [30] , we found itaconic acid had an inhibitory effect on L . pneumophila growth in liquid culture ( Fig 8F ) , however , at concentrations previously found insufficient to restrict bacterial growth . Importantly , itaconic acid , but not related organic acids , was capable of killing L . pneumophila as well as multidrug-resistant Gram-positive and -negative isolates ( Fig 8G–8J ) at concentrations that have recently been measured in activated RAW264 . 7 macrophages [30] . We thus conclude that IRG1 restricts L . pneumophila in their LCVs in macrophages through catalyzing the production of the broadly bactericidal metabolite itaconic acid .
IFNs execute antimicrobial functions by stimulating the expression of hundreds of ISGs . Recent large-scale examinations of ISGs have shed light on their antiviral activities [32–38] , and individual ISGs , including immunity-related GTPases and GBPs , are known to localize to microbial vacuoles and to restrict bacterial infection [39–44] . However , the function of IFNs and ISGs during bacterial infections have not been systematically examined , and the molecular mechanism of IFN-mediated restriction of many bacterial infections remains unknown . We therefore globally profiled the effects of type I and II IFNs on the transcriptome and subcellular proteome during L . pneumophila infection , thereby providing an important resource for IFN-mediated effects on basic cellular functions during infection . We demonstrate that both IFNs are master regulators of gene expression . Within macrophages , IFNs induce extensive remodeling of bacterial vacuoles , thereby altering their permissiveness for bacterial growth ( Fig 9 ) . Neither types of IFNs disturbed the establishment of the LCV but targeted several proteins involved in nucleic acid detection , antigen presentation and antibacterial defense to bacterial vacuoles . Furthermore , we demonstrate that IFN-dependent activation of CD11c+ cells ( most likely alveolar macrophages ) is critical for restricting infection in vivo . Interestingly , the comparison of our proteomic data with the INTERFEROME database as well as our transcriptomic data indicates that a large subset of IFN-dependently LCV-enriched proteins is not transcriptionally regulated . This reveals a hitherto unknown mode of action of IFNs , and suggests that IFNs are able to control the spatial distribution of a subset of proteins , potentially via the activation of signaling molecules that control protein recruitment . Mitochondrial proteins account for almost one third of identified LCV proteins , whereas the previously reported proteome of latex bead-phagosomes contained only 3–4% mitochondrial proteins [45] . Mitochondria have long been known to attach to LCVs as well as to other microbe-containing vacuoles [6 , 46 , 47] . The mechanisms underlying mitochondrial attachment , and most importantly , their biological function at the LCV remain , however , largely unknown . L . pneumophila secretes a mitochondrial carrier protein ( LncP ) through its T4SS [48] and might thereby actively recruit the organelles as a source of energy or nutritional metabolites . Alternatively , mitochondria are actively recruited to phagosomes that contain bacteria by a Toll-like receptor ( TLR ) -dependent mechanism [49] . The TLR-mediated recruitment of mitochondria and the concomitant IFN-dependent up-regulation of antimicrobial proteins within this organelle might thus represent a combined strategy of the immune system to counteract intravacuolar pathogens . We demonstrate for the first time that the mitochondrial protein IRG1 localizes in close association with a microbial vacuole , and indicate that IRG1 activity is able to restrict Legionella inside their LCVs . As discussed above , this IRG1 accumulation on LCVs is most likely dependent on IFN-mediated upregulation of IRG1 within mitochondria and Legionella- and/or TLR-dependent recruitment of those organelles to LCVs . IRG1 has recently been identified as an enzyme catalysing the production of itaconic acid following decarboxylation of the tricarboxylic acid cycle metabolite cis-aconitate [30] . Itaconic acid was found to exert bacteriostatic effects on Mycobacteria and Salmonella in liquid bacterial cultures [30] . We demonstrate that itaconic acid production in vivo is entirely dependent on IFN signals . In addition , we report for the first time a directly bactericidal effect of itaconic acid on different bacterial pathogens , which contrasts to the merely bacteriostatic activity recently reported [30] . Such different effects might be reflective of metabolic differences amongst these bacterial species . In M . tuberculosis , itaconic acid is thought to inhibit bacterial growth by inhibiting the glyoxylate shunt [30 , 50] . However , this pathway is believed to be absent in L . pneumophila [51] . We speculate that the bactericidal activity of itaconic acid on L . pneumophila involves the accumulation of toxic propionyl-CoA concentrations following inhibition of isocitrate lyase or methylisocitrate lyase [30 , 52] , or the blocking of other pathways . We further assume that the IRG1/itaconic acid pathway acts in concert with established antibacterial factors such as IRGM1 and GBPs to completely eliminate L . pneumophila in IFN-activated macrophages and mice [15 , 53] . Type II IFN is well-known for its activation of antibacterial immunity to most intravacuolar bacteria , whereas type I IFNs have been shown to either enhance or inhibit those responses [54] . Using an unbiased , quantitative approach we show here that , in principle , both types of IFNs are able to induce most known antibacterial ISGs ( e . g . GBPs , immunity-regulated GTPases , IRG1 ) . One could speculate that the differential roles of type I IFNs in various bacterial infections might be explained by differences in the architecture of the bacterial vacuoles , and the relative contribution of IFN-dependent defense systems versus other intracellular or extracellular immune mechanisms . In conclusion , our study provides for the first time comprehensive insight into the transcriptional and spatial regulations induced by type I and II IFNs that lead to critical modifications in the proteome of bacterial vacuoles , and it identifies a novel IFN-controlled defense pathway against L . pneumophila infection . In the future , therapeutic stimulation of IRG1 or delivery of encapsulated itaconic acid might be useful to fight intracellular and multidrug-resistant bacteria .
The L . pneumophila serogroup 1 strain JR32 , the ΔflaA mutant , JR32 expressing eGFP or DsRed and the culture conditions have been described previously [11 , 55] . ΔflaA mutants expressing eGFP or DsRed have been generated in this study using plasmids published recently [55] . A clinical isolate of methicillin resistant Staphylococcus aureus ( MRSA ) has been obtained from the Charité microbiology department , the multidrug-resistant Acinetobacter baumannii isolate A9703 was kindly provided by Harald Seifert , University Cologne , Germany . S . aureaus and A . baumannii were cultured on LB agar . All animal experiments were approved by institutional ( Charité –Universitätsmedizin Berlin ) and governmental animal welfare committees ( LAGeSo Berlin; approval IDs G0446/08 , G0278/11 , G0440/12 ) . Casp11-/- , Ifnar-/- , Ifngr-/- , Rip3-/- , Ifnar/Ifngr-/- and Nos2-/- mice were on a C57BL/6 background [15 , 56 , 57] . C57BL/6 CD45 . 1 mice and transgenic CD11c-DTR-GFP [58] mice were bred and maintained at the University of Melbourne . All mice used were on C57BL/6J background , 8–10 weeks old and female . Anaesthetized mice were intranasally infected with 1 × 106 L . pneumophila in 40 μl of PBS [15] . Lungs were flushed via the pulmonary artery with sterile saline and homogenized in Trizol ( Life Technologies ) [59] . Homogenized lungs were pooled ( 5 mice per group ) and RNA extraction was carried out according to manufacturer’s instructions . RNA amounts were estimated with a NanoDrop 1000 UV-Vis spectrophotometer ( Kisker ) and RNA integrity was confirmed using an Agilent 2100 Bioanalyzer with a RNA Nano 6000 microfluidics kit ( Agilent Technologies ) . Microarray analysis was performed as dual-color hybridizations . In order to compensate dye-specific effects and to ensure statistically relevant data , color-swap dye-reversal hybridizations were performed [60] . RNA labeling was done with a two-color Quick Amp Labeling Kit according the supplier’s recommendations ( Agilent Technologies ) . In brief , mRNA was reverse transcribed and amplified using an oligo-dT-T7 promoter primer , and labeled with Cyanine 3-CTP or Cyanine 5-CTP . After precipitation , purification , and quantification , 1 . 25 μg of each labeled cRNA was fragmented and hybridized to whole mouse genome 4x44K multipack microarrays ( Design ID 014868 ) according to the manufacturer’s protocol ( Agilent Technologies ) . Scanning of microarrays was performed with 5 μm resolution using a G2565CA high-resolution laser microarray scanner ( Agilent Technologies ) with XDR extended range . Microarray image data were analyzed and extracted with the Image Analysis/Feature Extraction software G2567AA v . A . 10 . 10 . 1 . 1 ( Agilent Technologies ) using default settings and the protocol GE2_1010_Sep10 . The extracted MAGE-ML files were subsequently analyzed with the Rosetta Resolver , Build 7 . 2 . 2 SP1 . 31 ( Rosetta Biosoftware ) . Ratio profiles comprising single hybridizations were combined in an error-weighted fashion to create ratio experiments . A 1 . 5-fold change expression cut-off for ratio experiments was applied together with anti-correlation of ratio profiles , rendering the microarray analysis highly significant ( p < 0 . 01 ) , robust , and reproducible . The data discussed in this publication have been deposited in NCBI's Gene Expression Omnibus and are accessible through GEO Series accession number GSE60085 . Genes identified to be significantly up-regulated upon L . pneumophila infection ( > 2-fold increase , p < 0 . 05 in infected versus PBS treated WT mice; S3 Dataset ) were analyzed for their predicted upstream regulators using the Ingenuity Pathway Analysis ( IPA ) software ( Ingenuity System ) . Only upstream regulators with an activation z-score > 2 ( predicted activators ) were considered and further categorized in respective groups . Chimeras were generated as described recently [61] ( Fig 2A ) . Briefly , CD45 . 1+ mice were lethally irradiated twice with 550 cGy and reconstituted with a 1:1 mix of 1 . 5 × 106 bone marrow cells from C57BL/6 WT or Ifnar/Ifngr-/- and transgenic CD11c-DTR-GFP mice ( all CD45 . 2+ ) . Chimeric mice were allowed to reconstitute for at least 10 weeks . Only those mice that contained < 10% host cells were included in experiments . Depletion of CD11c+ cells was achieved by injection of CD11c-DTR-GFP chimeric mice intraperitoneally three times with 100 ng diphtheria toxin ( Sigma-Aldrich ) on days -2 , +1 and +4 prior to and during infection . Preparation of lung cells for the determination of cell exchange in chimeric mice has been described [62] . The lung cell suspension was labelled with anti-panCD45 ( 30-F11 , eBioscience ) , anti-CD45 . 1 ( A20 , BD Pharmingen ) , anti-CD45 . 2 ( 104 , BD Pharmingen ) , anti-Ly6G ( 1A8 , BD Pharmingen ) , anti-CD11c ( N418 , eBioscience ) anti-MHC-II ( M5/114 . 15 . 2 , eBioscience ) , anti-Siglec-F , ( E50-2440 , BD Pharmingen ) and anti-CD64 ( X54-5/7 . 1 , BD Pharmingen ) . Cells were analyzed on a Becton Dickinson LSRFortessa flow cytometer using FACSDIVA software ( BD Biosciences ) . Bone marrow-derived macrophages ( BMMs ) were prepared from femurs and tibiae . Alveolar macrophages ( AMs ) were isolated by extensive bronchoalveolar lavage and purity was checked by flow cytometry . BMMs and AMs were transfected with control non-silencing or a mix of two gene-specific siRNAs ( S1 Table ) 24 h prior to infection ( Life Technologies ) using HiPerfect ( Qiagen ) , and with EGFP ( pEGFP-N1 , Clontech ) or full-length murine IRG1 ( NM_008392 ) with a carboxy-terminal TurboGFP ( pCMV6-AC-GFP , OriGene ) or Myc-DDK ( pCMV6-Entry , OriGene ) tag 48 h prior to infection using ViaFect ( Promega ) . BMMs and AMs were infected with L . pneumophila wt or ΔflaA , centrifuged at 200 g for 5 min and incubated for the indicated time intervals . Where indicated , cells were incubated either with IFNβ , IFNγ or both 16–18 h prior to and during infection at a concentration of 50 U/ml unless stated otherwise . For intracellular replication assays , cell death and mitochondrial ROS assays , BMMs or AMs were infected with L . pneumophila , centrifuged at 200 g for 5 min and incubated at 37°C for 30 min . Cells were washed with PBS and were further incubated with RPMI with 15% L cell supernatant and 10% FCS ( BMMs ) or 10% FCS ( AMs ) containing 50 μg/ml gentamicin for 1 h in order to kill extracellular bacteria . Total RNA was isolated from BMMs or lung homogenates using the PerfectPure RNA purification system ( 5 Prime ) or Trizol ( Life Technologies ) , respectively , reverse-transcribed using high capacity reverse transcription kit ( Applied Biosystems ) , and quantitative PCR was performed using TaqMan assays ( Life Technologies ) or self-designed primer sets , respectively ( S2 Table ) , on an ABI 7300 instrument . The input was normalized to the average expression of GAPDH and relative expression ( relative quantity , RQ ) of the respective gene in untreated cells or PBS-treated mice was set as 1 . Cells were detached using ice cold PBS containing 2 mM EDTA and stained for mitochondrial ROS ( MitoSOX Red , Life Technologies ) or cell death ( 7-AAD , eBioscience or LIVE/DEAD fixable red dead cell stain , Life Technologies ) . Proportions of mitochondrial ROS producing ( MitoSOX+ ) or dead ( 7AAD+ or LIVE/DEAD+ ) cells were determined in infected ( GFP+ ) and uninfected ( GFP- ) cell populations by flow cytometry ( FACScan , BD or MACSQuant , Miltenyi Biotec ) . Data analysis was done using FlowJo software ( Tree Star ) . For immunoblotting cells were lysed in SDS- and 1% NP40-containing lysis buffer , cleared extracts were separated by SDS-PAGE and SDS-gels were blotted onto Hybond nitrocellulose membranes . Antibodies against IRG1 ( HPA040143 , Sigma-Aldrich ) and actin ( sc-1616 , Santa Cruz ) followed by respective fluorophore-linked secondary antibodies ( Rockland ) were used and blots analyzed using an Odyssey infrared imaging system ( Li-Cor ) . BMMs were seeded onto glass coverslips or high precision glass coverslips for structured illumination microscopy . Where indicated , cells were stained with MitoTracker Orange ( Life Technologies ) approximately 2 h prior infection . Cells were fixed with 3% PFA or aceton/methanol if MitoTracker was used . For homogenization , cells were seeded in 6 well plates , infected and 2 h p . i . washed with PBS , scraped in homogenization buffer ( 20 mM Hepes , 250 mM sucrose , 0 . 5 mM EGTA , pH 7 . 2 ) and homogenized using a ball homogenizer ( Isobiotec ) . Homogenates were centrifuged onto glass coverslips and fixed with 3% PFA . For intracellular staining , cells were permeabilized with 0 . 1% triton x-100 and blocked with 5% FCS . Where indicated cells were stained with an affinity purified rabbit anti-SidC [63] and a mouse anti-Legionella pneumophila ( ab69239 , Abcam ) antibody followed by the respective anti-rabbit Alexa Fluor 633-conjugated ( Molecular Probes ) and anti-mouse DyLight 405-conjugated ( Thermo Scientific ) secondary antibody . Samples were mounted on slides using PermaFluor ( Thermo Scientific ) containing DAPI or Mowiol ( Sigma-Aldrich ) . For confocal laser scanning microscopy samples were examined using a LSM 780 microscope ( objective: Plan Apochromat 63×/1 . 40 oil DIC M27 , Carl-Zeiss ) . Structured illumination microscopy was performed on an ELYRA PS . 1 microscope ( objective: Plan Apochromat 63×/1 . 40 oil DIC M27 , Carl-Zeiss ) . Data sets were acquired with five grating phases and three rotations , post-processed in ZEN ( Carl-Zeiss ) using automatically determined parameters , and colour channels were subsequently aligned based on parameters determined from control measurements with multispectral beads performed with identical instrument settings . For time-lapse confocal microscopy cells were seeded in 8 well μ-slides ( Ibidi ) and stained with MitoTracker Orange ( Life Technologies ) approximately 2 h prior infection . Images were acquired on a LSM 780 microscope ( objective: Plan Apochromat 63×/1 . 40 oil DIC M27 , Carl-Zeiss ) at 37°C/5% CO2 . All images were processed using ZEN ( Carl-Zeiss ) and ImageJ software ( http://imagej . nih . gov/ij/ ) . BMMs were left untreated or treated with 50 U/ml IFNβ or IFNγ for 16–18 h prior to and during infection and subsequently infected with L . pneumophila ΔflaA for 2 h at a m . o . i of 50 . LCVs from BMMs were isolated as described previously for RAW264 . 7 cells and amoeba [27 , 64] . Briefly , cells were washed with PBS , scraped in homogenization buffer ( 20 mM Hepes , 250 mM sucrose , 0 . 5 mM EGTA , pH 7 . 2 ) , homogenized using a ball homogenizer ( Isobiotec ) , and incubated with an anti-SidC antibody [63] followed by a secondary anti-rabbit antibody coupled to magnetic beads ( Miltenyi Biotec ) . The LCVs were separated in a magnetic field and further purified by a Histodenz density gradient centrifugation step . Isolated LCV from 4 IFNβ , 5 IFNγ , and 6 untreated biological replicates were analyzed . LCV samples were solubilized in 1% RapiGest ( Waters ) in 50 mM Tris pH 8 . 0 , reduced with 10 mM tris ( 2-carboxyethyl ) phosphine ( TCEP ) ( Pierce ) , and heated at 70°C for 10 min . After cooling , proteins were alkylated in 10 mM iodoacetamide ( Sigma-Aldrich ) , and alkylation was quenched in 20 mM DTT . Protein concentrations were measured by the EZQ assay ( Life Technologies ) , and 8 μg of protein was digested by trypsin overnight at 30°C , after diluting the Rapigest concentration to 0 . 1% . Rapigest was removed from the sample by acidification to 2% trifluoroacetic acid ( TFA ) and incubation at 37°C for 1 h , followed by centrifugation at 14k g for 30 min . Peptides were then desalted with Microspin C18 solid phase extraction columns ( The Nest Group ) . After drying down , peptides were redissolved in 1% TFA . For each sample , 2 μg of peptides were analyzed on an Orbitrap Velos Pro mass spectrometer coupled to an Ultimate 3000 UHPLC system with a 50 cm EasySpray analytical column ( 75 μm ID , 3 μm C18 ) in conjunction with a Pepmap trapping column ( 100 μm x 2 cm , 5 μm C18 ) ( Thermo-Fisher Scientific ) . Acquisition settings were: lockmass of 445 . 120024 , MS1 with 60 , 000 resolution , top 20 CID MS/MS using Rapid Scan , monoisotopic precursor selection , unassigned charge states and z = 1 rejected , dynamic exclusion of 60s with repeat count 1 . 6 h linear gradients were performed from 3% solvent B to 35% solvent B ( solvent A: 0 . 1% formic acid , solvent B: 80% acetonitrile 0 . 08% formic acid ) with a 30 min washing and re-equilibration step [65] . Protein identification and quantification were performed using MaxQuant Version 1 . 4 . 1 . 2 [66] with the following parameters: stable modification carbamidomethyl ( C ) ; variable modifications of methionine oxidation , and protein N-terminal acetylation , and 2 missed cleavages . Searches were conducted using a Uniprot-Trembl Mus musculus database downloaded May 1 , 2013 , Legionella pneumophila strain Philadelphia 1 downloaded December 4 , 2013 , and common contaminants . Identifications were filtered at a 1% false-discovery rate ( FDR ) at the protein level , accepting a minimum peptide length of 7 . Quantification used only razor and unique peptides , and required a minimum ratio count of 2 . “Re-quantify” and “match between runs” were enabled . The host proteins identified in all six LCV samples from untreated macrophages ( S1 Dataset ) were analyzed for overrepresented cellular components using g:Profiler ( http://biit . cs . ut . ee/gprofiler/ ) [67] with default settings including g:SCS algorithm for multiple testing correction . All overrepresented child terms of the GO term intracellular membrane-bounded organelle ( GO:0043231 , p = 3 . 55 × 10−300 ) were extracted . The total result of GO enrichment analysis for cellular components ( CC ) can be found in S1 Dataset . To identify and visualize biological processes that are overrepresented at LCVs of untreated cells , the same list of proteins was analyzed with BiNGO [68] for Cytoscape [69] using default settings including hypergeometric testing and Benjamini & Hochberg FDR correction . Significance level cut-off was set to < 10−10 ( terms with p-values > 10−10 are depicted if p-value of final child term was < 10−10 ) . The total result of GO enrichment analysis for biological process ( BP ) can be found in S1 Dataset . Proteins identified in IFN-treated ( in 4 of 4 IFNβ or 5 of 5 IFNγ treated samples ) but not in untreated samples ( ≤ 1 of 6 samples; hereafter called “qualitative changers” ) and proteins with significant higher abundance in IFN-treated versus untreated samples ( log2 LFQ intensity ratio ≥ 1 , p < 0 . 05; hereafter called “quantitative changers” ) ( S2 Dataset ) were combined and analyzed for protein-protein interaction networks using STRING database ( http://string-db . org/ ) . The identified network was extracted and loaded into Cytoscape [69] for visualization; only interactions with a minimum STRING combined score of 0 . 400 , which represents the default medium confidence level in STRING , were kept . For identification of subnetworks of overrepresented biological functions , the combined protein list was analyzed by g:Profiler ( http://biit . cs . ut . ee/gprofiler/ ) [67] . Protein lists of overrepresented GO terms were extracted and subnetworks were built using STRING and Cytoscape . To identify proteins within the networks that were also transcriptionally induced by IFNs upon in vivo L . pneumophila infection , the combined list of qualitative and quantitative changing proteins was compared to genes with a >2-fold change ( p < 0 . 05 ) in L . pneumophila infected Ifnar/Ifngr-/- versus WT mice ( S3 Dataset ) . To cross-reference gene names from transcriptome analysis and Uniprot identifier from proteome analysis , both lists were uploaded to STRING and respective output lists were compared against each other . For identification of ISGs the protein list was also compared against the INTERFEROME database [28] . 106 BMMs per well were left untreated , were incubated either with 50 U/ml IFNβ or IFNγ for 16–18 h or were infected with L . pneumophila for 24 h . Where indicated cells were transfected with control non-silencing or a mix of two gene-specific siRNAs as described above 24 h prior to infection . After washing with PBS , metabolism was stopped adding 200 μl cooled 50% MeOH ( -20°C ) and cells were collected by scraping in the MeOH solution . Cells from 6 wells were pooled , 240 μl chloroform were added , samples centrifuged for 10 min at 10k g and supernatant containing polar metabolites was dried under vacuum overnight . For in vivo experiments mice were infected with L . pneumophila wt or left untreated . 2 d p . i . lungs were flushed with sterile PBS , shock frozen in liquid nitrogen and stored at -80°C . Lung tissue was homogenized using a Precellys24 bead homogenizer in chloroform ( 6 mL/g ) , methanol ( 6 mL/g ) , and distilled water ( 4 mL/g ) . Samples were centrifuged for 10 min at 10k g and supernatant containing polar metabolites was dried under vacuum overnight . For GC/MS analysis samples were processed using protocols and machine settings described elsewhere [70] . Data were analyzed using ChromaTOF ( Leco ) and the custom software MetMax [71] . Data were normalized on mean of total area of all analyzed metabolites ( in vitro samples ) or on internal standard ( in vivo samples ) and average amount of itaconic acid in untreated cells or control mice was set as 1 . For growth inhibition bacteria were grown in AYE broth containing indicated amounts of itaconic acid . OD600 was determined over time . For killing assays bacteria were resuspended in PBS and respective amounts of itaconic acid , acetic acid or citric acid were added . Bacteria were incubated at 37°C and plated at indicated time points to assess viability . Data were statistically analyzed using GraphPad Prism software . Groups were compared with two-tailed Mann-Whitney U test or , for multiple-group comparisons with Kruskal-Wallis analysis of variance followed by Dunn’s multiple comparison test . Differences with p < 0 . 05 were considered statistically significant . GEO: GSE60085 .
|
Numerous intracellular bacterial pathogens replicate in specialized vacuoles within macrophages . We systematically study the molecular mechanism and the impact of macrophage-intrinsic antibacterial defense . Using L . pneumophila , an important cause of pneumonia and model organism for intracellular bacteria , we found that type I and II interferons critically modify the proteome of bacterial vacuoles to restrict infection . We identify IRG1 and demonstrate a bactericidal activity of its metabolite itaconic acid on bacteria in their vacuole . Moreover , our study provides evidence for the impact of this cell-autonomous defense pathway in alveolar macrophages to restrict lung infection . We speculate that vacuolar IRG1 or its product itaconic acid could serve as future therapeutic targets to fight intracellular and drug-resistant bacteria .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"blood",
"cells",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"pathology",
"and",
"laboratory",
"medicine",
"vacuoles",
"intracellular",
"pathogens",
"pathogens",
"immunology",
"microbiology",
"animal",
"models",
"model",
"organisms",
"mitochondria",
"bioenergetics",
"cellular",
"structures",
"and",
"organelles",
"bacteria",
"bacterial",
"pathogens",
"research",
"and",
"analysis",
"methods",
"legionella",
"pneumophila",
"white",
"blood",
"cells",
"animal",
"cells",
"legionella",
"medical",
"microbiology",
"alveolar",
"macrophages",
"gene",
"expression",
"microbial",
"pathogens",
"mouse",
"models",
"biochemistry",
"macrophages",
"cell",
"biology",
"genetics",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"energy-producing",
"organelles",
"organisms"
] |
2016
|
IFNs Modify the Proteome of Legionella-Containing Vacuoles and Restrict Infection Via IRG1-Derived Itaconic Acid
|
Although enteroparasites are common causes of diarrheal illness , few studies have been performed among children in Tanzania . This study aimed to investigate the prevalence of Cryptosporidium parvum/hominis , Entamoeba histolytica and Giardia lamblia among young children in Dar es Salaam , Tanzania , and identify risk factors for infection . We performed an unmatched case-control study among children < 2 years of age in Dar es Salaam , recruited from August 2010 to July 2011 . Detection and identification of protozoans were done by PCR techniques on DNA from stool specimens from 701 cases of children admitted due to diarrhea at the three study hospitals , and 558 controls of children with no history of diarrhea during the last month prior to enrollment . The prevalence of C . parvum/hominis was 10 . 4% ( 84 . 7% C . hominis ) , and that of G . lamblia 4 . 6% . E . histolytica was not detected . The prevalence of Cryptosporidium was significantly higher in cases ( 16 . 3% ) than in controls ( 3 . 1%; P < 0 . 001; OR = 6 . 2; 95% CI: 3 . 7–10 . 4 ) . G . lamblia was significantly more prevalent in controls ( 6 . 1% ) than in cases ( 3 . 4%; P = 0 . 027; OR = 1 . 8; 95% CI: 1 . 1–3 . 1 ) . Cryptosporidium infection was found more often in HIV-positive ( 24 . 2% ) than in HIV-negative children ( 3 . 9%; P < 0 . 001; OR = 7 . 9; 95% CI: 3 . 1–20 . 5 ) , and was also associated with rainfall ( P < 0 . 001; OR = 2 . 41; 95% CI: 1 . 5–3 . 8 ) . Among cases , stunted children had significantly higher risk of being infected with Cryptosporidium ( P = 0 . 011; OR = 2 . 12; 95% CI: 1 . 2–3 . 8 ) . G . lamblia infection was more prevalent in the cool season ( P = 0 . 004; OR = 2 . 2; 95% CI: 1 . 3–3 . 8 ) , and more frequent among cases aged > 12 months ( P = 0 . 003; OR = 3 . 5; 95% CI: 1 . 5–7 . 8 ) . Among children aged 7–12 months , those who were breastfed had lower prevalence of G . lamblia infection than those who had been weaned ( P = 0 . 012 ) . Cryptosporidium infection is common among young Tanzanian children with diarrhea , particularly those living with HIV , and infection is more frequent during the rainy season . G . lamblia is frequently implicated in asymptomatic infections , but rarely causes overt diarrheal illness , and its prevalence increases with age .
Diarrheal disease is a leading cause of mortality and morbidity in young children , estimated to cause more than 760 000 annual deaths among children < 5 years of age [1] , with 72% of these deaths occurring in children < 2 years of age [2] . Globally , diarrheal diseases take more lives than malaria and HIV together [3] . While malaria is the leading cause of child deaths in the African region , diarrheal diseases still contribute to more than one tenth of deaths in African children [3] , and sub-Saharan Africa accounts for half of all global childhood deaths from diarrheal diseases [2] . Diarrheal diseases can be caused by various bacteria , viruses and parasites . Among the main infectious diarrheagenic pathogens , Cryptosporidium spp . results in the most deaths among children < 5 years of age [4] . Two other enteric protozoan parasites , Giardia lamblia ( synonymous with G . intestinalis , G . duodenalis ) and Entamoeba histolytica also contribute , but to a lesser extent [5] . The genus Cryptosporidium consists of approximately 20 different species , with C . hominis and C . parvum being the major species infecting humans . Transmission occurs via the fecal-oral route from human and animal reservoirs . In immunocompetent hosts , cryptosporidiosis is usually self-limiting , but in developing countries it contributes to persistent diarrhea in children and is a major enteric pathogen causing chronic diarrhea in HIV-positive patients [6] . G . lamblia is a known cause of diarrheal disease world-wide , but is more frequently encountered in developing countries [6] . It causes the diarrheal illness giardiasis , but can also be asymptomatic [7] . E . histolytica causes amoebiasis , with a wide spectrum of clinical presentations , ranging from asymptomatic infection to diarrhea , amoebic colitis , amoebic dysentery and abscesses in the liver , lungs or brain . It is endemic in several parts of the world . However , while symptomatic disease is rare , the outcome is often severe [8 , 9] . All these three parasites can cause waterborne outbreaks , and also foodborne outbreaks have been reported [6 , 10] . Although varying in designs and settings , other studies from sub-Saharan Africa have found prevalence ranging up to 30 . 5% for Cryptosporidium spp . [11] , 10 . 7% for E . histolytica/ dispar [11] and 60 . 1% for G . lamblia [12] in children < 5 years of age with diarrhea . Few studies of enteroparasites among young children with diarrheal illness have been performed in Tanzania [13–17] , and most of these had limited study populations . Seasonal differences for several pathogens causing diarrheal disease has been reported [11 , 13 , 18] . The objectives of the present study were to investigate the prevalence of C . parvum/ hominis , E . histolytica and G . lamblia among young children in Dar es Salaam , Tanzania , and to identify risk factors for infection . The results of this study may contribute useful information about prevalence and risk factors for these intestinal parasites in Tanzania .
The study was approved by the Senate Research and Publication Committee of Muhimbili University of Health and Allied Sciences in Dar es Salaam , Tanzania , by the Regional Committee for Medical and Health Research Ethics ( REK ) in Norway , and by the respective hospital authorities at the three study hospitals . Written informed consent was obtained from the parents or guardian on behalf of all the children enrolled in the study . The study population and data collection have previously been described [19] . Briefly , this prospective study was performed between August 2010 and July 2011 , in Dar es Salaam , Tanzania , covering both the dry and the wet seasons . A total of 1266 children < 2 years of age were recruited . Diarrhea was defined as three or more watery stools within 24 hours . An episode of diarrhea was considered over when two consecutive days pass without diarrhea . An episode of acute diarrhea was defined as duration between 24 hours and less than 14 days . Persistent diarrhea was defined as diarrhea for 14 days or more . Cases ( N = 705 ) were children admitted due to diarrhea at one of the three major hospitals in Dar es Salaam; Muhimbili National Hospital , Amana and Temeke Municipal district hospitals . Controls ( N = 561 ) were children with no history of diarrhea during the last month prior to enrollment . A standardized questionnaire and patient files were used for collection of demographic and clinical information . Weight for age ( WAZ ) , length for age ( LAZ ) and weight for length ( WLZ ) Z-scores were calculated using EPI Info ( USD , Inc . , Stone Mountain , GA ) . Children were categorized to have normal nutritional status , mild or severe malnutrition using Z-scores according to WHO criteria . Meteorological data for the region of Dar es Salaam for each month of the study period were collected from Global Historical Climatology Network ( GHCND ) Monthly Summaries database , available at http://www . ncdc . noaa . gov . The rainy season was defined as the months with the heaviest rainfall in mm precipitation; October—December and March—May . The dry season was defined as the months with least rainfall in mm precipitation; August–September , January–February and June–July . The hot season was defined as the months with the highest mean temperature; October–March . The cool season was defined as the months with the lowest mean temperature; August—September and April–July . One stool specimen from each child was collected on inclusion in the study , and frozen at -70°C on the day of collection . For extraction of DNA , 50 mg of the stool sample was mixed 1:10 with Bacterial Lysis Buffer ( Roche Diagnostics , Mannheim , Germany ) , and centrifuged at 13 000 x g for 3 min . DNA was extracted from 200 μl supernatant using the Magna Pure LC High Performance Total Nucleic Acid Isolation Kit ( Roche Applied Science , Mannheim , Germany ) . DNA was eluted and stored at—70°C until PCR analysis . A multiplex real-time PCR for C . parvum/ hominis , E . histolytica and G . lamblia , using Phocid Herpes Virus 1 ( PhHV–1 ) as an internal control , was performed with previously published primers and probes , with some changes for the labelling . All oligonucleotides used are listed in Table 1 . They were purchased from Applied Biosystems , Cheshire , UK ( primers and probes for C . parvum/ hominis , E . histolytica ) , and from TIB MOLBIOL , Berlin , Germany ( primers and probes for G . lamblia and PhHV–1 ) . Each PCR test was performed in a 25 μl reaction mixture . The reaction mixture included: 1 x HotStarTaq Plus Master Mix ( Qiagen , Hilden , Germany ) , 0 . 5 μg/ μl BSA ( New England Biolabs , Inc . , Ipswich , MA ) additional 3 . 5 mM of MgCl2 , concentrations of primers and probes as previously published [20] , and water . One μl of PhHV–1 ( diluted to give a Cq-value of approximately 32 ) and 4 μl of DNA sample were added to the reaction mixture . The fourplex real-time PCR assay was performed using a LightCycler 480 Instrument II ( Roche Diagnostics ) , with cycling conditions as follows: 95°C for 5 min , followed by 45 cycles at 95°C for 15 s , 60°C for 30 s and 72°C for 30 s each , and then cooled to 40°C for 30 s . All samples were run on LightCycler 480 Multiwell Plate 96 , white ( Roche ) , and sealed with LightCycler 480 Sealing Foil ( Roche ) . Each run included duplicate of a positive mixed control and multiple no-template controls . Dilution series , five-fold or ten-fold , of DNA extracted from each of the four pathogens were used to make a standard curve to determine the efficiency of the PCR . The PCR was repeated for samples with weak positive or uncertain results . A unidirectional workflow pre- to post-PCR was enforced , and preparation of PCR reaction mixture , DNA preparations and PCR were carried out in facilities physically separate from each other . To identify the Cryptosporidium isolates as C . parvum or C . hominis , primers and probes as described by Hadfield et al . and Lange et al . , and purchased from TIB MOLBIOL , were applied [21 , 22] . LNA converted probes were used and some changes performed for the labelling , see Table 1 . Each reaction contained 1x of LightCycler FastStart DNA Master HybProbe ( Roche ) , 0 . 5 μM of each of the five primers , 0 . 2 μM of each of the three probes , additional 1 . 5 mM of MgCl2 , 5 μl of template and water adjusted to a total volume of 20 μl . The triplex real-time PCR assay was performed using the LightCycler 480 Instrument II ( Roche Diagnostics ) , with cycling conditions as follows: 95°C for 10 min , followed by 55 cycles at 95°C for 15 s , 60°C for 30 s and 72°C for 1 min each , and then cooled to 40°C for 30 s . All samples were run on LightCycler 480 Multiwell Plate 96 , white ( Roche ) and sealed with LightCycler 480 Sealing Foil ( Roche ) . Each run included positive controls and multiple no-template controls . For samples that were negative , had an uncertain or very weak positive result , the PCR was repeated without the genus-specific ( SSU rRNA ) primers and probe . To identify the G . lamblia isolates as assemblages A or B , a nested-PCR method targeting the triosephosphate isomerase ( TPI ) gene as described by Sulaiman et al . , but with doubled concentration of MgCl2 , was used [23] . PCR cycling conditions for the first PCR were 95°C for 5 min , followed by 35 cycles at 95°C for 45 s , 50°C for 45 s and 72°C for 1 min each , and a final extension at 72°C for 7 min . For the second PCR , cycling conditions were 95°C for 5 min , followed by 40 cycles at 95°C for 30 s , 59°C for 30 s and 72°C for 15 s each , and a final extension at 72°C for 1 min . Primers were purchased from TIB MOLBIOL . PCR products were analyzed by gel electrophoresis and both strands sequenced using BigDye Terminator v1 . 1 Cycle Sequencing Kit ( Applied Biosystems , Foster City , CA , USA ) and an ABI PRISM 3730 DNA Analyzer ( Applied Biosystems ) . A consensus sequence was created for each PCR-product and aligned with reference sequences . PCR was repeated for negative samples , and sequencing repeated for inconclusive results . Univariate analysis was performed using Chi square test to compare proportions . For comparison of continuous variables , including meteorological data ( monthly median rainfall and monthly median temperatures ) , we used two-sample Wilcoxon rank-sum ( Mann-Whitney U ) test , since the data did not display a normalized distribution . A P-value ≤ 0 . 05 was considered statistically significant . Multivariate analysis of characteristic features for infection with Cryptosporidium and G . lamblia included the following nine variables; sex , age , place of residence , parent level of education , duration of diarrhea , hydration status , underweight , stunting and wasting . In multivariate analysis of characteristic features for carriage among controls , we included the same factors except for hydration status and duration of diarrhea . Statistical analysis was performed using Stata 13 ( Stata Corp , College Station , TX , USA ) .
DNA for PCR testing was available for 1259 patients; 701 cases and 558 controls . DNA for PCR testing was insufficient and not available for 4 of the cases and 3 of the controls , and these children were omitted in further analyzes . The distribution of children from each of the study sites was 639 from Ilala , 373 from Kinondoni and 247 from Temeke Municipal district hospital . Of these , 523 were females and 736 were males . The age distribution was as follows: 322 children between 0 to 6 months , 558 children between 7 to 12 months , 248 children between 13 to 18 months , and 131 children between 19 to 24 months . HIV testing results were available for 420 of the children , of whom 33 had a positive test result and 387 had a negative test result . The primers and probes have been extensively tested previously [12 , 20 , 24] , but to assure quality , we tested the assay with known positive samples , for cross-reactivity against other pathogens in the assay , for detection of pathogens in mixed samples , and for efficiency for each target . The multiplex PCR assay resulted in amplification curves for the correct targets as expected . No cross-reactivity was detected , and crosstalk was corrected by applying color compensation . The efficiency ( using the formula E = 10−1/slope– 1 ) and Error value ( E ) of the assay for each of the four different targets were: 99 . 7% and E = 0 . 013 for C . parvum/ hominis , 97 . 9% and E = 0 . 008 for E . histolytica , 98 . 5% and E = 0 . 004 for G . lamblia and 96 . 4% and E = 0 . 018 for PhHV–1 . No amplification of any of the no-template controls was detected . The overall prevalence of protozoans in the study population was 14 . 9% ( 187/1259 ) , of which 19 . 7% ( 138/701 ) of the cases and 8 . 8% ( 49/558 ) of the controls tested positive for one or two of the three protozoans . The prevalence of C . parvum/ hominis was significantly higher in cases ( 16 . 3% , 114/701 ) than in controls ( 3 . 1% , 17/558; P < 0 . 001; OR = 6 . 2; 95% CI: 3 . 7–10 . 4 ) . The prevalence of G . lamblia was significantly higher in controls ( 6 . 1% , 34/558 ) than in cases ( 3 . 4% , 24/701; P = 0 . 027; OR = 1 . 8; 95% CI: 1 . 1–3 . 1 ) . Two samples were positive for both Cryptosporidium and G . lamblia , but all samples were negative for E . histolytica . Species identification of the C . parvum/ hominis positive samples resulted in 10 C . parvum samples ( 7 . 6% ) and 111 C . hominis samples ( 84 . 7% ) . For ten of the samples it was not possible to identify whether they were C . parvum or C . hominis . A PCR product that could be sequenced was only achieved for 13 of the 58 G . lamblia positive samples . Of these 13 PCR products , only 8 gave sequences that could be analyzed , of which 4 was identified as assemblage A and 4 as assemblage B . Among study participants with known HIV-status , including both cases and controls , Cryptosporidium infection was significantly more prevalent in HIV-positive ( 24 . 2% , 8/33 ) than in HIV-negative ( 3 . 9% , 15/387 ) children in univariate analysis ( P < 0 . 001; OR = 7 . 9; 95% CI: 3 . 1–20 . 5 ) . In multivariate analysis of this same part of the study population , including both HIV- and Cryptosporidium infection , HIV-positive status was still significantly associated with Cryptosporidium infection ( P = 0 . 001; OR = 5 . 6; 95% CI: 2 . 1–15 . 3 ) , while stunting was not ( P = 0 . 07 ) . All the ten C . parvum positive samples were from HIV-negative children . The characteristics of cases and controls tested for C . parvum/ hominis are shown in Table 2 . When we analyzed cases and controls separately , multivariate analysis of cases showed that stunting was the only characteristic significantly associated with Cryptosporidium infection ( P = 0 . 007; OR = 2 . 12; 95% CI: 1 . 2–3 . 8 ) . The prevalence of Cryptosporidium was higher in the rainy months ( 12 . 9% , 105/812 , median rainfall 105 mm per month ) than in the dry months ( 5 . 8% , 26/447 , median rainfall 78 mm rainfall per month ( P <0 . 001; Fig 1 ) . There was no significant difference in prevalence of Cryptosporidium between the cool months ( 8 . 9% , 53/594 , median temperature 27°C ) and the hot months ( 11 . 7% , 78/605 , median temperature 27°C; P = 0 . 92 ) . The median age for both Cryptosporidium infected and uninfected children was 10 months ( P = 0 . 399 ) . The prevalence of Cryptosporidium in different age groups is illustrated in Fig 2A . Information on breastfeeding was only available for the age group 7–12 months , in which data were collected for 475/ 558 children ( both cases and controls ) . There was no difference in prevalence of Cryptosporidium infection between children who were breastfed or those who were not , neither among cases , controls or the total study population . G . lamblia infection was more prevalent in the cool months ( 6 . 4% , 38/594 ) , than in the hot months ( 3 . 0% , 20/665; P = 0 . 004; OR = 2 . 2; 95% CI: 1 . 3–3 . 8 ) , but with a prevalence of 4 . 2% ( 34/812 ) in rainy season and 5 . 4% ( 24/447 ) in dry season , it was not significantly affected by rainfall ( P = 0 . 338 ) . As seen in Table 3 and Fig 2B , the prevalence of G . lamblia infection in controls increased with age , and among cases age > 12 months was significantly associated with a higher G . lamblia prevalence ( P = 0 . 003; OR = 3 . 5; 95% CI: 1 . 5–7 . 8 ) . Among children < 12 months , G . lamblia infection was more prevalent in controls than in cases , ( P = 0 . 02 ) , but this association was not significant in children > 12 months . G . lamblia infection was significantly less frequent in children on breastfeeding ( 1 . 9% , 7/368 ) than those not on breastfeeding ( 6 . 5% , 7/107; P = 0 . 012 ) . Among the cases in this same age group , breastfeeding was significantly associated with lower prevalence of G . lamblia infection , 0 . 4% ( 1/225 ) , versus 6 . 0% ( 3/50; P = 0 . 003 ) . Among the controls , the prevalence of G . lamblia infection did not differ significantly among children who were breastfed ( 4 . 2% , 6/143 ) and those not breastfed ( 7 . 0% , 4/57; P = 0 . 409 ) . The median age for children with G . lamblia infection was higher than for those without ( median 12 months vs . 10 months; P = 0 . 001 ) . Characteristics for cases and controls that had a G . lamblia infection are shown in Table 3 .
Protozoans such as Cryptosporidium , E . histolytica and G . lamblia are all common causes of diarrheal illness worldwide , particularly in children . In this case-control study we targeted young children in Dar es Salaam , Tanzania , and found overall a quite high prevalence of these intestinal parasites . The prevalence of Cryptosporidium infection is comparable to that found in children in other studies in sub-Saharan Africa , both in the Global Enteric Multicenter Study ( GEMS ) , and in the study of Mbae et al . in Kenya [5 , 11] . The higher prevalence in cases than in controls concurs with findings from Kenya and supports the notion that the parasite causes symptomatic diarrhea [11] . In contrast , the study by Vargas et al . found that only one out of 451 hospitalized children < 5 years of age in Kilombero district in Tanzania had infection with Cryptosporidium [13] . This could be due to differences in methodology , using microscopy which is generally known to be less sensitive , although regional differences may exists . Cryptosporidiosis is often linked to impaired immunity , particularly HIV-associated immunosuppression , hence many of the studies on Cryptosporidium prevalence have been performed on HIV-positive patients , most often including adults , but also children [25] . In our study we found an association between Cryptosporidium infection and HIV-status in small children . Although the HIV-status was only known for approximately one third of the children , HIV-positive children were almost eight times more likely to have Cryptosporidium than those who were HIV-negative . The study among Kenyan children also found this association , but with an odds ratio of 3 . 1 [11] . A study of Ugandan children with persistent diarrhea , found that HIV-positive children were 18 times more likely to have Cryptosporidium than those who were HIV-negative [26] . However , Cryptosporidium should not be ignored as a cause of diarrhea in small children not known to be HIV-positive , as the GEMS-study found that it was an important pathogen at all sites regardless of HIV-prevalence , and the second most common pathogen causing diarrhea in infants [5] . The interaction between diarrhea and malnutrition is complex and multifactorial [2 , 27] . In univariate analysis , we found persistent diarrhea to be significantly more prevalent among cases infected with Cryptosporidium , and this concurs with other studies reporting an association between cryptosporidiosis and prolongation of diarrhea [26 , 28] . Stunting affects millions of children in developing countries [29] . Persistent diarrhea increases the risk of stunting . Molloy et al . found association between stunting and Cryptosporidium infection among Nigerian children [30] , and Yones et al . found the same association among Egyptian children [31] , which supports our findings that stunted children had significantly higher risk of being infected with Cryptosporidium . However , any causal relationship between stunting and Cryptosporidium infection could not be established in the current study . Indeed , when analyzing the part of the study population with known HIV-status , cryptosporidiosis was only associated with HIV-positive status , and not with stunting . This should however be interpreted with caution , as the HIV-status was known for only a limited number of the children infected with Cryptosporidium . Species identification of the isolates showed that the majority were C . hominis , while only 7 . 6% were C . parvum . Ten remaining isolates could not be identified to the species level , likely due to low amount of target . The prevalence of C . hominis and C . parvum varies in different parts of the world [25] , and risk factors are also reported to differ [25 , 30 , 32] . A predominance of C . hominis , with a ratio between C . hominis and C . parvum not much different from what we report , has also been reported from pediatric populations in other developing countries and C . hominis seems to be the dominating species in sub-Saharan Africa [26 , 33–36] . Our finding of C . hominis predominating contrasts a previous study from the same area , where only C . parvum was reported , with a prevalence of 18 . 9% in children < 5 years of age [14] . In that study a rapid test was used for detecting Cryptosporidium , for which the manufacturer claims it detects C . parvum . However , this rapid test also detects C . hominis , thus some C . parvum isolates in that study may have been misclassified isolates of C . hominis [37 , 38] . The probe we used in the multiplex-PCR only target C . hominis and C . parvum , hence other species , like C . meleagridis , C . felis and C . canis , which are also known to cause infection in humans [39] , could not be detected . Cryptosporidium infection was significantly more prevalent in the rainy season than in the dry season , and this is supported by previous studies [11 , 18 , 40] . The majority of Cryptosporidium positive samples were from children younger than one year of age , but age was not a significant factor . This might not be very surprising since all the study participants were below 2 years of age , and other studies which also included older children have reported a higher prevalence of infection among younger children , those below 2 years of age in particular [5 , 11 , 34] . The prevalence of G . lamblia in different regions , including sub-Saharan Africa , shows large differences [11–13 , 15 , 18 , 31] . The prevalence in our study population ( 4 . 6% ) was higher than reported in another study from the same area ( 1 . 9% ) [14] . However , considering the different study population ( children with diarrhea only ) , and use of a less sensitive detection method ( rapid test ) , the prevalence may not be significantly different from that of 3 . 4% found among the cases in our study . Several studies have reported that G . lamblia infection shows seasonality , with a higher prevalence during the rainy season [13 , 18 , 41] . We did not find this association with rainfall in our study . In contrast , we found a significantly higher prevalence of G . lamblia infection in the cool season than in the hot season , though with a low odds ratio . In a study of children in Thailand , Wongstitwilairoong et al . reports a higher prevalence of intestinal parasites in the cool season than in the hot season , however the authors did not provide data for G . lamblia alone [42] . Since most studies relate seasonality with rainfall and not temperature , there is some uncertainty whether any effect of temperature was considered . The prevalence of G . lamblia was significantly higher in controls than in cases . This is in agreement with other studies in pediatric populations [5 , 11 , 15 , 43] . The controls did not have diarrhea for the last month prior to enrollment , and hence were asymptomatic carriers . In our study we did not see any statistical difference in nutritional status between infected cases and controls , although the prevalence of underweight and stunting was slightly higher among cases , while the prevalence of wasting was almost equal . Several studies report that the prevalence of G . lamblia infection in children increases with age [11 , 12 , 43] . Although our study included children within a limited age range , age > 1 year was significantly associated with a higher prevalence among cases , and the prevalence also appeared to increase with age among controls , though this relationship was not statistically significant . Information on breastfeeding was available for the majority of the children aged 7–12 months . With a significantly lower prevalence of G . lamblia infection among those on breastfeeding , as well as a significantly lower prevalence among the cases in this age group , breastfeeding seems to protect children with diarrhea from symptomatic G . lamblia infection . The fraction of G . lamblia infection was higher among controls regardless of being breastfed or not , suggesting that breastfeeding does not prevent asymptomatic infection . However , Mahmud et al . found that breastfeeding prevents both symptomatic and asymptomatic G . lamblia infection among children < 1 year in Egypt [44] . Ignatus et al . also found a protective effect of breast-milk in a study of Rwandan children , but did not relate to symptomatic or asymptomatic infection [12] . Identification of the isolates as assemblages A or B using the TPI gene were , after several attempts , unfortunately only achievable for 8 of the 58 isolates , showing equal prevalence of these two assemblages . With this low outcome , conclusions related to type cannot be drawn . The multiplex Real-Time PCR , targeting a multicopy gene , showed high Cq-values for the majority of isolates . The TPI gene used for typing is a single copy gene , and this could contribute to the low sensitivity . Other explanation could be sequence variability , leading to primer mismatches . In this study population we did not find any E . histolytica . Studies from sub-Saharan countries have reported different prevalence of E . histolytica [11 , 13 , 45 , 46] . However , several of these studies may overestimate the prevalence as they depended on microscopy , which cannot distinguish between E . histolytica and the non-pathogenic E . dispar . A study from the Kilimanjaro district , Tanzania , reported an E . histolytica prevalence of 0 . 8% [45] . Prevalence of E . histolytica/dispar has been reported to increase with age [11 , 45] , and this agrees with the low prevalence found in young children in our study , as well as other studies , like the study by Nesbitt et al . , and the probably very low prevalence found by Vargas et al . . The GEMS-study did not report on any E . histolytica among children < 2 years of age from the sub-Saharan African countries , neither did Krumkamp et al . in a study of children 0–13 years from Ghana , which also supports our results [5 , 43] . This is the first study from Tanzania reporting on the prevalence of protozoans in a large study population of children < 2 years of age , with and without diarrhea , and not many exist from other sub-Saharan countries . The seemingly protective effect of G . lamblia in healthy children needs further elucidation , as it could give important knowledge into the potential immunomodulatory host-pathogen interaction of this microbe . Further studies should also include other causes of diarrhea , such as other parasites , bacteria and viruses , including even larger study groups and older children , in order to broaden our understanding of childhood diarrhea .
|
Diarrheal diseases are a leading cause of disease and deaths among young children . In Africa they contribute to more than one tenth of childhood deaths . Parasites like Cryptosporidium , Entamoeba histolytica and Giardia lamblia are all common causes of diarrheal illness , but there are few studies on these enteroparasites among Tanzanian children . In this case-control study , we included 701 cases and 558 controls , all < 2 years of age , in Dar es Salaam , Tanzania . We assessed the prevalence of C . parvum/hominis , E . histolytica and G . lamblia by PCR , and the association with potential risk factors such as demographic data , clinical symptoms , HIV status and seasonality . One or more parasites were found in 14 . 9% of the samples . C . parvum/ hominis and G . lamblia were found in 10 . 4% and 4 . 6% , respectively , while E . histolytica was not found in any of the samples . The prevalence of Cryptosporidium was high , particularly in children with HIV , and its prevalence increased during the rainy season . Among cases , Cryptosporidium was found more frequently in stunted children , although any causal association could not be established in the current study . G . lamblia was more often implicated in asymptomatic infections than in overt diarrheal illness . The prevalence of G . lamblia increased with age , and breastfeeding seemed to protect the children from G . lamblia . This study presents relevant information about the prevalence and clinical characteristics of these intestinal parasites in Tanzanian children .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[] |
2015
|
Prevalence of Cryptosporidium parvum/hominis, Entamoeba histolytica and Giardia lamblia among Young Children with and without Diarrhea in Dar es Salaam, Tanzania
|
Phosphoribosyl pyrophosphate synthetase ( PRPS ) is a rate-limiting enzyme whose function is important for the biosynthesis of purines , pyrimidines , and pyridines . Importantly , while missense mutations of PRPS1 have been identified in neurological disorders such as Arts syndrome , how they contribute to neuropathogenesis is still unclear . We identified the Drosophila ortholog of PRPS ( dPRPS ) as a direct target of RB/E2F in Drosophila , a vital cell cycle regulator , and engineered dPRPS alleles carrying patient-derived mutations . Interestingly , while they are able to develop normally , dPRPS mutant flies have a shortened lifespan and locomotive defects , common phenotypes associated with neurodegeneration . Careful analysis of the fat body revealed that patient-derived PRPS mutations result in profound defects in lipolysis , macroautophagy , and lysosome function . Significantly , we show evidence that the nervous system of dPRPS mutant flies is affected by these defects . Overall , we uncovered an unexpected link between nucleotide metabolism and autophagy/lysosome function , providing a possible mechanism by which PRPS-dysfunction contributes to neurological disorders .
Phosphoribosyl pyrophosphate synthetase ( PRPS ) is a rate-limiting enzyme in the biosynthesis of purine , pyrimidine , and , pyridine nucleotides ( Fig 1A ) . Purine and pyrimidine nucleotides are the building blocks of RNA and DNA while pyridine nucleotides , such as NAD+ and NADP , are important co-factors in many enzymatic reactions . PRPS produces phosphoribosyl pyrophosphate ( PRPP ) , a common precursor of the five-carbon sugar subunit of nucleotides [1] . The essential role of PRPS in nucleotide metabolism is illustrated through its conservation among all free-living organisms , ranging from E . coli to humans . In humans , three PRPS orthologs exists . Notably , PRPS1 mutations were identified in a number of X-linked neurological disorders: Arts syndrome , Charcot-Marie-Tooth disease ( CMT ) , and nonsyndromic sensorineural deafness [2 , 3] . All PRPS1 mutations identified from patients are missense mutations affecting enzymatic activity to varying degrees . Arts syndrome is the most severe form of PRPS1 disease , and is characterized by infant death , sensorineural hearing loss , intellectual disability , hypotonia , and ataxia . The fact that no nonsense mutations were identified suggests that PRPS1 is essential for embryonic development and patients with missense mutations retain a certain level of PRPS1 function . Knockout mutants of PRPS orthologs have been generated in a number of animal models . PRPS1 knockout mice were generated as a part of a high-throughput screening of mouse genes important for skeletal muscle development [4] . Not surprisingly , PRPS1 was identified as an X-linked gene required for animal viability , supporting the notion that the gene is essential for embryonic development . In contrast , PRPS2 knockout mice are viable and fertile with no discernable developmental defects . This suggests that other mouse PRPS orthologs , PRPS1 and PRPS1L1 , can compensate for the loss of PRPS2 [5] . Interestingly , while PRPS2 is non-essential for development , PRPS2 knockout mice are resistant to Eμ–Myc-driven cancer development , suggesting that PRPS2 is specifically required for tumorigenesis [5] . A zebrafish model of PRPS deficiency has also been recently generated [6] . The mutant animals of the two zebrafish PRPS orthologs , PRPS1a and PRPS1b , fail to properly develop and show some phenotypic similarities to human PRPS1-associated diseases . Notably , only null alleles of PRPS were generated in both mouse and zebrafish models and the biological consequence of patient-derived PRPS1 mutations have not been directly examined . Retinoblastoma ( RB ) family proteins are evolutionarily conserved regulators of the cell cycle [7 , 8] . They are best known for their ability to control S-phase entry by directly binding and regulating E2F family transcription factors . Cell cycle-dependent phosphorylation of RB ensures the timely activation of E2F , whose target genes are required for proper cell cycle progression [8] . Notably , RB is also involved in other cellular processes including cell metabolism [9] . For instance , studies in RB mutant flies and mice have shown that the loss of RB alters glutamate catabolism to compensate for an increased demand for DNA synthesis [9 , 10] . Moreover , a recent study using a proteomics appoach demonstrated that RB inactivation profoundly affects protein levels that are associated with mitochondrial function [11] . While theses studies clearly demonstrated metabolic reprogramming in RB-deficient cells , genes and pathways that are crucial for these changes are yet to be determined . Studies in mammals suggest that PRPS activity is regulated by signals that control cellular growth . In addition to being regulated by Myc [5] , the enzymatic activity of PRPS is regulated by 5′ AMP-activated protein kinase ( AMPK ) [12] . The hexameric form of PRPS , the active configuration of the enzyme , is converted to inactive monomers by AMPK phosphorylation . This study demonstrated that AMPK , which senses energy stress , directly regulates the rate of nucleotide synthesis by controlling PRPS activity . Interestingly , a recent study in Drosophila using DNA adenine methyltransferase identification ( DamID ) technique identified CG6767 , which codes the only Drosophila PRPS ( dPRPS ) , as one of the Capicua ( Cic ) targets [13] . Cic is a transcriptional repressor downstream of the EGFR/Ras pathway , which our lab previously identified as an important determinant of proliferation and survival of rbf1 mutant cells , a Drosophila ortholog of RB [14] . In addition , the CG6767 locus was identified as a locus occupied by a Drosophila E2F , dE2F1 , in S2 cells [15] . Given its rate-limiting function in nucleotide metabolism and possible role downstream of RBF1/dE2F1 , we sought to investigate the in vivo function of dPRPS . Since dPRPS has 89% protein sequence identity with the mammalian PRPS1 ( S1A Fig ) and the in vivo consequence of the patient-derived mutations has not been directly tested , we generated dPRPS alleles that carry mutations identified from Arts syndrome , dPRPSQ165P and dPRPSR228W [2 , 16] . Interestingly , while these flies are viable and fertile , they are highly sensitive to nutrient deprivation . We discovered that their susceptibility to starvation is in part caused by failure to mobilize their lipid reserves due to profound defects in autophagy and lysosome function . Further analysis also revealed that dPRPSQ165P and dPRPSR228W have defects during cellular response to oxidative stress and accumulate lipid droplets and protein aggregates in the brain . Our findings revealed an unexpected link between nucleotide metabolism and autophagy/lysosome function and provide a possible explanation by which PRPS1 dysfunction results in neurological disorders .
We first tested if dPRPS expression is indeed regulated by Cic and RBF1 . For this task , we used a protein trap allele of dPRPS , wherein a GFP coding sequence is inserted in-frame at the N-terminal region of the protein ( S1B Fig ) . Comparing the expression of GFP-trap in eye imaginal discs containing cic mutant clones confirmed that dPRPS expression is normally repressed by Cic ( Fig 1B ) . In addition , a similar experiment in eye imaginal discs containing rbf1 mutant clones also showed that RBF1 normally represses dPRPS expression ( Fig 1C ) . RT-qPCR using RNA isolated from eye imaginal discs mostly composed of mutant cells of cic and rbf1 further demonstrated that dPRPS expression is indeed regulated by Cic and RBF1 ( Fig 1D and 1E ) . We also confirmed that dPRPS is a direct RBF1/dE2F1 target by chromatin immunoprecipitation ( ChIP ) , which showed a recruitment of RBF1 and dE2F1 to the promoter region of CG6767 ( Fig 1F ) . Taken together , our results demonstrate that dPRPS is a Cic- and RBF1-regulated gene and suggest that PRPS may link nucleotide metabolism to the rate of proliferation . Mi09951 is a publicly available insertional dPRPS allele that disrupts splicing between the exon 1 and 2 of dPRPS and places a polyadenylation site in the intron ( S1B Fig ) . As a consequence , the expression of dPRPS exons downstream of exon 1 is severely disrupted in Mi09951 ( S1C Fig ) . Trans-heterozygous between a deficiency line covering the dPRPS locus and the Mi09951 or GFP-Trap allele resulted in early larval lethality ( S1D Fig ) , making them difficult to analyze . Therefore , to better understand the in vivo function of PRPS and to directly assess the physiological consequence of patient-derived mutations , we engineered two dPRPS alleles that carry the mutations identified from Arts syndrome via CRISPR/Cas9: dPRPSQ165P and dPRPSR228W ( S2A Fig and Fig 2A ) . These patient-derived mutations did not largely affect dPRPS transcript levels ( S2B Fig ) and unlike Mi09951 or GFP-Trap mutants , dPRPSQ165P and dPRPSR228W flies are viable and fertile with no discernable developmental defects . In addition , developmental timing of dPRPSQ165P and dPRPSR228W is not affected and the size of wings and eyes of adult flies are normal ( S2C–S2E Fig ) . However , enzymatic function of dPRPS is compromised in both mutants , validated by measuring the relative ATP levels in control and dPRPS mutant adult flies ( Fig 2B ) . Compromised dPRPS function in dPRPSQ165P and dPRPSR228W flies was further demonstrated by their ability to suppress a Ras-induced phenotype . Since Cic is a crucial factor in regulating proliferation downstream of the EGFR/Ras pathway , we reasoned that dPRPS activity may be required for the Ras-induced tumor phenotype . To test this hypothesis , both mutations were introduced to a genetic background overexpressing an activated form of Ras ( RasV12 ) in the eye , which results in a characteristic hyperplastic overgrowth [17] . Strikingly , the RasV12-induced phenotype is strongly suppressed in the dPRPSQ165P or dPRPSR228W mutant background , indicating that dPRPS is required for Ras-induced hyperplastic overgrowth ( Fig 2C ) . Molecular markers such as phospho-Histone H3 and cleaved caspase revealed that dPRPS is required by RasV12 to overcome developmentally-regulated cell cycle arrest and apoptosis ( S2F and S2G Fig ) . Overall , these results demonstrate that dPRPSQ165P and dPRPSR228W are hypomorphic alleles of dPRPS that maintain a sufficient amount of PRPS function to support animal development . Because the mutant alleles are derived from Arts syndrome , we examined whether dPRPSQ165P and dPRPSR228W adult flies display any signs of neurological defects . Supporting the notion that dPRPS is important for the nervous system , a recent study identified CG6767 as a gene required for proper olfactory behaviour [18] . Consistent with other fly models of neurodegenerative disorders that have a shortened lifespan [19] , both patient-derived PRPS mutations demonstrate reduced lifespan . The lifespan of control versus dPRPS mutant flies were compared and at 50 days after eclosion , approximately 25% of dPRPSQ165P and dPRPSR228W flies survived while more than 50% of control flies were still alive ( Fig 2D ) . In addition , climbing tests were performed to examine locomotive and nervous system function [20] . Strikingly , dPRPSQ165P and dPRPSR228W flies display locomotive defects as early as three days after eclosion . While most control flies were able to reach a target line by 50 seconds , only about 75–80% of dPRPS mutants were able to climb to the threshold ( Fig 2E left panel ) . In addition , at 30 days after eclosion , a significant fraction of dPRPS mutant flies were unable to reach the target line while most control flies were able to do so in two minutes ( Fig 2E right panel ) . Overall , patient-derived mutations compromise dPRPS function and produce phenotypes associated with neurological defects . The Drosophila fat body functions as an energy reserve in the form of lipid droplets and plays a role similar to the mammalian liver and adipose tissue [21] . Interestingly , publicly available expression data from microarrays and RNA-sequencing analyses indicate that dPRPS is highly expressed in the Drosophila fat body ( S3A Fig ) . This was confirmed by the GFP-trap allele that showed strong expression in the fat body ( S3B Fig ) . Because animal survival during starvation is tightly linked to fat usage [21] , we examined the sensitivity of dPRPSQ165P and dPRPSR228W adult flies to starvation . Indeed , upon complete nutrition withdrawal , dPRPSQ165P and dPRPSR228W flies die about two days faster than age-matched control flies ( S3C Fig ) . We next visualized the lipid droplets in the fat body during the progression of starvation . While the size of lipid droplets decreases in control flies , it remains relatively unchanged in dPRPSQ165P and dPRPSR228W flies ( Fig 3A and 3B ) . Additionally , quantification of triglyceride ( TAG ) levels , the primary lipid form in which the fat is stored , revealed a similar trend wherein TAG levels are significantly reduced in the control flies during starvation but are relatively unchanged in the dPRPS mutants ( Fig 3C ) . This observation suggests that the increased sensitivity to starvation in dPRPSQ165P and dPRPSR228W flies may be caused by failure to use their lipid reserves . To test this hypothesis , we overexpressed Lipase 4 ( Lip4 ) , a lipase used in starvation-mediated lipolysis [22] , in control and dPRPS mutant fat bodies using lsp2-Gal4 ( Fig 3D ) . Lip4 overexpression had little to no effect on the survival of control flies during starvation . However , Lip4 overexpression in dPRPSQ165P and dPRPSR228W flies significantly improved their survival , suggesting that defects in lipid mobilization contribute to their hypersensitivity to nutrient withdrawal . During starvation , autophagy is one of the primary mechanisms to mobilize lipids [23] . Intracellular lipids are sequestered in double membrane vesicles called autophagosomes and delivered to the lysosome for their eventual degradation for energy production and macromolecular synthesis [24 , 25] . To determine whether autophagy is deregulated in dPRPS mutants , we expressed GFP-Atg8a , a molecular marker of autophagy , in the fat body [26] . In control flies , GFP-Atg8a signals under fed condition show low number of basal puncta that increase with starvation ( Fig 4A ) . Strikingly , in dPRPS mutant flies , not only are the basal GFP-Atg8a signals lower than the control , but the number of GFP-Atg8a puncta induced by starvation does not increase to the level observed in the control flies under fed condition ( Fig 4A lower panel ) . Fundamental defects in autophagy under fed condition was further demonstrated by abnormal accumulation of Ref ( 2 ) p/p62 , which is normally turned over by autophagy [27] . Even under normal growth conditions , Ref ( 2 ) p/p62 puncta were readily observed in dPRPSQ165P fat bodies , while being largely absent in control flies ( Fig 4B ) , supporting the notion that basal autophagy is deregulated in dPRPS mutant flies . To functionally confirm the autophagy defect revealed by the molecular markers , we examined the effect of depleting core autophagy proteins atg8a and atg16 in dPRPS fat bodies [26 , 28] . We reasoned that atg8a or atg16 depletion should have minimal effects in dPRPS mutant flies if autophagy is already compromised . Indeed , while atg8a or atg16 knockdown decreases the survival rate of control flies upon starvation , it has no significant effect on the survival of dPRPS mutant flies ( Fig 4C and 4D ) . Overall , these data suggest that dPRPS dysfunction caused by patient-derived mutations affects autophagy and may explain why dPRPS mutants are unable to breakdown lipid droplets during starvation . The endosomal system plays a critical function during macroautophagy [29] . Autophagosomes are generated at the endoplasmic reticulum ( ER ) and undergo a series of fusion processes with various endolysosomal compartments and lysosomes [29] . To determine whether an obvious abnormality exist in the endolysosomal system of dPRPS adult mutant flies , we visualized the ER , early endosome , and lysosome using anti-Calnexin ( S4A Fig ) , anti-Rab5 ( S4B Fig ) , and GFP-tagged Lamp1 ( GFP-Lamp1 ) ( Fig 5A ) , respectively . Strikingly , while the anti-Calnexin and anti-Rab5 did not show obvious differences , GFP-Lamp1 revealed that the lysosomal compartment is almost undetectable in dPRPS mutant fat bodies ( Fig 5A ) . Two additional molecular markers were used to monitor lysosome function: LysoTracker , a membrane-permeable dye marking acidic compartments such as the lysosome , and Magic Red , a substrate-based fluorescence marker of Cathepsin B [30] . Similar to what was observed with the GFP-Lamp1 , signals from both LysoTracker and Magic Red were significantly diminished in dPRPSQ165P and dPRPSR228W fat bodies while present in control fat bodies ( Fig 5B and 5C ) . Importantly , mitochondria were readily detected in both the control and dPRPS mutant fat bodies ( S4C Fig ) , indicating that the lysosome is specifically affected in dPRPS mutants . We also treated dPRPS mutant flies with chloroquine , a known inhibitor of lysosome function ( Fig 5D ) [31] . dPRPSQ165P and dPRPSR228W flies are more sensitive to chloroquine than control flies , indicating that lysosome function is indeed compromised in the mutant flies . Overall , our results indicate that the dPRPS is required for lysosome homeostasis and possibly contributes to the autophagy defect described in Fig 4 . To identify critical metabolic pathways contributing to the dPRPS mutant phenotypes , key enzymes in purine and pyrimidine biosynthetic pathways were knocked down in the fat body . Among the genes tested , adenylosuccinate synthase ( dAdS ) depletion reduced the LysoTracker signals , recapitulating the effect of dPRPS knockdown ( Fig 6A ) . Moreover , fat body-specific depletion of dAdS or dPRPS resulted in an increased sensitivity to starvation , similar to what was observed in dPRPSQ165P and dPRPSR228W flies ( Fig 6B ) . Importantly , depletion of an E2F-target important for dNTP synthesis , ribonucleotide reductase small subunit ( dRNRs ) , had no effect , indicating the effect of dAdS knockdown is specific ( Fig 6A and 6B ) . dAdS is a crucial enzyme in purine biosynthesis that converts Inosine monophosphate ( IMP ) to Adenylossucinate , which eventually becomes Adenosine monophosphate ( AMP ) . Interestingly , S-adenosylmethionine ( SAM ) , which can be converted to AMP and replenish purine nucleotides independently of PRPP , is given to patients with Arts syndrome hoping to restore a deficiency caused by PRPS dysfunction [32] . To determine if SAM treatment can have an effect on dPRPS-associated phenotypes , we fed dPRPSQ165P and dPRPSR228W flies with a sub-lethal dosage of SAM throughout the larval stage and determined its effect in adult flies . We also fed the mutants with a sub-lethal dosage of PRPP , the direct product of a PRPS-catalyzed reaction . Strikingly , SAM treatment partially and PRPP treatment virtually rescued the increased sensitivity of dPRPSQ165P and dPRPSR228W flies ( Fig 6C and S5B Fig ) . Moreover , analysis of the adult fat body revealed that SAM or PRPP treatment partially restored LysoTracker signals , explaining why they can rescue the increased sensitivity of dPRPS mutant flies to starvation ( Fig 6D and S5C Fig ) . Importantly , while PRPP treatment also suppressed the accumulation of Ref ( 2 ) p/p62 puncta in dPRPS mutant fat bodies , SAM treatment failed to do so ( Fig 6E and S5C Fig ) . This suggests that PRPP-dependent metabolic pathways other than purine biosynthesis may be critical for the autophagy defect and are required to fully restore the deficiency caused by dPRPS dysfunction . Furthermore , this likely explains why SAM treatment cannot fully rescue the increased sensitivity of dPRPS mutant flies to starvation . We therefore concluded that the phenotype observed in dPRPSQ165P and dPRPSR228W flies are contributed by a deficiency in purine metabolism . Autophagy and thus proper lysosome function are needed to protect cells against oxidative stress by removing damaged proteins and organelles [25] . Interestingly , dPRPS was identified as one of the transcripts whose expression is induced by oxidative stress [13] . While the dPRPS transcript level is unchanged upon starvation , it is upregulated in response to paraquat ( PQT ) treatment , a Parkinsonian toxin that can induce oxidative stress ( S6A Fig ) . We utilized the dPRPS GFP-trap allele to monitor dPRPS expression in specific tissues and noted that GFP expression is correspondingly increased in the ovary of PQT-treated flies ( Fig 7A upper panel ) . However , in the fat body where the basal expression level of the dPRPS is high , PQT treatment did not have any effect ( Fig 7A lower panel ) . Nevertheless , our results indicate that dPRPS expression is regulated by the cellular level of reactive oxygen species ( ROS ) and suggests a critical function during oxidative stress . Supporting this notion , we observed that dPRPSQ165P and dPRPSR228W fat bodies have higher levels of ROS measured by a fluorescent marker a 2' , 7'-dichlorodihydrofluorescein diacetate ( H2DCFDA ) ( Fig 7B ) . This observation and its role in autophagy/lysosome led us to hypothesize that dPRPS is required for protection against oxidative stress . Thus , we examined the survival of dPRPSQ165P and dPRPSR228W flies treated with PQT ( Fig 7C ) and hydrogen peroxide ( Fig 7D ) . Strikingly , dPRPS mutants died faster than control flies when treated with each compound , demonstrating their hypersensitivity to oxidative stress . We next tested whether the hypersensitivity of dPRPS mutants to oxidative stress contributes to their susceptibility to starvation . Control , dPRPSQ165P and dPRPSR228W flies were fed with an antioxidant NAC , and its effect on survival upon starvation was determined . While NAC supplementation had little to no effect on control flies , the survival of dPRPS mutants were partially rescued ( Fig 7E ) , suggesting that oxidative stress contributes to the hypersensitivity of dPRPSQ165P and dPRPSR228W flies to starvation . Taken together , these results suggest that dPRPS is a ROS-regulated gene whose function is critical for cellular response to oxidative stress . Given that the dPRPS mutant alleles are derived from Arts syndrome patients , we next investigated whether lipid mobilization , autophagy , and lysosome function are affected in the nervous system . A previous study has demonstrated that an accumulation of lipid droplets in surrounding glial cells of photoreceptor units in the adult eye is a common feature of Drosophila models of neurodegeneration [33] . We examined if a similar phenotype could be observed in dPRPSQ165P and dPRPSR228W flies . Indeed , we observed that lipid droplets accumulate in dPRPSQ165P and dPRPSR228W pupal eyes but not in control ( Fig 8A ) . Additionally , mutations affecting autophagy in Drosophila often result in the accumulation of protein aggregates in the brain , caused by a failure to turnover damaged cytosolic proteins [34] . Since locomotive defects are observable in dPRPSQ165P and dPRPSR228W mutants as early as three days after eclosion ( Fig 2E ) , we examined adult brains of three-day old control and dPRPS mutant flies . Anti-Ubi immunostaining ( Fig 8B ) and immunoblot ( Fig 8C ) revealed that dPRPSQ165P and dPRPSR228W have a higher level of ubiquitinated protein , suggesting that dPRPS function is required for proper clearance of damaged proteins . Notably , immunostaining resulted in a more striking difference than immunoblot . This is likely due to the formation of protein aggregates within the cell that amplifies the signal intensity . To better determine the role of PRPS in the nervous system , we depleted dPRPS in the nervous system using a pan-neuronal Gal4 driver , ELAV-Gal4 . This was sufficient to produce accumulation of protein aggregates in the brain ( Fig 8D ) and to result in climbing defects ( Fig 8E ) while muscle-specific depletion had no effect ( Fig 8F ) . Overall , these data suggest that the autophagy and lysosome defects , caused by the patient-derived PRPS mutations , affects the nervous system and provide a possible mechanism by which PRPS dysfunction contributes to neuropathogenesis .
In this study , we have established a Drosophila model of a PRPS-dependent neurological disorder . Using CRISPR/Cas9 , mutations found in Arts syndrome patients were introduced to the only member of Drosophila PRPS . We uncovered that dPRPS is required for proper lipid mobilization , autophagy induction , lysosome function , and cellular response to ROS . Importantly , PRPS-dependent processes affect the nervous system , providing a possible mechanism by which PRPS dysfunction contributes to PRPS-associated neuropathology . PRPS is a critical enzyme in nucleotide biosynthesis responsible for producing PRPP which is the five-carbon sugar subunit of nucleotides . Recent studies have shown that mammalian PRPS1 and 2 are regulated by genes and pathways that govern growth and proliferation such as Myc and AMPK [5 , 12] . These studies demonstrated that PRPS is a key enzyme that couples metabolic demand to nucleotide production . It is unclear if dPRPS in Drosophila is regulated in a similar manner . However , our data showing that dPRPS is a Cic- and RBF1-regulated gene ( Fig 1 ) and is required for Ras-induced hyperplastic overgrowth ( Fig 2C ) , suggest its vital role in nucleotide metabolism downstream of signals controlling proliferation . Moreover , we found evidence indicating that PRPS expression is controlled by cellular level of ROS ( Fig 7A and S6A Fig ) . It is probable that PRPS couples various environmental cues , including developmental signals , to regulate the rate of nucleotide synthesis . Importantly , we showed that dPRPS is critical for cellular level of autophagy and lysosomal homeostasis . An interesting possibility is that dPRPS expression is regulated to set the cellular level of autophagy and lysosome . This may also explain why in the fat body , where the basal level of autophagy is high [21] , dPRPS expression is developmentally upregulated ( S3A Fig ) . Of note , dPRPS expression is not induced by starvation indicating that the basal level of dPRPS is important for starvation-induced autophagy ( S6B Fig ) . However , it is possible that dPRPS activity is post-transcriptionally regulated upon energy deprivation similar to what was observed with mammalian PRPS [12] . Two independent models of Arts syndrome-derived mutations we established in this study , dPRPSQ165P and dPRPSR228W , have profound defects in autophagy and lysosome function . Importantly , these defects can be recapitulated by PRPS depletion ( Figs 6A and 8D ) and transheterozygous between patient-derived dPRPS mutations and a deficiency line covering the dPRPS locus ( S7A–S7C Fig ) . This indicates that the patient-derived mutations decrease PRPS function and may not result in a gain of aberrant activity . Given the high degree of sequence homology ( S1A Fig ) and its essential role in nucleotide metabolism [1] , it is probable that patients with PRPS1-associated disorders have similar defects described in this study . However , we have yet to pinpoint exact molecular and cellular processes that are affected by dPRPS dysfunction . Our data herein suggest that defects in autophagy and lysosome homeostasis may be independently affected by dPRPS dysfunction . This is most evident when dPRPSQ165P and dPRPSR228W are treated with SAM . While PRPP can suppress both autophagy and lysosomal defects , SAM only partially improves lysosomal defects ( Fig 6D and 6E ) . In addition , lysosomal dysfunction alone cannot fully explain the autophagy defect described in this study . A recent study demonstrated that acidification is not required for autophagosome-lysosome fusion [35] . The v-ATPase-deficient non-acidic lysosomes can still fuse with autophagosomes , causing Atg8 to accumulate in abnormally large vesicles . This is certainly not the phenotype observed in dPRPS mutant fat bodies , where no Atg8a-puncta were detectable and indicative of defective autophagosome formation ( Fig 4 ) . Notably , our observations using various molecular markers of the lysosome all indicate that lysosomes are scarce in dPRPSQ165P and dPRPSR228W ( Fig 5 ) . The Drosophila mutants of MITF , a transcription factor important for lysosome biogenesis and homeostasis , show a similar phenotype as described in this study [36] . Detailed analysis comparing dPRPS and MITF mutant flies may tell us the role of lysosome in dPRPS-associated defects . Because PRPP , the product of PRPS ( Fig 1A ) , is a precursor for purines , pyrimidines , as well as pyridines that are important for a wide range of metabolic processes , it is unclear which dPRPS-dependent metabolic changes cause autophagy and lysosomal defects . We are in the process of performing extensive metabolic profiling and genetic dissection of the dPRPS-dependent pathways to determine the exact relationship between nucleotide metabolism and the endolysosomal system . However , we suspect that crucial metabolites missing in dPRPS-deficient cells can be non-cell autonomously provided . When clones of cells expressing the same dPRPS RNAi construct used in Fig 6A are generated in the fat body , they failed to show any decrease in LysoTracker signals ( S8 Fig ) . Since the same RNAi construct were able to decrease the LysoTracker signals when expressed in the entire tissue ( Fig 6A ) , this is unlikely due to inefficient knockdown . We hypothesize that neighboring wild-type cells produce and provide missing metabolites and/or their precursors to the dPRPS-deficient cells . While the identity of the exact metabolite is unclear , our data indicate that SAM or PRPP can be non-cell autonomously provided to dPRPS mutant cells and suppress their defects ( Fig 6C–6E ) . Importantly , autophagy and lysosomal defects in dPRPS mutant larvae are not as pronounced as those observed in the adult flies ( S7D and S7E Fig ) . It is possible that the difference in the feeding behavior and metabolic status of larvae compared to the adult may affect the overall consequence of PRPS dysfunction . Perhaps , constant intake of food during the larval stage can partly compensate for the PRPP deficiency in dPRPSQ165P and dPRPSR228W flies . Another factor that may explain why autophagy and lysosomal defects are more pronounced in adult flies is the role of macroautophagy during metamorphosis [37] . Several studies have demonstrated that remodelling of tissues such as fat body , salivary glands , and midgut during pupariation requires macroautophagy . PRPS dysfunction may lead to a failure of proper tissue remodelling during metamorphosis and the accumulative effects manifest in the adult stage . However , while macroautophagy is important for tissue remodelling , the Atg8 conjugation system is not essential for completion of pupariation . Null mutants of atg7 , which functions as an E1-like enzyme for Atg8 and Atg12 , complete metamorphosis despite an abnormal level of autophagy [38] . It is worth noting that the Atg8 conjugation machinery is not absolutely required for autophagy [39] and that our assays largely relies on GFP-Atg8a quantification . Further in-depth analysis such as transmission electron microscopy will determine the precise impact of dPRPSQ165P and dPRPSR228W mutations on autophagy and lysosome homeostasis . Many studies have clearly demonstrated that defects in autophagy are involved in the pathologies of various nervous system disorders , such as Alzheimer’s , Parkinson’s and Huntington’s disease [40–42] . In addition , there is a growing body of evidence suggesting that lysosome-mediated processes are important for neuronal homeostasis [41] . Notably , missense mutations in rab7 , a small GTPase that plays a regulatory role in the endolysosomal pathway and autophagy , were also identified in CMT , one of the neurological disorders associated with PRPS1 mutations [43] . To date , no treatment option is available for patients with PRPS1-associated neurological disorders , such as children with Arts syndrome [3 , 44 , 45] . Preliminary results from an open-label clinical trial revealed that dietary SAM supplementation stabilized neurodegenerative progression in one ARTS syndrome patient [16] . Likewise , in our study SAM treatment partially rescued the starvation sensitivity and lysosome dysfunction in dPRPSQ165P and dPRPSR228W flies ( Fig 6C and 6D and S5B Fig ) . This suggests that the dPRPS mutants described in this study may be used to identify future therapies for PRPS1-associated disorders . Moreover , identifying the exact mechanism by which PRPS regulates autophagy and lysosome homeostasis may uncover a novel treatment option for PRPS1-associated neurological disorders .
D . melanogaster stock cultures were maintained at 25°C . All crosses were performed at 25°C with the exception of the UAS-RASV12 crosses at 18°C . Following stocks were obtained from Bloomington stock center: PRPSMi09551 ( #53132 ) , PRPSMi09551-GFP-trap ( #59305 ) , UAS-RASv12 ( #4847 ) , UAS-Lipase ( #67142 ) , UAS-Atg8a ( #58309 ) , UAS-Atg16 ( #58244 ) , Lsp2-Gal4 ( #6357 ) , UAS-AdSRNAi ( #33993 ) , MEF2-Gal4 ( #27390 ) , ELAV-Gal4 ( #8765 ) and the BSC576 deficiency stock ( #26827 ) . The GFP-Atg8a and GFP-Lamp1 are from [46] . The cicQ474X mutant allele used in this project , including the method of generating clones in the eye has been previously described in [14] . The rbf1cas21 allele that normally contains rbf1 mutant clones in the eye , is used in this study . The full description of the allele can be found in [47] . To generate eyes mostly composed of rbf1cas21 mutant cells , an additional copy of eye-specific FLPs was introduced to the rbf1cas21 allele . The UAS-PRPSRNAi construct used in this project was obtained from the Vienna Drosophila Resource Center ( VDRC ) ( # 35112 ) . The UAS-RNRsRNAi construct used in this project was obtained from the Harvard Medical Centre Transgenic RNAi Project ( TRiP ) . dPRPSQ165P and dPRPSR228W mutants were generated through homology-directed protocol as previously described in [48] . Positive CRISPR lines were screened through T7 Endonuclease I assay following the manufacturer's instructions ( NEB MO302S ) and confirmed through sequencing their genomic DNA ( S2A Fig ) . To ensure that the control used in this study is within the same genetic background as the CRISPR generated mutants , an allele that does not contain the intended mutation was selected during the screening process . This control allele is used for starvation assays , feeding treatments , ATP determination , triglyceride quantification , immunofluorescence staining , immunoblot , developmental tracking , lifespan , and climbing assays . Other genotypes used as control are specified otherwise . Donor DNA for dPRPSQ165P: TCCTTTGCATCTCTTTCTACGCTGGCCAATGCACAGAGTCGTGCGCCCATCTCGGCCAAATTGGTGGCCAACATGCTGTCCGTTGCTGGAGCGGATCACATCATCACCATGGATCTGCACGCCTCACCGATTCAGGTAAGTCAGCCCATCAACAACATTTGTATATTTATCTTTGATATTAGAGTGATTTCTTATCGTGC Donor DNA for dPRPSR228W: TGCGTAATAGATTAATAATGCTATTAATATTTCACTTTAAGTGTCACCTCAATTGCCGATCGACTGAACGTGGAGTTCGCTCTGATACACAAGGAGTGGAGAAGGCCAACGAAGTGGCCTCTATGGTACTGGTGGGTGATGTCAAGGACAAGATTGCCATTCTGGTCGATGACATGGCCGACACATGCGGCACCATTGTG Guide RNA ( gRNA ) as follows: dPRPSQ165P forward: GTCGCCAACATGCTGTCCGTTGC dPRPSQ165P reverse: AAACGCAACGGACAGCATGTTGGC dPRPSR228W forward: GTCGCGCAAGAAGGCCAACGAA dPRPSR228W reverse: AAACCTTCGTTGGCCTTCTTGCGC For each genotype , triplicate batches of 25 female flies ( 2–3 days old ) were transferred to vials of either normal cornmeal medium ( fed ) or water supply only ( 1% agarose in H2O ( starved ) ) . Survival rates were determined by counting the number of dead flies diagnosed by lack of a sit-up response every 12–14 hours . Triplicate sets were performed . Eggs were laid in cornmeal medium treated with PBS , 10mg/mL of NAC ( Sigma ) , 30mM of PRPP ( Sigma ) or 30mM of SAM ( NEB ) to allow larvae to develop in the presence of each compound . Triplicate batches of 25 female flies , for SAM or PRPP , or 35 female flies , for NAC , were then collected and kept in the same media until they are 2–3 days old . Starvation assays were performed as described above . Triplicate batches of 25 2–3 day old females were transferred to cornmeal medium treated with 20mM of chloroquine ( Sigma ) , 30mM of paraquat ( Sigma ) or 1% hydrogen peroxide ( Sigma ) . Survival rates were determined by counting the number of dead flies diagnosed by lack of a sit-up response every 12–14 hours . Triplicate sets were performed . RNA was extracted using the RNeasy Mini Kit ( Qiagen ) according to manufacture specifications . RNA was collected from whole yw flies ( n = 5 ) for the measurement of dPRPS transcript levels in adult females fed normal yeast medium supplemented with 1X phosphate buffered saline ( PBS ) , N-acetyl cysteine ( NAC ) , or paraquat ( 24 hours ) . 500 ng of RNA was reverse transcribed using the DyNamo cDNA synthesis kit ( ThermoScientific ) . Reverse transcriptase qPCR ( RT-qPCR ) experiments were done using DyNAmo Flash SYBR Green qPCR kit ( ThermoScientific ) according to the manufacture specifications . Threshold cycle ( CT ) was determined using the Bio-Rad CFX Manager software . rp49 and β-tubulin were both used for normalization . Each primer reaction was performed in triplicates and the three biological replicates were averaged . Primers were designed using Primer3 ( Whitehead Institute for Biomedical Research Primer3 shareware , Frodo . wi . mit . edu/Primer3 ) . The following primers were used for qPCR reactions: β-tubulin forward: ACATCCCGCCCCGTGGTC β-tubulin reverse: AGAAAGCCTTGCGCCTGAACATAG dPRPS Exon4 forward: CTTAGCAAGGGGTGATTTGG dPRPS Exon4 reverse: CCTTGATCCACTTGAGTACC dPRPS-E1-E3 forward: TTCAGCAACTTGGAGACCTG dPRPS-E1-E3 reverse: CCATGGTGATGATGTGATCCD dRNRs forward: CGTCCAAGGAAAACATTGCTG dRNRs reverse: TGGTGCTATCCGTCAGAATCTT dAdS forward: CCGGCTTACTCCAGCAAGG dAdS reverse: GGCCACAATCGACTTGAACTTTT Five female flies ( 2–3 days old ) were homogenized in 100 μl of 6 M guanidine-HCl in extraction buffer ( 100 mM Tris and 4 mM EDTA , pH 7 . 8 ) to inhibit ATPases . Homogenized samples were frozen in liquid nitrogen , followed by boiling for 5 minutes . Samples were centrifuged and diluted with extraction buffer followed by the addition of luminescent solution ( Invitrogen ) . Luminescence was measured on a luminometer ( Turner Biosystems ) . Relative ATP levels were normalized to the total protein concentration determined by Bradford assay . Triglyceride levels were measured using a coupled colorimetric enzymatic triglyceride kit following manufacture instructions ( Stanbio ) . 5 adult flies per genotype were used for each sample measurement in triplicates . Protein levels were measured in conjunction with TAG levels using Bradford assay ( BioRad ) . For whole-mount staining of adult fat bodies ( < 36 hours after eclosion ) , were dissected in 1X PBS and fixed with 4% formaldehyde in PBS for 20 minutes at room temperature . The fixed samples were then washed twice with 0 . 1% triton in 1X PBS ( 0 . 1% PBST ) for 10 minutes following incubation in 1X DAPI in 0 . 1% PBST at room temperature for 15 minutes . After washing with 0 . 1% PBST , samples were incubated in 1:500 HCS LipidTOX Red Neutral Lipid Stain ( ThermoFisher H34476 ) . Samples were visualized within 2 hours by Leica SP8 point-scanning confocal system on a Leica DMI6000B inverted microscope ( provided by the Cell Imaging Analysis Network ( CIAN ) in the Core Facility for Life Sciences at McGill University ) . Adult fat bodies ( 2–3 days old ) were dissected in 1X PBS and stained with either LysoTracker Red ( 1:1000 , ThermoFisher L7528 ) or Magic Red ( 1:150 , BioRad IT937 ) Mitotracker ( 1:300 , ThermoFisher M7512 ) for 5 , 15 and 30 minutes respectively . Samples were directly visualized using Zeiss Axio Imager fluorescent microscope . Adult fat bodies ( 2–3 days old ) were dissected in Schneider’s Media and immediately incubated in 10 μm 5- ( and-6 ) -carboxy-2′ , 7′-dichlorodihydrofluorescein diacetate , acetyl ester ( CM-H2DCFDA , Molecular Probes , Invitrogen C6827 ) for 5 minutes . The samples were then washed twice with 0 . 1% triton in 1X PBS ( 0 . 1% PBST ) for 10 minutes following incubation in 1X DAPI in 0 . 1% PBST at room temperature for 15 minutes . Samples were visualized within 30 minutes using Zeiss Axio Imager fluorescent microscope . The following antibodies were used in this study: anti-Calnexin antibody ( 1:100 , Abcam 75801 ) , anti-Rab5 ( 1:100 , Abcam 18211 ) , anti-Fasciclin II ( 1:100 , DHSB 528235 ) , anti-Ref ( 2 ) p/p62 ( 1:100 , Abcam 78440 ) , anti-PH3 ( 1:100 , Abcam 5176 ) , anti-DCP1 ( 1:100 , Cell Signalling 9578 ) and anti-GFP-FITC ( 1:200 , Abcam 25052 ) . Adult fat bodies were dissected in 1X PBS and fixed with 4% formaldehyde for 20 minutes . Fixed samples were washed twice with 0 . 3% PBST ( 0 . 3% Triton-X-100 in PBS ) for 10 minutes . Next , samples were incubated with the indicated antibody . Fat body images were taken using Zeiss Axio Imager fluorescent microscope . Adult brain images were taken using Leica SP8 point-scanning confocal system on a Leica DMI6000B inverted microscope ( provided by the Cell Imaging Analysis Network ( CIAN ) in the Core Facility for Life Sciences at McGill University ) . Fat bodies and lipid droplet size were analysed using ImageJ software . Quantification of puncta was performed with ImageJ using the particle intensity and 3D objects counter plugin [49] . Climbing tests were performed as previously described [20] . Briefly , 20 adult flies were transferred into a 250 mL glass graduated cylinder and gently tapped to the bottom of the cylinder . The number of flies reaching the 190 mL line ( 17 . 5 cm ) of the 250 mL graduated cylinder was measured every ten seconds for two minute increments . Ten trials were performed for each genotype . Error bars indicate standard deviation from triplicated experiments . ChIP-qPCR experiments were performed as previously described in [50] . Briefly , chromatin was collected from w1118 flies . Presented data are averages of triplicated ChIP experiments which consisted of experimental duplicates followed by quantitative real time PCR reactions . Presented data for target loci enrichment is represented by percentage of input chromatin not subjected to immunoprecipitation . All primers were designed by Primer 3 and primers used for ChIP-qPCR analyses are following: Images were taken on FEI Quanta 450 Environmental Scanning Electron Microscope ( FE-ESEM ) located at the Facility for Electron Microscopy Research at McGill University . All blots were blocked with 5% skim milk powder in 0 . 1% PBS-Tween20 . To measure ubiquitinated protein levels , adult fly heads were collected from control and dPRPS mutants at day 3 . Anti-Ubiquitin ( 1:1000 , Invitrogen and Anti- β-tubulin ( 1:1000 , DSHB ) were used as primary antibodies followed by anti-mouse HRP ( 1:2000 , GE Healthcare ) . Adult eye size was examined by using the nail polish imprinting technique as previously described [51] . Adult heads were covered in clear nail polish and dried at room temperature for 1 hour . The nail polish imprint was peeled off using tungsten needles and immediately mounted in 100% glycerol . Minimum of 5 adult heads were analyzed .
|
Phosphoribosyl pyrophosphate synthetase ( PRPS ) is an important enzyme in nucleotide synthesis: the building blocks of DNA and RNA and other important metabolites . Importantly , while PRPS is mutated in neurological disorders such as Arts syndrome , Charcot-Marie-Tooth disease , and nonsyndromic sensorineural deafness , it is currently unclear why PRPS dysfunction leads to neurological disorders . In this study , we engineered a Drosophila model of ARTS syndrome and discovered that PRPS mutations result in defects in lysosome-mediated and autophagy processes , which are known to be important for neuronal homeostasis . Our study provides crucial insights into the way that PRPS mutations contribute to neurological disorders .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"cell",
"death",
"invertebrates",
"lysosomes",
"autophagic",
"cell",
"death",
"medicine",
"and",
"health",
"sciences",
"oxidative",
"stress",
"cell",
"processes",
"animals",
"animal",
"models",
"drosophila",
"melanogaster",
"model",
"organisms",
"experimental",
"organism",
"systems",
"eyes",
"cellular",
"structures",
"and",
"organelles",
"drosophila",
"research",
"and",
"analysis",
"methods",
"lipids",
"animal",
"studies",
"fats",
"head",
"insects",
"arthropoda",
"biochemistry",
"eukaryota",
"cell",
"biology",
"anatomy",
"neurology",
"biology",
"and",
"life",
"sciences",
"ocular",
"system",
"organisms"
] |
2019
|
Pleiotropic role of Drosophila phosphoribosyl pyrophosphate synthetase in autophagy and lysosome homeostasis
|
Poly ( A ) -binding proteins ( PABPs ) regulate mRNA fate by controlling stability and translation through interactions with both the poly ( A ) tail and eIF4F complex . Many organisms have several paralogs of PABPs and eIF4F complex components and it is likely that different eIF4F/PABP complex combinations regulate distinct sets of mRNAs . Trypanosomes have five eIF4G paralogs , six of eIF4E and two PABPs , PABP1 and PABP2 . Under starvation , polysomes dissociate and the majority of mRNAs , most translation initiation factors and PABP2 reversibly localise to starvation stress granules . To understand this more broadly we identified a protein interaction cohort for both T . brucei PABPs by cryo-mill/affinity purification-mass spectrometry . PABP1 very specifically interacts with the previously identified interactors eIF4E4 and eIF4G3 and few others . In contrast PABP2 is promiscuous , with a larger set of interactors including most translation initiation factors and most prominently eIF4G1 , with its two partners TbG1-IP and TbG1-IP2 . Only RBP23 was specific to PABP1 , whilst 14 RNA-binding proteins were exclusively immunoprecipitated with PABP2 . Significantly , PABP1 and associated proteins are largely excluded from starvation stress granules , but PABP2 and most interactors translocate to granules on starvation . We suggest that PABP1 regulates a small subpopulation of mainly small-sized mRNAs , as it interacts with a small and distinct set of proteins unable to enter the dominant pathway into starvation stress granules and localises preferentially to a subfraction of small polysomes . By contrast PABP2 likely regulates bulk mRNA translation , as it interacts with a wide range of proteins , enters stress granules and distributes over the full range of polysomes .
Gene expression is regulated by multiple transcriptional and post-transcriptional mechanisms . At the post-transcriptional level , regulation of protein synthesis by modulation of translation initiation is a major contributor . The first step in mRNA cap-dependent translation initiation is assembly of the eIF4F complex at the m7G cap of the mRNA 5’ end [1] . The eIF4F complex consists of a large ( ~180kDa ) scaffold protein , eIF4G , bound to the cap-binding protein eIF4E and an RNA helicase , eIF4A . The latter is involved in secondary structure unwinding of the target mRNA , facilitating 40S subunit scanning , together with a further factor eIF4B . Significantly , eIF4G and eIF4B directly interact with the poly ( A ) -binding protein ( PABP ) associated with the poly ( A ) tail at the 3’ end of the target mRNA , to increase translation efficiency by mRNA circularisation and ribosome recycling . Most higher eukaryotes have several paralogs of eIF4F complex subunits [2] and PABP [3]; increasing evidence suggests that these different paralogs can assemble into distinct eIF4F complexes , facilitating modulation of translation to distinct environmental and developmental conditions [4] . For example , in metazoa there is one eIF4F complex specialised to mediate cap-dependent translation under low oxygen conditions [5 , 6] , and specific eIF4F complexes select distinct sets of mRNAs during development in C . elegans germ cells [7] . The specific functions of distinct eIF4F complexes are mediated by the properties of the individual subunits , for example H . sapiens eIF4E paralogs differ in their ability to localise to P-bodies and stress granules [8] , ribonucleoprotein granules ( RNA granules ) with important functions in mRNA storage , regulation and quality control [9] . The presence of multiple PABP paralogs further increases the combinatorial complexity of this system . Arabidopsis thaliana has eight PABP paralogs [10] that differ in domain structure and expression patterns , with both overlapping and distinct functions [10–16] . Xenopus laevis has three paralogs that are all independently essential [17] . Many protozoa also possess several paralogs of each of the eIF4F complex subunits , but these are the product of lineage-specific expansions and hence unrelated to the paralogs found in higher eukaryotes . Very little is known about their specific functions [18 , 19] . Kinetoplastids , including the animal and human pathogens Leishmania , Trypanosoma cruzi and T . brucei , rely almost completely on post-transcriptional gene regulation [20] . mRNAs are transcribed poly-cistronically and processed by trans-splicing of a miniexon to the 5’ end , a process coupled to polyadenylation of the upstream transcript [21–26] . Furthermore , the mRNA cap structure is a highly unusual type four , with ribose 2’-O methylations at the first four transcribed nucleotides ( AACU ) and additional base methylations at the first ( m26A ) and fourth ( m3U ) positions [27 , 28] . This cap requires a kinetoplastid-specific decapping enzyme for degradation [29] . Hence , translational control is a major contributor to gene regulation [30] . As a possible consequence of this kinetoplastids possess a large number of translation initiation factor paralogs [31]: six for eIF4E ( eIF4E1-6 ) , five for eIF4G ( eIF4G1-5 ) and two for eIF4A ( eIF4A1-2 ) , of which only one , eIF4A1 , is known to be involved in translation [32] . Trypanosomes have two PABP paralogs ( PABP1 , PABP2 ) , while Leishmania has an additional paralog ( PABP3 ) . Multiple studies have addressed the composition of kinetoplastid translation initiation complexes , and whilst data are equivocal in some cases , several distinct eIF4F complexes were described ( recently reviewed in [31] . The best characterised complex comprises eIF4E4 , eIF4G3 , eIF4A1 and PABP1 in both Leishmania and trypanosomes [33–37] . Evidence of a direct physical interaction between eIF4E4 and eIF4G3 was obtained in L . major using yeast two hybrid [37] , but direct binding between LmPABP1 and eIF4G3 was not observed [35 , 37] . Instead , LmPABP1 interacted directly with eIF4E4 , mediated by the non-conserved N-terminal extension of eIF4E4 [35] , an interaction critical for the function of eIF4E4 [38] . The current assumption that eIF4E4/eIF4G3/PABP1 is the major translation initiation complex is predicated on the following: i ) all proteins are of high abundance , ii ) PABP1 has greater specificity for poly ( A ) than PABP2 [36 , 39] , iii ) eIF4E4 binds the type 4 cap with the highest affinity of all eIF4E4 paralogs [40–42] and iv ) silencing of eIF4E4 , eIF4G3 and PABP1 in at least some T . brucei life cycle stages is lethal [33 , 34 , 36] and eIF4E4 cannot be deleted in L . infantum [38] . At least three additional translation initiation complexes are known . The first consists of eIF4E5 bound to either eIF4G1 or eIF4G2 [41] . Two further proteins specifically interact with the eIF4G1 version of this complex: TbG1-IP ( Tb927 . 11 . 6720 ) and TbG1-IP2 ( Tb927 . 11 . 350 ) . TbG1-IP is an mRNA cap guanine-N7 methyltransferase , suggesting involvement in nuclear mRNA capping [43] , but such a function is unlikely , as the protein is cytoplasmic and localises to starvation stress RNA granules [44] , and nuclear cap methylation is known to be performed by TbCGM1 [45 , 46] . TbG1-IP2 is an RNA binding protein with unknown function . The second complex consists of eIF4G5 , which specifically interacts with eIF4E6 and one further protein TbG5-IP ( Tb927 . 11 . 14590 ) [42] . Interestingly , similarly to TbG1-IP1 , this protein contains a nucleoside triphosphate hydrolase and a guanylyltransferase domain in common with enzymes involved in cap formation . The third complex consists of eIF4G4 , eIF4E3 and eIF4A1 [33] . However , neither PABP was identified in any of these complexes . Several studies have directly addressed function , substrate specificity and localisation of kinetoplastid poly ( A ) -binding proteins . PABP1 and PABP2 are highly abundant and in excess of the total number of mRNA molecules , at least in the procyclic life cycle stage of T . brucei [36] . RNAi in T . brucei revealed that both isoforms are essential [36] and both isoforms stimulate translation when tethered to the 3’ end of a reporter mRNA [47 , 48] . Both PABPs are cytoplasmic in untreated cells , but differentially localise under stress conditions: PABP2 , but not PABP1 , localises to the nucleus under certain conditions [36 , 49] and only PABP2 localises to starvation stress granules [49 , 50] , while PABP1 and its interacting partners eIF4E4 and eIF4G3 do not [49] . Both PABPs localise to polysomes [49 , 51] , but PABP1 is mainly located in small polysomes while PABP2 is more equally distributed across all polysomes [49] . There is some evidence that PABP2 may have a function unrelated to poly ( A ) binding . PABP2 binds poly ( A ) with lower specificity ( in comparison to PABP1 ) in Leishmania [36 , 39] and binds to the CAUAGAAG element present in cell-cycle regulated mRNAs of Crithidia fasiculata [52] and to the U-rich RNA binding protein UBP1 [53] , which mediates instability of the T . cruzi SMUG mucin mRNA [54] . To probe for distinct roles of PABPs we examined their protein interactomes in T . brucei procyclic forms . PABP1 co-precipitates eIF4E4 and eIF4G3 and RNA-binding protein RBP23 , but few additional proteins . In contrast , PABP2 co-precipitated a large number of RNA binding proteins , including all proteins that co-precipitated with PABP1 except RBP23 . Most eIF4F paralogs co-precipitated with PABP2 , most significantly the eIF4G1/eIF4E5 complex and its two interacting partners TbG1-IP and TbG1-IP2 . These data , together with analysis of the localisations of PABP1 and PABP2 complex components challenge the current paradigm that PABP1 is the major poly ( A ) -binding protein in trypanosomes and an alternative model is discussed .
To isolate PABP complexes we used two previously published cell lines expressing C-terminal eYFP fusions of each PABP paralog from their endogenous locus; the second allele remained unaltered [49] . PABP1-eYFP is fully functional as deletion of the wild type allele has no phenotype , while RNAi that targets both alleles is lethal . In the cell line expressing PABP2-eYFP , the second allele could not be deleted , but cells have normal growth rates and localisation of the protein to various types of RNA granules was indistinguishable from that determined with specific antiserum against PABP2 [49] . This indicates that most functions of PABP2-eYFP are essentially identical to the wild type protein . Protein expression and localisation to the cytoplasm was demonstrated by fluorescence microscopy ( Fig 1A ) . Wild type cells served as negative controls . Cultures of each cell line were snap-frozen and subjected to cryomilling to generate a powder [55] . Aliquots of this powder were used to systematically optimise conditions for isolation of PABP complexes . Protein complexes were captured with polyclonal anti-GFP antibodies covalently coupled to magnetic Dynabeads and analysed by SDS-PAGE . In the optimised protocol , the cell powder was solubilised using CHAPS detergent with two different buffers: a low salt buffer and a high salt buffer , the latter contained 150 mM KCl but was otherwise identical to the low salt buffer . Coomassie-stained gels revealed clearly visible PABP bait proteins and several protein bands specific to one or both PABPs , but absent from the control pull down ( Fig 1B ) . For each cell line , protein complexes were isolated in two independent experiments for each buffer condition and the captured proteins analysed by liquid chromatography tandem mass spectrometry ( LC-MS2 ) and subjected to label free quantification using MaxQuant [56] . 1901 distinct protein groups ( peptides assigned to a specific coding sequence , but where these cannot be assigned to a single gene in the case of close paralogs ) were identified ( S1 Table ) ; this list was reduced to 1224 after removing all protein groups with less than three unique peptides ( S1B Table ) . For each protein group from each experiment we determined the enrichment ratio in relation to the wild type control cell line , based on quantification by unique peptides only . To avoid division by zero , a constant ( 0 . 001 ) was added to each LFQ value; such ‘infinite ratios’ are clearly distinguishable from genuine ratios by being significantly larger , smaller or exactly 1 . 0 ( S1B Table ) . For PABP1 , we identified 25 proteins at least two-fold enriched in each of the two low salt replicates ( S1C Table ) and 66 proteins at least two-fold enriched in both high salt replicates ( S1D Table ) . For PABP2 , 77 and 170 proteins were enriched in both replicates under low salt and high salt conditions , respectively ( S1E and S1F Table ) . Ribosomal proteins were exclusively co-precipitated under high salt conditions and not detected under low salt , consistent with intact ribosomes requiring physiological potassium concentrations and dissociating upon potassium depletion [57 , 58] . Interestingly , the number of co-precipitated ribosomal proteins differed between the PABP1 and PABP2 pull-downs: 43 proteins were co-purified with PABP2 ( 25% of all precipitated proteins ) , but only 7 ribosomal proteins with PABP1 ( 11% of all precipitated proteins ) . This could reflect differences in polysomal association between the two isoforms: PABP2 associates with heavier sucrose fractions than PABP1 on polysome fractionation gradients [49] . Alternatively , these differences could be explained by the RNA-binding ability of PABP2 being less specific to poly ( A ) tails in comparison to PABP1 , as has been previously found for Leishmania orthologues [36]: unspecific binding of PABP2 to ribosomal RNA could cause co-precipitation of intact ribosomes under high salt conditions , resulting in the presence of ribosomal proteins in the proteomics data . Evidence for the second hypothesis is provided by the large number of nucleolus-localised proteins in the PABP2 pull-down with high salt buffer: 20 of the 127 non-ribosomal proteins purified with PABP2 are known to entirely or predominantly localise to the nucleolus , in comparison to only 4 of 59 non-ribosomal proteins purified with PABP1 [59] . PABP2 does not localise to the nucleolus , at least not to detectable levels , thus , these interactions are likely non-physiological . All PABP interacting proteins were judged for their possible function in mRNA metabolism . A protein was classified as having a known role in mRNA metabolism ( indicated as ‘yes’ in S1C–S1F Table ) , if it possesses an RNA-binding domain , or if there is direct experimental evidence for involvement in RNA metabolism ( for example validated localisation to RNA granules ) . A protein was classified as having a predicted role in mRNA metabolism ( indicated as ‘ ( yes ) ’ in S1C–S1F Table ) if it was identified in one out of three large scale experiment that screened for posttranscriptional activators , repressors and RNA-binding proteins [47 , 48] , without further experimental validation . The low salt precipitations contained mostly proteins with a known or predicted function in mRNA metabolism for both PABP1 ( 19/25 proteins ) and PABP2 ( 62/77 proteins ) and few obvious contaminants . High salt precipitations were still enriched in mRNA metabolism proteins ( PABP1 28/66 and PABP2 54/170 ) but also contained a large fraction of likely or obvious contaminants , including mitochondrial , nucleolar and ribosomal proteins . To obtain a high confidence list , we filtered for proteins that were at least two-fold enriched in all four experiments . In a second step , all protein groups with more than one infinite ratio were removed , and three further proteins were manually removed because they were obvious contaminations; two mitochondrial RNA-binding proteins ( Tb927 . 7 . 2570 , Tb927 . 2 . 3800 ) and one glycosomal protein ( Tb927 . 10 . 5620 ) . Average enrichment ratios were calculated for each protein , excluding ‘infinite ratios’ ( S1G Table , Fig 2A ) together with a PAPB1/PABP2 enrichment ratio , to determine the specificity of each interaction ( S1G Table , Fig 2B ) . All 27 PABP-interacting proteins have a known or predicted function in mRNA metabolism . 12/27 proteins were more than 2-fold enriched in both pull-downs . For 6 of the 27 proteins the interaction with PABP ( s ) had been independently validated in at least one of the Kinetoplastids: ALBA1-3 co-precipitate both T . brucei PABPs [60] . Both T . brucei PABPs were found in a large scale yeast 2-hybrid screen to interact with PBP1 [61] . Several studies have identified PABP1 as part of the eIF4G3/eIF4E4 complex in Leishmania [34–36] , with an unusual direct interaction between PABP1 and eIF4E4 [35] . For Leishmania eIF4E4 , additional interaction with PABP2 was shown [35] . Moreover , while this manuscript was in revision , a PABP1 interactome for Leishmania infantum was published [62] and is in agreement with our data: seven proteins consistently co-precipitated with Leishmania PABP1 , of which six correspond to the six most enriched proteins in the T . brucei PABP1 pulldown ( eIF4E4 , eIF4G3 , PABP1 , RBP23 , Tb927 . 7 . 7460 , ZC3H41 ) and only one protein ( Tb927 . 10 . 13800 ) was not identified with our conditions . As a further control , we performed reverse pull-downs of the proteins mostly enriched in either the PABP1 pull-down ( eIF4E4 ) or the PABP2 pull-down ( G1-IP2 ) ( Fig 2C ) . For this , eIF4E4 and G1-IP2 were expressed as eYFP fusion proteins from their endogenous loci , in cell lines also expressing PABP1 or PABP2 C-terminally fused to a tandem of four Ty1 epitopes . Precipitations of eIF4E-eYFP and G1-IP2-eYFP were performed as above , using low salt buffer conditions . Co-precipitated PABP-Ty1 proteins were detected by western blot probed with anti-Ty1 . Both PABP proteins were enriched in these pull-downs in comparison to the negative control . However , in agreement with the mass spectrometry , PABP1 had a much higher enrichment ratio than PABP2 in the eIF4E4 pull-down ( 64-fold/3-fold on average for PABP1/PABP2 , n = 2 ) while the opposite was found for the G1-IP2 pull-down ( 4-fold/19-fold on average for PABP1/PABP2 , n = 2 ) . Two proteins were particularly enriched in the PABP1 pull-down , with a more than 100-fold enrichment and more than 20-fold enrichment against PABP2 . These are the two known PABP1 interactors eIF4E4 and eIF4G3 , confirming the specificity of the pull-down . Only three further proteins had average enrichment ratios of >10 in comparison to the negative control , namely the RNA binding protein RBP23 , a hypothetical protein Tb927 . 7 . 7460 and the CCCH type zinc finger protein ZC3H41; all experimentally uncharacterised . RBP23 was the only protein that was solely precipitated with PABP1: all other PABP1 interacting proteins also interact with PABP2 , albeit in most cases with lower enrichment ratios . In contrast , PABP2 co-precipitated a larger number of proteins than PABP1 , but with much lower enrichment ratios , possibly reflecting greater promiscuity and interactions with a larger number of heterogenic target mRNAs and hence likely representing isolation of multiple PABP2 complexes . Of the seven proteins most specific to the PABP2 pull-down , three were members of the previously characterised eIF4G1/eIF4E4 complex [41] , namely eIF4G1 , the RNA-binding protein Tb927 . 11 . 350 ( G1-IP2 ) and Tb927 . 11 . 6720 , an mRNA cap guanine-N7 methyltransferase . One of the specific PABP2 targets with high enrichment ratio , CBP110 , is localised to the nucleoplasm [63] ( Fig 3B ) ; this could be a true interacting protein given that PABP2 shuttles between the nucleus and the cytoplasm [36 , 49] . Among the PABP2 interacting proteins were 14 proteins that had enrichment rates of <2 in the PABP1 pull-down and thus appeared specific to PABP2 ( Fig 2B ) . As we observed major differences between the two PABPs in localisation to RNA granules , we analysed the localisation to RNA granules for all 27 proteins that interact with either or both PABPs ( Fig 2B , S1G Table ) . We used published data that used either DHH1 , PABP2 or poly ( A ) as stress granule markers [44 , 49 , 60 , 64] or co-expressed several proteins as eYFP fusions with a mChFP fusion of the stress granule marker protein PABP2 ( Fig 3 and S1 Fig ) . In addition , we obtained information from the genome-tagging project TrypTag [59] ( http://www . tryptag . org , with permission ) . For the TrypTag project , cells are washed in amino acid free buffer prior to imaging and starvation stress granules are therefore visible . The majority of proteins ( 20/27 ) localised to RNA granules , for one protein the localisation remains unknown as tagging failed , and only six proteins did not localise to RNA granules . At least four of the five proteins mostly enriched in the PABP1 pull-down were excluded from granules; these include the unique PABP1-interacting protein RBP23 ( Fig 3A and S1D Fig ) . In contrast , for the majority of the PABP2-interactors there is evidence or proof for stress granule localisation . Only two proteins of the PABP2 interacting proteins are excluded from granules , one is the nuclear protein CBP110 , which is not expected to localise to RNA granules and the other the zinc finger protein ZC3H28 . Thus , the PABP1 complex appears largely excluded from granules , while most of the PABP2 interacting proteins localise to granules , similar to the majority of mRNAs [44] . The data above confirm the strong association of PABP1 with eIF4G3 and eIF4E4 . PABP2 in contrast interacts with eIF4E4 , eIF4G3 , eIF4G1 and two further proteins of the eIF4G1/eIF4E4 complex indicating multiple binding abilities to different paralogs of the eIF4F complex . For a more comprehensive picture , we analysed the enrichment ratios for all members of the translation initiation complex of the four individual experiments ( Fig 4 ) . PABP1 shows strong interactions with eIF4G3 , eIF4E4 in all four experiments and , in particular in low salt conditions , also interaction with eIF4A1 . Interactions with other translation initiation factors and with the two proteins known to interact with eIF4G1 and eIF4E5 [41] are absent or have very small enrichment factors . In contrast , PABP2 co-precipitated all five isoforms of eIF4G under low salt conditions and eIF4G1 and eIF4G3 also under high salt conditions . Similarly , all eIF4E subunits co-precipitated with PABP2 at least under low salt conditions , with the exception of eIF4E2 that has very low abundance [33] . Interestingly , both PABPs clearly co-precipitate each other , indicating that there may be complexes containing both PABPs on the same mRNA protein complex . For Leishmania PABPs such an interdependency has not been observed [36] . Our data contribute towards better understanding of translation initiation control mechanisms in trypanosomes . Demonstration of highly distinct interactomes for the two paralogs of PABP in African trypanosomes indicates discrete functions . Specifically , PABP1 has a small interactome , comprising eIF4E4 and eIF4G3 , and the hypothetical RNA-binding proteins RBP23 and Tb927 . 7 . 7460 . PABP1 , eIF4E4 , eIF4G3 and RBP23 are largely absent from stress granules; the localisation of Tb927 . 7 . 7460 remains unknown . In contrast , PABP2 has a rather more extensive interactome that includes all proteins precipitated with PABP1 , except RBP23 , and most subunits of the eIF4F complex . PABP2 and the majority of its interaction partners localise to starvation stress granules . The impairment in stress granule localisation of the entire PABP1 complex challenges the previous assumption that this complex is the major translation initiation complex involved in bulk mRNA translation . At starvation , polysomes largely dissociate and most mRNAs and proteins involved in mRNA metabolism localise to starvation stress granules [44] . Localisation to granules is the default pathway , and impairment in stress granule localisation is the exception . As the PABP1/eIF4E4/eIF4G3 complex does not locate to stress granules , it is unlikely to regulate translation of bulk mRNAs . Instead , we propose that the PABP1/eIF4E4/eIF4G3 complex is specialised for the regulation of a small subgroup of mRNAs . It is tempting to speculate that this group of mRNAs could be those encoding ribosomal proteins , because these mRNAs are the only group of mRNAs that were found to be excluded from starvation stress granules [44] and these are of small size , consistent with a localisation of PABP1 to lower molecular weight polysomes [49] . The interaction of PABP2 with most eIF4F subunits and many mRNA metabolism proteins indicates a wider substrate specificity for this PABP subunit . The data are consistent with PABP2 being distributed over a range of different translation initiation complexes and mRNAs and thus being responsible for bulk mRNA translation . The eIF4F complex that was identified with highest confidence to bind to PABP2 is eIF4G1/eIF4E5 with its previously identified interactors G1-IP and G1-IP-2 . The fact that both PABPs co-precipitate each other indicates that a separation of the two PABPs to a distinct group of mRNA targets is potentially not strict . A model of the PABP target mRNAs , consistent with the data , is shown in Fig 5 . One limitation of this study is that only one life cycle stage , the procyclic stage , was examined and we can not exclude that the PABP interactomes and their localisations are different in other life cycle stages , for example in blood stream forms . Notably , the functions of many eIF4F complex subunits ( for example eIF4E1 , 2 , 6 ) , and their association with the PABPs remains unsolved . The reason could be that all studies to date , including this one , focus only on the proliferating life cycle stages of the parasites . Translational control may , however , be particular important in G1-arrested stages or during differentiation processes and an analysis of these stages , albeit experimentally challenging , may be highly informative for a more comprehensive picture of the eIF4F/PABP complexes .
T . brucei procyclic Lister 427 cells were cultured in SDM79 medium ( containing fetal bovine serum from Sigma ) . The generation of transgenic trypanosomes was done using standard methods [65] . For starvation , parasites were washed once in one volume PBS and stored in PBS for two hours; the starvation time started at the first contact with PBS . Cell lines expressing PABP1-eYFP or PABP2-eYFP from endogenous loci were previously described [49] . Proteins were expressed as C-terminal ( RBP23 ) or N-terminal ( all others ) eYFP fusion proteins by transfecting trypanosomes with PCR products obtained with the template plasmid pPOTv7-blast-blast-eYFP ( RBP23 ) with oligonucleotides designed as described [66] . All transfected cell-lines co-expressed PABP2-mChFP from the endogenous locus [49] as a marker for starvation stress granules . The plasmid for the expression of a C-terminal 4Ty1 fusion protein was previously described for PABP1 [49] ) and made accordingly for PABP2 [67] . Cells were washed with serum-free SDM79 , fixed with 2 . 4% paraformaldehyde overnight , washed once in PBS and stained with 4′ , 6-diamidino-2-phenylindole ( DAPI ) . Z-stacks ( 100 images , 100-nm spacing ) were recorded with a custom-built TILL Photonics iMIC microscope equipped with a 100× , 1 . 4 numerical aperture objective ( Olympus , Tokyo , Japan ) and a sensicam qe CCD camera ( PCO , Kehlheim , Germany ) using exposure times of 500 ms for fluorescent proteins and 50 ms for DAPI . Images were deconvolved using Huygens Essential software ( SVI , Hilversum , The Netherlands ) and are presented as Z-projections ( method sum slices ) produced by ImageJ [68] . Procyclic trypanosomes were grown to a density of 5–8 x 106 cells/ml . Four litre cultures were harvested in a F14S-6x 250 Y rotor at 1500g at room temperature in four subsequent centrifugations and washed once with 250 ml serum free SDM-79 . Finally , the cells were sedimented by centrifugation ( 1500*g ) into a capped 20 ml syringe placed in a 50 ml Falcon tube . After discarding all supernatant , inserting the plunger and removing the cap the cells were passed slowly into liquid nitrogen in order to form small pellets suitable for subsequent cryomilling . Frozen cells were processed by cryomilling into a fine powder in a planetary ball mill ( Retsch ) [55] . For precipitation , aliquots of approximately 50 mg powder ( corresponding to ~2 x 108 cells ) were mixed with 1 ml ice-cold buffer ( low salt buffer: 20 mM HEPES pH 7 . 4 , 50 mM NaCl , 1 mM MgCl2 , 100 μM CaCl2 , 0 . 1% CHAPS; high salt buffer: 20 mM HEPES pH 7 . 4 , 50 mM NaCl , 1 mM MgCl2 , 100 μM CaCl2 , 150 mM KCl , 0 . 1% CHAPS ) complemented with protease inhibitors ( Complete Protease Inhibitor Cocktail Tablet , EDTA-free , Roche ) . After sonication with a microtip sonicator ( Misonix Utrasonic Processor XL ) at setting 4 ( ~20 W output ) for 2 x 1 second , insoluble material was removed by centrifugation ( 20 , 000 g , 10 min , 4°C ) . The clear lysate was incubated with 3 μl polyclonal anti-GFP llama antibodies covalently coupled to surface-activated Epoxy magnetic beads ( Dynabeads M270 Epoxy , ThermoFisher ) for two hours on a rotator . Beads were washed three times in the respective buffer ( low salt or high salt buffer ) and finally incubated in 15 μl 4 x NuPAGE LDS sample buffer ( ThermoFisher ) , supplemented with 2 mM dithiothreitol , at 72°C for 15 minutes to elute the proteins . The precipitates were analysed on an SDS-PAGE gel stained with Coomassie . For subsequent proteomics analysis six pullout samples were pooled after the final washing step and eluted in 30 μl 4 x NuPAGE LDS Sample buffer , then run 1 . 5 cm into a NuPAGE Bis-Tris 4–12% gradient polyacrylamide gel ( ThermoFisher ) under reducing conditions . The respective gel region was sliced out and subjected to tryptic digest and reductive alkylation . For the precipitation of eIF4E4 and G1-IP2 , essentially the same protocol was used starting from 2 L cultures at a density 8 x 106 cells/ml . The immunoprecipitation was carried out in low salt buffer using 5 ul recombinant , monoclonal dimeric fusion anti-GFP nanobody LaG16-LaG2 [69] coupled to magnetic beads . The same beads , where the antibody coupling step was omitted were used as a control . Eluates were run on a NuPAGE Bis-Tris 4–12% gradient polyacrylamide gel ( ThermoFisher ) under reducing conditions , then subjected to western blotting using standard procedures . 4Ty1 tagged fusion proteins were decorated with monoclonal anti-Ty1 antibody clone BB2 ( Sigma ) at 1:10 , 000 dilution . Quantitation was performed on raw images gathered under nonsaturating conditions using ImageJ [68] and enrichment ratios calculated comparing against uncoupled control beads . Liquid chromatography tandem mass spectrometry ( LC-MS2 ) was performed on a Dionex UltiMate 3000 RSLCnano System ( Thermo Scientific , Waltham , MA , USA ) coupled to an Orbitrap VelosPro mass spectrometer ( Thermo Scientific ) at the University of Dundee FingerPrints Proteomics facility and mass spectra analysed using MaxQuant version 1 . 5 [56] searching the T . brucei brucei 927 annotated protein database ( release 8 . 1 ) from TriTrypDB [70] . Minimum peptide length was set at six amino acids , isoleucine and leucine were considered indistinguishable and false discovery rates ( FDR ) of 0 . 01 were calculated at the levels of peptides , proteins and modification sites based on the number of hits against the reversed sequence database . Ratios were calculated from label-free quantification intensities using only peptides that could be uniquely mapped to a given protein . If the identified peptide sequence set of one protein contained the peptide set of another protein , these two proteins were assigned to the same protein group . The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE [71]ƒ partner repository with the dataset identifier PXD008839 .
|
Poly ( A ) -binding proteins ( PABPs ) bind to the poly ( A ) tails of eukaryotic mRNAs and function in regulating mRNA fate . Many eukaryotes have several PABP paralogs and the current view is that each PABP binds a specific subset of mRNAs . Trypanosoma brucei has two PABPs , and to understand the differential functionality of these paralogs we identified interacting proteins for each . We found unique interactors for both PABPs , and significant differences between the two interaction cohorts . Our data indicate that the two PABP paralogs of trypanosomes have very distinct roles in mediating mRNA fate .
|
[
"Abstract",
"Introduction",
"Results",
"and",
"discussion",
"Materials",
"and",
"methods"
] |
[
"cell",
"physiology",
"rna-binding",
"proteins",
"messenger",
"rna",
"parasitic",
"protozoans",
"organisms",
"trypanosoma",
"brucei",
"fractional",
"precipitation",
"protozoans",
"translation",
"initiation",
"research",
"and",
"analysis",
"methods",
"proteins",
"gene",
"expression",
"precipitation",
"techniques",
"biochemistry",
"rna",
"trypanosoma",
"eukaryota",
"cell",
"biology",
"nucleic",
"acids",
"protein",
"translation",
"genetics",
"biology",
"and",
"life",
"sciences",
"trypanosoma",
"brucei",
"gambiense",
"salting",
"out",
"cell",
"fusion"
] |
2018
|
Comparative proteomics of the two T. brucei PABPs suggests that PABP2 controls bulk mRNA
|
Severe fever with thrombocytopenia syndrome virus ( SFTSV ) , the causative agent for the fatal life-threatening infectious disease , severe fever with thrombocytopenia syndrome ( SFTS ) , was first identified in the central and eastern regions of China . Although the viral RNA was detected in free-living and parasitic ticks , the vector for SFTSV remains unsettled . Firstly , an experimental infection study in goats was conducted in a bio-safety level-2 ( BSL-2 ) facility to investigate virus transmission between animals . The results showed that infected animals did not shed virus to the outside through respiratory or digestive tract route , and the control animals did not get infected . Then , a natural infection study was carried out in the SFTSV endemic region . A cohort of naïve goats was used as sentinel animals in the study site . A variety of daily samples including goat sera , ticks and mosquitoes were collected for viral RNA and antibody ( from serum only ) detection , and virus isolation . We detected viral RNA from free-living and parasitic ticks rather than mosquitoes , and from goats after ticks’ infestation . We also observed sero-conversion in all members of the animal cohort subsequently . The S segment sequences of the two recovered viral isolates from one infected goat and its parasitic ticks showed a 100% homology at the nucleic acid level . In our natural infection study , close contact between goats does not appear to transmit SFTSV , however , the naïve animals were infected after ticks’ infestation and two viral isolates derived from an infected goat and its parasitic ticks shared 100% of sequence identity . These data demonstrate that the etiologic agent for goat cohort’s natural infection comes from environmental factors . Of these , ticks , especially the predominant species Haemaphysalis longicornis , probably act as vector for this pathogen . The findings in this study may help local health authorities formulate and focus preventive measures to contain this infection .
Severe fever with thrombocytopenia syndrome ( SFTS ) is an emerging infectious disease caused by SFTS bunyavirus ( SFTSV ) [1] . This virus was originally identified in 6 provinces of central and northeastern China [1] . Later on , 10 more provinces have been added to the list of endemic regions [2] . Moreover , similar viruses have recently been found to circulate in the United States , South Korea and Japan [3 , 4 , 5] , indicating the genus phlebovirus worldwide distribution . Although most human cases in China are sporadic , SFTS constitutes a threat to public health in China because of its epidemic potential , high fatalities ( 10–16% ) , potential for family cluster or nosocomial outbreaks by means of direct infectious blood or secretion contact , and the difficulties in treatment and prevention [6–10] . Ticks have been implicated as the primary host vector for SFTSV based on several lines of evidence . First , most of the index patients had histories of tick bites before illness onset [1 , 10] , second , the living environments of the patients were heavily infested by ticks [11] , third , several research groups have detected SFTSV-specific nucleotide sequences , or isolated virus from ticks collected from animals , or the environment [12–14] , fourth , spatial and temporal distributions of human cases are consistent with the fluctuation of certain species of ticks in a given endemic area [6] . However , most of the data supporting SFTSV transmission by ticks were obtained from molecular epidemiological surveys , and the vector for SFTSV remains unsettled . In order to explore the role of ticks in the natural cycle of this virus , an experimental as well as a natural infection studies were conducted . Goats were chosen as the study subjects based on the following evidence: 1 ) goats graze on meadow all year round with a potential for close contact to ticks , and they are usually infested with ticks during spring and summer , correlating with the seasonal distribution of human cases [6]; 2 ) several studies have shown a high SFTSV-specific sero-prevalence in goats in the endemic regions ranging from 12 to 80% [11 , 15–17] , indicating goats’ high exposure to the virus under the natural environment . In this study , we detected the viral RNA in both ticks and goats , observed the sero-conversion in goats , and isolated 2 viral strains with the same identity , providing new evidence to support a relationship between ticks and their naïve susceptible hosts in SFTSV transmission .
This study protocol was approved by the animal care committee of Jiangsu Provincial Center for Disease Prevention and Control ( JSCDC ) with permit number 2011–015 . This protocol was designed in accordance with the experimental animal management regulation of People’s Republic of China [18] . All efforts were made to minimize suffering of animals . According to the guideline for prevention and control of SFTS issued by China Ministry of Health [19] , and the written approval from JSCDC , the operation was performed under bio-safety level-2 containment conditions . The strain of SFTSV used in this study , Jiangsu-014 , was originally isolated from a patient in 2010 in Jiangsu Province . It was propagated at 37°C in Vero cells at a multiplicity of infection ( m . o . i . ) of 1 . 0 and cultivated for 10 days . Supernatants containing viral particles were harvested , aliquoted and stored at -70°C until use . The plague-forming unit technique was used to measure the virus titer on Vero cells . Experimental infection study was conducted in a bio-safety level-2 laboratory in JSCDC [19] . A cohort of ten 6-month-old male white goats ( Yangtze river delta ) was imported from SFTS-free region , and no-previous SFTSV exposure was confirmed by lack of detection of viral RNA and antibodies from their sera as described previously[15 , 20] . Of the 10 goats , 5 were subcutaneously inoculated with doses calculated to contain 107 plague-forming units ( p . f . u . ) of virus in a 3-ml volume respectively , and the other 5 were injected with equal volume of phosphate buffer solution ( PBS ) . After inoculation , all animals were reared together , and monitored daily for clinical symptoms . At a set time point ( 9:00 am ) from Day 1 to Day 8 after infection , a daily serum ( 1ml ) , nasopharyngeal and anal swab samples ( 2ml ) were separately taken from each animal in the cohort , and the samples were subjected to RNA extraction by using RNeasy mini kit ( Qiagen ) . TaqMan quantitative real-time RT-PCR was performed as previously described [20] . For antibody detection , a double-antigen sandwich ELISA kit ( Xinlianxin Bio-Tech , China ) was used for all serum samples to test total antibodies ( IgG and IgM ) against SFTSV [15] . The natural infection study site called Xiaogang is a natural village administered by Xuyi County , Jiangsu Province . It is located in the north-west of province capital , Nanjing and in the boundary between Jiangsu and Anhui Provinces , with a latitude of 33°N , and a longitude of 118 . 05°E , respectively ( Fig 1 ) . This village consists of about 20 hills with an average altitude of 150m above the sea level . The subtropical-temperate monsoon climate provides a mean annual temperature of 14 . 7°C and an average annual precipitation of 1005 . 4mm . The well-developed herb and patchy shrub flora constitute the local plant community . Most of the residents in this village are farmers , mainly involved in domestic animal herding , wild herbs collecting and crop planting . The domestic animals , including goats and cattle , are severely infested by ticks in spring and summer . A recent serological survey in this area demonstrated a SFTSV infection prevalence of 3 . 06% for humans and 34 . 80% for animals ( goat , cattle , pig , chicken , goose ) , respectively [16] . This natural infection study started from 19 April , 2012 . Local temperature and relative humidity were 19–25°C and 60–65% , respectively . The vegetation was completely restored , and questing ticks were found on the grass in the open habitat of the study site . Similar to the experimental infection , a cohort of five naïve 6-month-old male white goats ( Yangtze river delta ) was imported from SFTS-free region , and each individual was numbered by installation of ear tag . The goats were allowed to graze on meadows freely in the daytime from 8:00am to 5:00pm , and held in stable at night . To monitor new infection of SFTSV , a daily serum sample ( 1ml ) and a parasitic tick sample ( up to ten ticks collected from host skin ) were taken from each animal . All the samples were stored in 1 . 5-ml micro-centrifuge tubes at -20°C and immediately sent to laboratory for testing . For serum samples , SFTSV specific RNA and antibody were detected by the methods described above . Collected parasitic ticks from each animal were sorted as described previously [21] , ground up by a mixer mill ( Tissue Lyser LT , Qiagen , USA ) according to the manufacturer’s instructions . After a brief centrifugation , the supernatant of lysate ( 300μl ) was used for RNA extraction ( RNeasy mini kit , Qiagen ) , which would be used as template for viral RNA amplification by real-time RT-PCR [20] . The natural infection study was terminated when all goats sero-converted . During the natural infection study period , two dominant arthropods in the area , the free-living ticks and mosquitoes , were collected as described by Schwarz [22] and Turell [23] , respectively , from the grassland where goats grazed . All the samples were subjected to SFTSV RNA detection as described above . For all the collected samples , including goat sera , ticks , and mosquitoes , once the RT-PCR proved positive , the corresponding sample was used for virus isolation [1] . The isolated virus was sequenced and subjected to phylogenetic analysis by using MEGA 5 . 05 software and compared with the published SFTSV strains sequences .
After challenge with SFTSV , only 3 of 5 goats showed a transient viremia on Day 3 post-infection which lasted for less than 24 hours long ( Fig 2 ) . Antibodies became demonstrable on all inoculated animals from Day 4 ( Fig 2 ) . No viral RNA was detected from either nasopharyngeal or anal swab samples in the every inoculated animal ( Fig 2 ) , suggesting that infected animals did not shed virus to the outside through respiratory or digestive tract route . For the control animals , no viral RNA was detected from either swab or serum samples , consistent with a lack of antibody response over the whole testing period ( Fig 2 ) , demonstrating no evidence for virus transmission between animals by close contact mode . No visible clinical signs of infection were observed in the cohort of goats over the testing period . The goats were observed to be gradually infested by ticks when they were farmed in the study site from Day 0 ( April 19 , 2012 ) . After 13 days of incubation , the first goat ( goat-2 ) was found to be infected on Day 14 as both viral-specific RNA and antibody from a serum sample were detected ( Fig 3 ) . Then the viral RNA and sero-conversion were sequentially observed in other animals studied , and the last goat’s ( goat-4 ) infection was on Day 34 ( Fig 3 ) . Similar to the result of experimental infection , the goats were viremic over a very short period ( less than 24hr ) after viral infection , soon occupied by a timely mounting antibody response which effectively controlled the infection . The whole cohort did not show any specific clinical signs of illness , and all survived infection . The fact that all the goats got infected in a short period of time ( within 21days , from Day 14 to Day 34 ) indicates that the pathogen density is high in the local site . Of the 2500 ticks collected from natural environment and goats in the study site , only two species , Haemaphysalis longicornis ( H . longicornis ) and Haemaphysalis doenitzi ( H . doenitzi ) were identified with H . longicornis being the dominant one ( 96 . 04% , 2401/2500 ) ( Fig 4 ) . Viral RNA was detected from H . longicornis , but not H . doenitzi ( Fig 4 ) . 12 samples including 102 H . longicornis ticks ( 4 . 25% , 102/2401 ) were viral RNA positive ( Figs 4 and 5 ) . The time that the viral RNA was first detected from free-living H . longicornis was Day 9 ( Fig 5 ) , 5 days earlier than that from the first infected goat ( Day 14 ) ( Fig 3 ) . Additionally , the viral RNA positivity of parasitic ticks collected from goat-1 , -2 , -4 , and -5 was temporally close to the time point when the hosts were in their transient viremic phases ( Figs 3 and 5 ) , suggesting an effective clearance of circulating virus by serum antibody . No viral RNA was detected from mosquitoes ( 323 in total , including Aedes albopictus and Culex pipiens species ) throughout this study ( Fig 4 ) . Only two isolates were obtained from one infected goat ( goat-1 ) serum and its parasitic tick ( H . longicornis ) sample , respectively , collected at the same day ( Day 22 ) , although virus isolation was attempted on all viral RNA positive samples . Phylogenetic analysis of the S segment of two SFTSV isolates in this study is genetically close to the 8 SFTS patient-derived isolates in 2011–2012 from Jiangsu Province by sharing more than 95% identity ( Fig 6 ) .
In this study , an experimental as well as a natural infection studies in goats were conducted to investigate the role of ticks in the natural cycle of SFTSV . In the natural infection study , we were able to detect the viral RNA in ticks and goats , observe sero-conversion in goats , and also isolate two virus strains from one infected goat serum and one of its parasitic tick samples , respectively , which share a high sequence homology . These findings help to understand the virus transmission mechanism in the natural settings . Unlike human infections , which can cause overt clinical manifestation , and transmit virus by means of human-to-human contact [6] , the goats inoculated with SFTSV exhibited no signs of disease , didn’t shed virus to the outside through respiratory or digestive tract route as described in the experimental infection study , suggesting that the viral transmission cycle could not be established effectively without certain species of arthropods as vehicle in the natural settings . In the natural infection study , the results that the naïve animals were infected after ticks’ infestation and two recovered viral isolates shared high sequence homology , demonstrate that etiologic agent for goat cohort’s natural infection comes from environmental factors . Of the two dominant arthropods in the local site , viral RNA was only detected in ticks rather than mosquitoes . Based on all these findings , we speculate that ticks , especially the predominant species , H . longicornis , act as SFTSV vector , and transmit the virus to naïve goats . However , we could not rule out the possibility that other arthropods besides ticks , e . g . fleas and sandflies , may also be enzootic SFTSV vectors , considering the fact that our natural infection study was carried out in an open environment . The existence of SFTSV in ticks and the transmission between ticks and goats indicate that SFTS might well be a zoonotic disease , although the viral reservoir remains to be discovered . Conceivably , SFTSV is maintained in an enzootic transmission cycle among ticks and wild animals . Any accidental virus “spillover” from this cycle by ticks bite may generate infection or outbreak in humans . Ticks are obligate blood-feeders that require an animal host to survive and reproduce . Although some species of ticks feed on specific host animals , some have a wide range of hosts and transmit many human or animal pathogens [24] . In this study , of the two identified tick species , only H . longicornis was found to carry and transmit SFTSV , but we cannot exclude H . doenitzi’s possible role as SFTSV vector . Since H . doenitzi accounted for only 4% of all the ticks collected , more samples are needed for the assessment of the competence of this tick species in virus transmission . Similar to other vector-borne diseases [25] , the increase in human infection of SFTSV in China is mainly caused by anthropogenic interventions . In China , farmers are recommended by local governments to breed domestic animals , such as sheep , goats , and cattle for economic purposes . The husbandry of these free-ranging animals in SFTSV endemic areas has dramatically promoted ticks population growth and virus expansion [16] . Additionally , ruminant trade may also facilitate the spread of some viruses as seen in the case of the Rift Valley Fever virus in Saudi Arabia and Yemen [26] , we have also found a SFTSV-specific sero-prevalence of 8% , and severe ticks infestation in a flock of goats in SFTSV-free region in China ( personal communication ) , highlighting the threat of this virus expansion into other parts of China and world by live animal trade . In conclusion , as an emerging pathogen circulating mainly in East Asia , with a tendency of spreading to other parts of the world , and in the absence of an effective drug or vaccine , the findings in this study may help local health authorities formulate and focus preventive measures to contain SFTSV infection .
|
Severe fever with thrombocytopenia syndrome virus ( SFTSV ) , a newly identified bunyavirus , has been found to circulate in mainland China , South Korea , and Japan since 2009 . This virus is the etiologic agent for an emerging fatal hemorrhagic fever , severe fever with thrombocytopenia syndrome ( SFTS ) with high fatality . Although ticks have been implicated as the primary host vector indicated by epidemiological surveys , their role in transmitting this virus to the susceptible hosts , including humans , has not been validated . In this study , we conducted experimental and natural infections of goats with SFTSV to explore the role of ticks for this pathogen’s transmission . In the experimental infection study , we have not found any viral transmission within the cohort by close contact between animals . However , in the natural infection study , every member of a naïve goat cohort was observed to get infected sequentially when they were farmed in a SFTSV-endemic site . We detected viral RNA from free-living and parasitic ticks rather than mosquitoes , and from goats after ticks’ infestation . We also observed sero-conversion in all members of the animal cohort subsequently . More importantly , in the natural infection study , two virus strains isolated from one infected goat and its parasitic ticks showed identical S segment sequences of the viral genome . All these findings indicate that ticks , especially the dominant species Haemaphysalis longicornis , probably act as viral vector for this emerging pathogen , SFTSV .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[] |
2015
|
Experimental and Natural Infections of Goats with Severe Fever with Thrombocytopenia Syndrome Virus: Evidence for Ticks as Viral Vector
|
Hepatitis C virus ( HCV ) infection is a leading cause of liver transplantation and there is an urgent need to develop therapies to reduce rates of HCV infection of transplanted livers . Approved therapeutics for HCV are poorly tolerated and are of limited efficacy in this patient population . Human monoclonal antibody HCV1 recognizes a highly-conserved linear epitope of the HCV E2 envelope glycoprotein ( amino acids 412–423 ) and neutralizes a broad range of HCV genotypes . In a chimpanzee model , a single dose of 250 mg/kg HCV1 delivered 30 minutes prior to infusion with genotype 1a H77 HCV provided complete protection from HCV infection , whereas a dose of 50 mg/kg HCV1 did not protect . In addition , an acutely-infected chimpanzee given 250 mg/kg HCV1 42 days following exposure to virus had a rapid reduction in viral load to below the limit of detection before rebounding 14 days later . The emergent virus displayed an E2 mutation ( N415K/D ) conferring resistance to HCV1 neutralization . Finally , three chronically HCV-infected chimpanzees were treated with a single dose of 40 mg/kg HCV1 and viral load was reduced to below the limit of detection for 21 days in one chimpanzee with rebounding virus displaying a resistance mutation ( N417S ) . The other two chimpanzees had 0 . 5–1 . 0 log10 reductions in viral load without evidence of viral resistance to HCV1 . In vitro testing using HCV pseudovirus ( HCVpp ) demonstrated that the sera from the poorly-responding chimpanzees inhibited the ability of HCV1 to neutralize HCVpp . Measurement of antibody responses in the chronically-infected chimpanzees implicated endogenous antibody to E2 and interference with HCV1 neutralization although other factors may also be responsible . These data suggest that human monoclonal antibody HCV1 may be an effective therapeutic for the prevention of graft infection in HCV-infected patients undergoing liver transplantation .
More than 180 million people worldwide are infected with hepatitis C virus ( HCV ) [1] , [2] with over 80% developing chronic disease marked by progressive hepatitis , fibrosis and cirrhosis that often results in liver failure [3] , [4] , requiring transplantation . Unfortunately nearly all donor allografts transplanted into HCV-positive patients become infected with HCV in the early post-transplant period . Standard treatment with interferon–alpha ( IFN–α ) and ribavirin is poorly tolerated and of limited efficacy in liver transplant recipients [5]–[8] and the recently licensed protease inhibitors have not been extensively studied in this population . Many lines of evidence suggest broadly-neutralizing antibody preparations may protect from infection with HCV . Before the identification of HCV as the primary cause of non-A , non-B hepatitis , several randomized trials demonstrated that immune serum globulin ( ISG ) prevented non-A , non-B hepatitis following blood transfusion or sexual exposure [9]–[11] . More recent studies suggest that an early neutralizing antibody response may assist in controlling HCV in the acute phase of infection [12]; however , polyclonal ( Civacir ) [13] and monoclonal antibody ( HCV-AbXTL68 ) [14] have been tested for efficacy in preventing allograft infection in humans without success possibly due to insufficient dose or neutralizing potency . Furthermore , hepatitis B immune globulin ( HBIG ) and cytomegalovirus ( CMV ) immune globulin ( Cytogam ) have both been used successfully to prevent hepatitis B virus ( HBV ) and CMV infection , respectively , after liver transplantation [15] , [16] . In both instances , combination of antibody plus small molecule anti-viral treatment has been shown to be most effective . Lastly , receipt of HBIG preparations containing anti-HCV antibody was correlated with reduced risk of HCV recurrence in patients undergoing liver transplantation [17] . There is clinical precedence , therefore , for the use of antibody-based therapies to prevent recurrence of viral hepatitis after liver transplantation . HCV is a member of the Flaviviridae family and contains a 9 . 6 kb positive-stranded RNA genome which encodes a single polypeptide that is cleaved post-translationally into at least ten different proteins . The major HCV surface glycoproteins , E1 and E2 , form a non-covalent heterodimer that mediates viral entry into target hepatocytes [18] . Defective lentivirus pseudotyped with E1/E2 envelope glycoproteins ( HCVpp ) has been shown to infect hepatocytes [19] , [20] and HCVpp has become a standard model for studying HCV entry inhibitors . Numerous cellular co-receptors including CD81 [21] , claudin-1 [22] , occludin [23] , scavenger receptor class B type I [24] and others [25] have been identified and shown to play essential roles in the interaction of HCV envelope glycoproteins with hepatocytes . The E2 glycoprotein has been shown to directly interact with cellular receptors [25] and provides an attractive target for monoclonal antibody neutralization . HCV1 is a human monoclonal antibody isolated using HuMAb mouse ( Medarex , Inc . , a wholly owned subsidiary of Bristol-Myers Squibb ) technology by immunizing with soluble E2 envelope glycoprotein consisting of amino acids 384–660 of E2 [26] . HCV1 recognizes amino acids 412–423 of the HCV E2 envelope glycoprotein ( E2 epitope I ) , a conserved linear epitope in the N-terminus of E2 . HCV1 neutralizes a broad range of HCV genotypes ( 1a , 1b , 2b , 3a and 4a ) using HCVpp [26] as well as cell culture-infectious HCV ( HCVcc ) JFH1/J6 genotype 2a [27] . Alanine scanning mutagenesis has identified positions 413 and 420 as critical for HCV1 binding . These residues are essentially invariable in the Los Alamos HCV sequence database suggesting that changes at these positions are detrimental to the virus . Of interest , antibody response to E2 amino acids 412–423 in chronically-infected HCV patients has been shown to be quite low or nonexistent [28] and nearly all E2 antibodies isolated from infected humans have been directed against conformational epitopes . Chimpanzees are the only animal other than humans that are permissive to HCV infection and remain the only natural experimental model of HCV infection although expense and ethical concerns limit their use . The chimpanzee provides a model for prevention of initial infection with HCV in the presence of a fully competent immune system that closely models that of humans , an important criteria for the evaluation of a human monoclonal antibody . Also , the levels of HCV replication in chimpanzee are significantly high enough to allow meaningful evaluation of entry inhibitors such as monoclonal antibodies . The infectious inoculum for transmitting HCV to chimpanzees has been carefully characterized [29] . Typically , animals develop viremia shortly after exposure to HCV and can develop long-term infection just as seen in humans . However , the rate at which exposed chimpanzees develop chronic infection is lower than that seen for humans . Chimpanzees have been used to demonstrate that polyclonal antibody has the capacity to prevent initial HCV infection; however , the antibody preparation had to be premixed with the viral inoculum to have a protective effect [30]–[32] . When immunoglobulin containing HCV antibodies were given to chimpanzees shortly after infection as post-exposure prophylaxis , clinical disease was delayed [30] . To our knowledge , no antibody ( monoclonal or polyclonal ) has ever protected a chimpanzee from initial infection when administered to the chimpanzee prior to viral inoculation . On the whole , chimpanzees are the most appropriate model for the study of HCV-directed human monoclonal antibody therapeutics and may also provide valuable toxicology data that can assess acceptability for human studies . To determine the potential of HCV1 as a therapeutic for the prevention of HCV infection , chimpanzees were infused with either 50 mg/kg or 250 mg/kg HCV1 followed 30 minutes later with exposure to HCV H77 genotype 1a . HCV1 , given at 250 mg/kg , prevented HCV infection of a chimpanzee . Also , HCV1 reduced the viral load of both an acutely-infected and long-term chronically-infected chimpanzee to below the limit of detection for 7 to 21 days followed by a rebound in viral titer . Rebounding virus was shown to harbor mutations in the 412–423 epitope that conferred resistance to HCV1 neutralization by preventing HCV1 binding to the E2 glycoprotein . Interestingly , two chronically-infected chimpanzees only modestly responded to HCV1 treatment and preliminary experiments suggested that components in the chimpanzee sera inhibited the effectiveness of HCV1 neutralization .
HCV1 neutralizes HCVpp pseudotyped with E1/E2 derived from a diverse group of genotypes but the mechanism of neutralization had not been determined [26] . Numerous antibodies have been shown to block interaction of E2 with CD81 and this effect correlated with HCV neutralization . The large extracellular loop ( LEL ) of CD81 was expressed in E . coli , purified and coated on ELISA plates . A soluble version of E2 comprising amino acids 384–660 ( E2660 ) containing a ( His ) 6 epitope tag was produced in CHO cells , purified and quantitated . E2660 was incubated with varying concentrations of antibody and the mixture was applied to the CD81 LEL-coated ELISA plates . Binding of E2660 to CD81 LEL was detected using an anti- ( His ) 6 antibody and the results , graphed as percent inhibition of the CD81 LEL/E2660 interaction , are shown in Figure 1 . HCV1 was able to inhibit E2660 interaction with CD81 LEL ( >90% at the highest antibody concentrations tested ) whereas an irrelevant human antibody had no impact on binding . These results demonstrate that HCV1 likely neutralizes HCV by blocking E2 interaction with the CD81 receptor found on target cells . HCV1 was chosen as the lead candidate antibody for use in chimpanzee studies . Three HCV-naïve , healthy chimpanzees were challenged with 32 chimp infectious doses ( CID ) of HCV genotype 1a strain H77 . Thirty minutes prior to infusion with the HCV challenge , chimpanzee #2 received a single intravenous infusion of 50 mg/kg HCV1 and chimpanzee #3 received a single intravenous infusion of 250 mg/kg HCV1 . Chimpanzee #1 did not receive HCV1 and served as the untreated control . It is important to note that the viral inoculum was not pre-incubated with HCV1 antibody ex vivo . Figure 2A is a schematic representation of the sampling and testing performed over the first 42 days of the 20 week study . HCV RNA in the chimpanzee sera was measured using RT-PCR with a lower limit of quantification ( LLQ ) equal to 500 genome equivalents ( Ge ) /ml . Infusion with HCV1 at both 50 mg/kg and 250 mg/kg was well tolerated and no infusion-related reactions were observed . There were no significant changes in hematology , serum chemistries or urinalysis following HuMAb infusion ( data not shown ) . Viral load measurements demonstrated that chimpanzee #1 ( untreated control ) and chimpanzee #2 ( 50 mg/kg HCV1 ) were infected with HCV by 14 days after challenge ( Figure 2B ) . HCV RNA was not detected in chimpanzee #3 ( 250 mg/kg ) during the initial 42 days shown in Figure 2B and remained below the limit of detection for the duration of the study ( 140 days ) at which time the experiment was terminated . The serum concentration of HCV1 was determined at multiple time points in chimpanzee #3 using an ELISA specific for E2 amino acids 412–423 and the terminal half-life of HCV1 was estimated to be 11 . 7 days . These results demonstrate that HCV1 was able to prevent initial HCV infection in one chimpanzee dosed with 250 mg/kg but was not protective in another chimpanzee given 50 mg/kg . To determine if the treatment failure of the chimpanzee that received 50 mg/kg HCV1 was due to mutation in the viral envelope glycoprotein that provided resistance to HCV1 , we employed RT-PCR to amplify and sequence the entire E1/E2 envelope glycoprotein coding region from HCV RNA in the serum of untreated chimpanzee #1 at 21 days after viral challenge and from the 50 mg/kg-treated chimpanzee #2 at 35 days after viral challenge . Different days were chosen for viral sequencing based on the availability of serum samples . A minimum of 8 distinct viral clones from each infected chimpanzee were sequenced and compiled . The derived sequences were translated and the amino acid sequences were compared to the consensus amino acid sequence present in the viral inoculum ( H77 genotype 1a ) . For each chimpanzee #1 and #2 , only three amino acid positions were altered from that of the HCV H77 genotype 1a virus ( Table 1 ) . Both the untreated chimpanzee #1 and the 50 mg/kg-treated chimpanzee #2 had no alterations in amino acids 412–423 when compared to the consensus H77 sequence . Untreated chimpanzee #1 had 3 positions , 394 , 434 and 444 , that were divergent from the H77 consensus sequence . Treated chimpanzee #2 also had three amino acid positions that did not match those found in the H77 consensus sequence and was different than those in chimpanzee #1 , residues 391 , 401 and 608 . Based on the epitope of HCV1 , none of the sequence alterations found in either chimpanzee #1 or chimpanzee #2 were predictive of HCV1-resistant virus . To determine if the lack of virologic control in chimpanzee #2 correlated with resistance to neutralization , lentiviral pseudovirus was generated . This was accomplished by co-transfection of HEK-293T/17 cells with the E1/E2 sequences from the untreated chimpanzee #1 or the 50 mg/kg-treated chimpanzee #2 and the lentiviral backbone with a defective native glycoprotein gene and an engineered luciferase reporter gene ( HCVpp ) . HCVpp was harvested from the culture supernatants , concentrated and stored frozen in aliquots . Varying dilutions of HCV1 were incubated with pseudovirus from chimpanzee #1 ( day +21 ) and chimpanzee #2 ( day +35 ) , CHP1+21–HCVpp and CHP2+35–HCVpp , respectively , and HCV1 neutralization capacity assessed . HCV1 was able to neutralize both CHP1+21–HCVpp and CHP2+35–HCVpp equivalently to H77–HCVpp ( data not shown ) . These data demonstrate that the inability of HCV1 to protect chimpanzee #2 from infection was not due to HCV1-resistant virus . To determine if HCV1 has the capacity to treat an acutely-infected animal , the untreated control chimpanzee #1 was administered 250 mg/kg HCV1 42 days following initial infection with 32 CID H77 virus . Viral load was measured at day +49 and HCV could not be detected in the serum ( Figure 3 ) . At day +56 , two weeks following HCV1 administration , viral load was found to have rebounded and continued to persist through day +72 , at which time the chimpanzee naturally cleared the virus ( Figure 3 ) . These results suggest that HCV1 provided strong neutralizing activity in the setting of recent infection and it is worth noting that HCV1 antibody levels in the serum were still high at the time of viral rebound ( data not shown ) . Given the complete suppression of circulating HCV followed by rapid rebound of the virus in this chimpanzee we assessed whether rebounding virus had developed resistance mutations that would allow for viral escape from HCV1 neutralization . RT-PCR was performed on serum samples from chimpanzee #1 on day +42 before HCV1 treatment and day +56 ( two weeks following treatment ) to determine sequence changes in the viral envelope glycoprotein gene following receipt of HCV1 antibody . The entire E1/E2 coding region was amplified and multiple clones were sequenced . The nucleotide sequence was translated and it was determined that nearly all amino acids were identical between the day +42 and day +56 sequences with the exception of position 415 . In all of the viral sequences identified at day +56 , N415 had been altered to either N415D or N415K , in equal distribution ( 10 of 20 sequenced clones for each mutation ) ( Table 2 ) . The N415K mutation had previously been identified in vitro and H77–HCVpp bearing N415K E1/E2 had been shown to be resistant to HCV1 neutralization [26] . Four E2 residues , 480 , 612 , 615 and 618 , showed minor alterations in that the amino acid usage was mixed at day +56 as compared to that seen at day +42 ( Table 2 ) . To determine if the N415K/D changes resulted in resistance to HCV1 , HCVpp were generated using E1/E2 sequences from chimpanzee #1 identified on both day +42 and day +56 . As expected HCVpp bearing envelope glycoproteins cloned from day +42 virus ( CHP1+42–HCVpp ) were neutralized by HCV1 equivalently to H77–HCVpp ( Table 3 ) . In contrast , HCVpp incorporating E1/E2 glycoproteins from day +56 virus ( CHP1+56–HCVpp ) were completely resistant to HCV1 neutralization even at the highest concentration of antibody tested ( 1000 nM ) . To confirm that resistance was a result of the N415K/D mutation in CHP1+56 E1/E2 envelope glycoprotein , site-directed mutagenesis was performed to revert the 415 position back to asparagine . As expected , this mutation restored sensitivity to HCV1 neutralization ( Table 3 ) . In addition , site-directed mutagenesis was performed to incorporate the N415K or N415D mutation into the E1/E2 gene from the H77 isolate . Both mutations at position 415 resulted in H77–HCVpp that were completely resistant to HCV1 neutralization . These results confirm that mutation at position 415 of the E1/E2 envelope glycoprotein following treatment with HCV1 allowed the virus present in chimpanzee #1 to escape neutralization . Of note , H77–HCVpp containing engineered E1/E2 alterations L480H , P612L , L615F or Y618F ( found in virus present at day +56 ) were all neutralized by HCV1 equivalently to wild–type H77–HCVpp ( data not shown ) . Although HCV1 was able to reduce viral load to below the level of detection in an acutely-infected chimpanzee , this protection was not durable and resulted in the emergence of resistant virus . To determine the efficacy of HCV1 in treating chronically-infected chimpanzees , a single intravenous infusion of HCV1 at 40 mg/kg was administered to three chronically-infected ( >5 years ) chimpanzees ( genotype 1a ) . This dose was the maximum allowable dose as determined by the IACUC at the institution where this specific study was performed . Each infusion was well tolerated and no adverse reactions were observed . Viral load was measured during the course of the study and the results were plotted through day +35 in Figure 4 . Two of the chimpanzees , A and C , had minor reductions in viral load immediately following the infusion , 0 . 5 log10 and 1 log10 , respectively . In contrast , the viral load in chimpanzee B rapidly diminished to below the level of detection by day +5 . HCV RNA remained below the level of detection until day +21 and returned to pre-treatment levels by day +24 . To determine if the rebound in viral load in chimpanzee B was a result of the emergence of escape virus resistant to HCV1 , the entire E1/E2 gene from chimpanzee B at day +35 was isolated and sequenced and compared to sequences obtained from this same chimpanzee prior to HCV1 treatment ( day −8 ) and the results are shown in Table 4 . 100% of the sequences identified at day +35 had two significant alterations that were not seen in any day −8 clones , one at position 417 ( N to S ) and the other at position 444 ( Q to R ) . The mutation N417S is within the epitope for HCV1 ( amino acids 412–423 ) and suggested that this alteration may provide resistance to HCV1 neutralization . To ascertain if the changes at position 417 and 444 conferred resistance to HCV1 neutralization , HCVpp was generated using E1/E2 sequence isolated from chimpanzee B at day +35 ( CHPB+35–HCVpp ) . Varying concentrations of HCV1 were applied to CHPB+35–HCVpp and neutralization potency determined . HCV1 was unable to neutralize CHPB+35–HCVpp at any concentration tested demonstrating that the virus was completely resistant to HCV1 ( Table 3 ) . A neutralizing HCV HuMAb ( 96-2 ) directed against E2 epitope II ( amino acids 432–443 ) was able to neutralize this virus equivalently to H77–HCVpp ( data not shown ) . Also , mutation of the S417 to N417 in CHPB+35–HCVpp restored sensitivity to HCV1 neutralization ( data not shown ) . To confirm that the N417S mutation was responsible for escape from HCV1 neutralization , we engineered this mutation into the H77 virus E1/E2 envelope glycoprotein and generated H77N417S–HCVpp . Interestingly , this mutation rendered the pseudovirus non-infectious and we were initially unable to assess the impact of this mutation on HCV1 neutralization ( Table 3 ) . Given that the N417S and Q444R both arose in the chimpanzee upon viral rebound we created H77 pseudovirus bearing both of these mutations ( H77N417S/Q444R–HCVpp ) . Addition of the Q444R mutation to the N417S mutation restored H77–HCVpp infectivity suggesting that the Q444R mutation was compensatory to N417S . HCV1 was unable to neutralize H77N417S/Q444R–HCVpp at the highest concentration tested ( 1000 nM , Table 3 ) whereas the epitope II-specific antibody , 96-2 , could potently neutralize this HCVpp ( data not shown ) . H77–HCVpp was also generated solely containing the Q444R mutation ( H77Q444R–HCVpp ) and HCV1 neutralized this HCVpp similarly to wild type H77–HCVpp ( Table 3 ) suggesting that the mutation at position 444 does not confer resistance to HCV1 . Since HCV1 was able to reduce virus to below the limit of detection in chimpanzee B but did not provide robust antiviral activity in either chimpanzee A or C , it was formally possible that HCV1 was unable to neutralize the predominant virus in these two chimpanzees . To test this hypothesis , E1/E2 gene sequences were isolated from serum obtained from each of the three chimpanzees 8 days prior to treatment with HCV1 . The dominant sequence from each chimpanzee was cloned into an expression vector and HCVpp were pseudotyped with each synthesized envelope glycoprotein . HCV1 was able to potently neutralize HCVpp harboring envelope glycoprotein from all day −8 chimpanzee sequences ( Table 3 ) suggesting that virus in all three chimpanzees was sensitive to HCV1 . We then examined whether the virus found in chimpanzees A and C had developed mutations that were resistant to HCV1 neutralization shortly after treatment . We sequenced isolated clones from each chimpanzee 8 to 10 days following HCV1 infusion . No significant alterations were found in the coding region for E1/E2 following treatment with HCV1 , and no changes were found in the 412–423 epitope ( data not shown ) . To demonstrate that there were not changes that conferred resistance , we generated HCVpp using E1/E2 isolated from chimpanzee A and C 10 days following treatment with HCV1 . HCV1 had comparable neutralization potency against HCVpp containing E1/E2 isolated either eight days prior or ten days following treatment with HCV1 from both chimpanzees ( data not shown ) . These results suggest that HCV1-resistant virus had not developed in the two chimpanzees with a minor reduction in viral load following HCV1 treatment . Because HCV1 neutralized HCVpp from chimpanzees A and C before and after treatment with equivalent potency , we speculated that components of the serum , present in the chimpanzees with low-level response to HCV1 ( A and C ) but not in serum from the chimpanzee with complete response ( B ) , inhibited HCV1 neutralization . H77–HCVpp was incubated with 100 nM HCV1 or an irrelevant human antibody in the presence of varying concentrations of serum from each of the three chronically HCV-infected chimpanzees as well as serum from an uninfected chimpanzee and infectivity was assessed . Chimpanzee serum alone had a profound and variable impact on H77–HCVpp infection by either enhancing or inhibiting infection depending on the dilution used ( data not shown ) . To overcome the variability introduced by serum in the assay , a ratio of infection in the presence of HCV1 as compared to an irrelevant antibody was determined and plotted ( Figure 5A ) . A ratio of 1 indicates that HCV1 was unable to neutralize HCVpp in the presence of chimpanzee serum and a ratio approaching 0 would indicate complete neutralization of HCVpp . When H77–HCVpp was incubated with 100 nM HCV1 ( fully neutralizing concentration ) in the presence of a 1∶16 serum dilution from chimpanzee A or C , the infection , expressed as a ratio of HCV1 to irrelevant antibody , was 0 . 72 and 0 . 4 respectively demonstrating significant HCVpp infection suggesting that HCV1 neutralization was impaired . In contrast , this ratio was only 0 . 15 for serum from both the uninfected chimpanzee and chimpanzee B , consistent with a low level of HCVpp infection , suggesting that HCV1 was able to neutralize HCVpp effectively . Infection in the absence of chimpanzee serum resulted in a ratio near 0 . 1 ( data not shown ) suggesting effective HCV1 neutralization . These data demonstrate that a component in the serum from both chimpanzees A and C interfered with HCV1 neutralization of H77–HCVpp and this factor is not present in the serum of chimpanzee B or an uninfected chimpanzee . We postulated that non-neutralizing antibody present in the serum of chimpanzees A and C may be responsible for abrogating the neutralizing activity of HCV1 . We performed an ELISA to measure the serum concentration of antibody directed against soluble E2 envelope glycoprotein ( E2660 ) . E2660 was coated on ELISA plates and varying dilutions of chimpanzee serum was applied and detected with goat anti-human polyclonal antibody . Plates were developed and the results plotted in Figure 5B . Chimpanzees A and C had high serum concentrations of antibody directed against E2660 whereas chimpanzee B had very low antibody levels to this protein ( Figure 5B ) . These data suggest that endogenous antibody specific to E2 in the serum of chimpanzees A and C may be capable of interfering with neutralization of HCV with the HCV1 antibody and potentially explains the differences in outcomes after treatment of the animals with chronic infection . Alternatively , the findings may be related to the dose of antibody given , and as was seen in the prevention study , a higher dose of antibody could have been effective for HCV neutralization . We previously reported that E2660 soluble glycoprotein harboring an N415K mutation produced by CHO cells in defined , serum-free media ( CD-CHO , Invitrogen ) was bound strongly by HCV1 yet HCV1 could not neutralize H77N415K–HCVpp [26] . This result led to speculation that HCV1 binding required amino acid residues distant to amino acids 412–423 , although we believed this to be unlikely . HCV1 binding to E2 harboring mutations at positions 415 and 417 and produced under various cell culture conditions was assessed . HCVpp are typically produced from HEK-293T/17 cells in bovine serum-containing media and we speculated that E2 glycoprotein produced in media containing serum was not equivalent to E2660 ( H77–derived ) produced in defined media . In fact , soluble glycoprotein produced in CD-CHO had significantly different mobility than E2660 produced in serum-containing media when analyzed by SDS-PAGE ( data not shown ) . We performed binding studies ( ELISA ) to determine if HCV1 had differential recognition of E2660 depending on the method of protein production . Wild-type E2660 and E2660 with an N415K mutation ( E2660–N415K ) were produced from CHO cells in either serum-containing media or serum–free CD-CHO media . The four proteins were coated on ELISA plates and HCV1 binding at varying concentrations assessed . HCV1 strongly bound to wild type E2660 regardless of the type of media used to grow the E2660–producing CHO cells ( Figures 6A and 6B ) . HCV1 bound strongly to E2660–N415K produced in defined media ( Figure 6C ) though less well than to wild-type E2660 produced under the same conditions . However , E2660–N415K produced from CHO cells in serum-containing media was bound very weakly by HCV1 showing minimal binding even at 5 µg/ml ( Figure 6D ) . Both the anti– ( His ) 6 specific antibody and the epitope II-specific 96-2 antibody recognized all proteins equivalently ( Figure 6A–D ) . These experiments were repeated using E2660–N415D and E2660–N417S ( data not shown ) and we observed the same result as shown for E2660–N415K . These data demonstrate that mutation at amino acids 415 or 417 of the E2 envelope glycoprotein abrogate HCV1 binding and thus neutralization capacity .
Chronic infection with HCV leads to liver failure in 5–10% of patients with liver transplant being the only treatment option . In the setting of liver transplantation for chronic hepatitis C nearly all liver allografts become infected with HCV and the long-term survival of the graft is compromised by infection [33]–[39] . The clinical course of recurrent hepatitis C is often aggressive resulting in accelerated allograft cirrhosis and increased risk of graft failure and death [34] , [40] , [41] . Treatment with pegylated-IFN–α and ribavirin is often attempted , but the limited efficacy is balanced against the significant risk of adverse events in this population [8] , [42] . The first FDA-licensed HCV direct-acting antiviral agents , boceprevir and telaprevir , are indicated for chronic HCV infection but their tolerability profile will likely limit their use in patients that have undergone liver transplantation . Novel , safe and efficacious therapies to prevent HCV recurrence are needed , particularly because of the limited supply of organs for transplantation . Treatment during liver transplantation with human monoclonal antibody capable of neutralizing HCV infection of the liver allograft is an attractive potential therapy because of the favorable tolerability profile of human monoclonal antibodies , IV administration , and precedent of HBV neutralization during liver transplant with human polyclonal antibodies . Because HCV does not integrate into the host genome , viral clearance and cure should be attainable in theory . However , past clinical and chimpanzee experiments have given mixed results for the success of HCV specific antibodies for prevention of infection . The HCV1 human monoclonal antibody can potently neutralize a broad range of HCV genotypes in the HCVpp system and recognizes a conserved linear epitope in the E2 envelope glycoprotein ( epitope I , amino acids 412–423 ) [26] . Neutralization of circulating virus with a monoclonal antibody prior to engraftment may prevent liver infection in the transplant patient population . As a surrogate for graft infection , we used a model of initial infection in chimpanzees . HCV1 , at doses of 50 mg/kg or 250 mg/kg , was infused into chimpanzees 30 minutes prior to challenge with 32 CID of H77 genotype 1a virus . The chimpanzee receiving 250 mg/kg was completely protected from HCV infection whereas the chimpanzee receiving 50 mg/kg HCV1 was not protected . It is unclear if the higher dose of HCV1 is required for protection from initial infection in chimpanzees given that only one chimpanzee per cohort was assessed . It is formally possible that HCV1 can only protect a subset of chimpanzees from initial infection regardless of dose and the limited number of chimpanzees tested did not allow for statistical significance . Also , it was formally possible that the chimpanzee that received the 250 mg/kg dose was refractory to HCV infection; however , in subsequent experiments , this chimpanzee was shown to be permissive to HCV genotype 1b infection ( data not shown ) . Sequence analysis of the virus from the chimpanzee that received the lower dose of HCV1 did not reveal any sequence alterations that would be predicted to resist HCV1 neutralization . In fact , the sequences identified in this chimpanzee were nearly an exact match for the H77 consensus sequence and any minor variation was also observed in the untreated chimpanzee . Given the expense and ethical issues associated with chimpanzee studies , true dose-ranging studies of HCV1 to determine a therapeutic level were not performed . To our knowledge , this is the first demonstration that pre-administered antibody can protect chimpanzees from initial infection with HCV . This study is the first report of an antibody directed against E2 amino acids 412–423 tested in chimpanzees and it is possible that this epitope is the appropriate target of a protective monoclonal antibody . This is supported by the fact that polyclonal antibody , which does not protect from initial infection in this model , has been shown to contain very little antibody directed against the HCV1 epitope . Another possibility is that we delivered a higher dose of antibody than has been delivered previously and this fact led to our success . At this time it is unclear which of these possibilities are correct and experiments directly comparing HCV1 to other neutralizing antibodies would need to be performed . The anti-viral neutralizing activity of HCV1 was also demonstrated in acutely-infected and chronically-infected chimpanzees . For the initially untreated chimpanzee that developed an acute infection ( chimpanzee #1 ) , a dose of 250 mg/kg was able to suppress viral load to an undetectable level before viral rebound was noted 14 days following HCV1 infusion . For the chronically-infected chimpanzees , a single dose of 40 mg/kg was able to suppress viral titer to below the level of detection for 21 days in one chimpanzee with the other two chimpanzees demonstrating a 0 . 5 and 1 . 0 log10 reduction in viral load within four days of treatment . For the two chimpanzees treated with HCV1 ( acutely- and chronically-infected ) in which the viral load was suppressed to undetectable levels , escape mutations were detected in the 412–423 E2 epitope upon viral rebound . Specifically , N417S and N415K/D mutations were observed in 100% of the HCV sequences isolated from the chronically-infected and acutely-infected chimpanzee , respectively , indicating strong selective pressure of the antibody on the virus of both responding chimpanzees . Not surprisingly , HCV1 was unable to neutralize N415K/D–HCVpp bearing glycoproteins derived from the acutely-infected chimpanzee or H77–HCVpp engineered to contain these mutations . HCV1 was also unable to neutralize HCVpp pseudotyped with E1/E2–N417S derived from the chronically-infected chimpanzee . Interestingly , N417S mutant H77–HCVpp was not infectious in our culture system . However , the introduction of the Q444R in addition to the N417S mutation in H77–HCVpp restored infectivity and HCV1 was then unable to neutralize this pseudovirus . The N417S mutation identified in chimpanzee B following viral rebound was accompanied by a Q444R mutation which presumably was compensatory to maintain infectivity of the virus bearing this N417S mutation . It is not clear why two of the chronically-infected chimpanzees had an incomplete virologic response to treatment with HCV1 . Virus isolated from these two poorly-responding chimpanzees had E1/E2 glycoprotein sequences nearly identical to the viral sequences prior to HCV1 infusion suggesting that escape virus did not develop in these chimpanzees . Also , HCV1 was able to neutralize HCVpp bearing envelope glycoproteins derived from both poorly-responding chimpanzees with a similar potency to H77–HCVpp indicating that HCV1 had the capacity to neutralize virus from these two chimpanzees . Interestingly , in the treatment model of chronic infection , sera from the two poorly-responding chimpanzees , but not the responding chimpanzee , were able to inhibit HCV1 neutralization of H77–HCVpp . These chimpanzees received a dose of 40 mg/kg HCV1 rather than the 250 mg/kg used in the prevention study due to IACUC restrictions . It is possible that a higher dose of HCV1 may have reduced the viral load in all three chronically-infected chimpanzees rather than one of three by overcoming the neutralization inhibition imposed by the poorly-responding chimpanzee sera . Given previous literature reports , the inhibition could be due to competing non-neutralizing antibodies , serum lipoproteins , or other as yet unidentified factors . Due to the small number of animals and the limited supply of serum from the animals , we were not able to determine the nature of the serum factor that inhibited the neutralization of HCVpp . It is possible that endogenous anti-HCV antibodies directed against epitope II of E2 may interfere with HCV1 potency . It has been suggested that epitope II-specific antibodies in both chimpanzees and humans inhibit epitope I-directed antibodies [43] , [44] though these data have recently been challenged [45] . Using ELISA we demonstrated that chronically-infected chimpanzees with an incomplete virologic response to HCV1 did possess high concentrations of anti-E2660 antibody in contrast to the low concentrations found in the chimpanzee with complete virologic response . It should be noted that we also performed ELISA on the chimpanzee serum to determine the titers of epitope I- and II-specific antibodies and none of the three chimpanzees had appreciable antibody to these two epitopes ( data not shown ) . If endogenous chimpanzee non-neutralizing antibodies do interfere with HCV1 , these antibodies are most likely not directed to epitope II but to other regions of the envelope glycoproteins . We were unable to deplete chimpanzee antibody from the serum and perform HCVpp assays due to the limited supply of chimpanzee serum from this study . It has also been hypothesized that serum lipoproteins coating HCV virions present in sera may interfere with epitope I-directed antibodies such as HCV1 [46] , [47] though we have not been able to reproduce these results in our laboratory ( Babcock , unpublished observations ) . The chimpanzees that did fully respond to HCV1 treatment would have to have differences in lipoprotein/virion coating from the poorly-responding chimpanzees and this appears to be unlikely . To determine the exact cause of the failure of HCV1 to suppress viral load in 2 of 4 chimpanzees ( antibody , lipoproteins , other factors , etc . ) , additional experiments would need to be performed on the chimpanzee serum; however , the supply has been exhausted . As HCV1 is intended for human therapeutic use , a relevant question is if some or all sera samples from HCV-infected humans are able to inhibit the neutralizing activity of HCV1 in HCVpp-based assays . Therefore , we conducted experiments on sera from multiple HCV-infected humans to assess interference with HCV1 neutralization of HCVpp ( data not shown ) . The HCVpp neutralizing titers in these sera were so high that dilutions >1∶1000 needed to be employed to distinguish HCV1 activity from endogenous neutralizing activity against HCVpp , i . e . the human sera neutralized HCVpp potently which masked HCV1 neutralization . Mixed with human serum at dilutions >1∶1000 , HCV1 neutralization of HCVpp was not effected ( Babcock , unpublished data ) . However , chimpanzee serum inhibition was only significant at dilutions of <1∶500 . As such , it is not clear if human sera contains factors that inhibit HCV1 neutralization at dilutions <1∶1000 . It is possible that human sera contains inhibiting factors similar to those present in two of the chronically-infected chimpanzees; however , these factors could be diluted in the setting of a liver transplant with large blood and fluid losses and replacements or may only be relevant for the in vitro HCVpp system . One limitation of this study was the sole use of HCVpp as a model in vitro system for understanding HCV1 neutralization capacity and the impact of chimpanzee serum factors on HCV1 neutralization . Cell culture-infectious HCV ( HCVcc ) has been developed which recapitulates all steps of the viral life cycle including entry [48]–[50] . HCVcc have been shown to be infectious in vivo [51] and as such are considered a more appropriate viral infection system than HCVpp . Numerous differences exist between HCVpp and HCVcc including , but not limited to , envelope glycoprotein density [52] , glycan incorporation [53] , and interaction with factors present in serum [52] , [54] . It is unclear how these differences between HCVpp and HCVcc would impact the current results . Clearly , the results demonstrating E2 glycoprotein escape variants using HCVpp would be expected to repeat using the HCVcc system . However , HCV1 neutralization capacity of HCVpp was adversely affected by the presence of poorly-responding chimpanzee serum . If HCVcc interact with serum factors differently than HCVpp it is possible that we may observe a different impact of serum on HCV1 neutralization . HCVcc may be more resistant to HCV1 than HCVpp in the presence of chimpanzee serum but the possibility exists that HCVcc may be more sensitive to neutralization . In either case , further study using HCVcc to validate and confirm the current results obtained with HCVpp is warranted . In the treatment study , using chronically-infected chimpanzees , we administered a dose of 40 mg/kg HCV1 rather than the 250 mg/kg used in the prevention study due to IACUC restrictions . It is possible that a higher dose of HCV1 may have reduced the viral load in all three chronically-infected chimpanzees rather than one of three by overcoming the neutralization inhibition imposed by the poorly-responding chimpanzee sera . 40 mg/kg was the highest dose allowed at this specific animal facility and it is not possible to test again with higher doses of HCV1 . We previously demonstrated that HCV1 was able to bind peptides and bacterially-expressed fusion proteins comprising E2 amino acids 412–423 . The strong binding of HCV1 to both peptides and bacterially-expressed proteins suggests that glycosylation at positions 417 and 423 ( two putative E2 glycosylation sites ) is not required for binding [26] . Interestingly , HCV1 was able to bind soluble E2660 harboring N415K mutations even though neutralization capacity against HCVpp bearing this same alteration was lost [26] . Here we show that HCV1 is able to bind E2660–N415K protein produced from defined medium but not from serum-containing medium , the same media used for the production of HCVpp . These data illuminate our previous results and demonstrate that resistance mutations in the 412–423 epitope do indeed abrogate HCV1 binding to E2 and likely HCVpp produced in serum-containing media . It is unclear why E2660–N415K produced in defined medium can be bound by HCV1 , but the altered migration of the protein in SDS-PAGE suggests E2660 is glycosylated differentially in defined medium . We speculate that complete and appropriate glycosylation at positions 417 and/or 423 impacts the ability of HCV1 to bind E2 envelope glycoprotein containing mutations at positions 415 and 417 , but not wild-type E2 . Of note , the N417S mutation eliminates the N-linked glycosylation site at position 417 but does introduce a potential O-linked glycosylation site . The binding of HCV1 to E2660–N417S produced in serum-containing media may be impacted by this possible O-linked glycosylation . Also , differential glycosylation of amino acids distant from amino acids 412–423 could possibly be in proximity to the HCV1 epitope when positions 415 and 417 are mutated disrupting HCV1 binding . Future studies of HCV1 ( and possibly other epitope I antibodies ) binding and epitope mapping should avoid the use of peptides and bacterially-produced proteins that lack glycosylation as this may lead to erroneous results . HCV1 effectively prevents acute HCV infection and significantly reduces viral load to undetectable levels in a proportion of chronically-infected chimpanzees at doses ranging from 40 mg/kg to 250 mg/kg . A phase 1 trial in healthy adults has been completed and HCV1 proved to be safe and well tolerated . Studies are now in progress to examine the efficacy of HCV1 in the prevention of HCV infection of the donor liver in the liver transplant setting .
All experiments using chimpanzees were performed in accordance with the Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committees at both the Southwest National Primate Research Center at the Texas Biomedical Research Institute ( Animal Welfare Assurance Number A3082-01 ) and New Iberia Research Center ( Animal Welfare Assurance Number A3029-01 ) . CHO-K1SV cells ( Lonza ) were grown in CD-CHO media ( Invitrogen ) or Dulbecco's modified eagle medium supplemented with 10% fetal bovine serum ( DMEM/FBS ) . HEK-293T/17 cells ( ATCC ) were grown in DMEM/FBS supplemented with 100 IU penicillin/streptomycin . CHO-K1SV cells in CD-CHO were grown in Erlenmeyer flasks with constant orbital shaking . All cells were cultured at 37°C in air supplemented with 5% CO2 . HCV1 heavy and light chain genes were cloned into a mammalian expression vector containing a glutamine synthetase ( GS ) gene as previously described [26] . CHO-K1SV cells were electroporated with the HCV1 expression vector and transfectants were selected using 50 µM methionine sulfoximine ( MSX ) . Transformants were screened for antibody expression and the highest expressing clone ( >500 mg/L ) was selected for antibody production in bioreactors . Culture supernatant from bioreactors was clarified using centrifugation and filtration and antibody was purified using protein A followed by ion exchange chromatography . Antibody was formulated at 10 mg/ml in 20 mM citrate/150 mM NaCl/0 . 025% Tween-80 and was determined to be >98% antibody monomer lacking detectable endotoxin . Antibody 96-2 is a human monoclonal antibody directed against HCV antigenic site II ( amino acids 432–443 ) which was generated and characterized during the isolation of HCV1 [26] . 96-2 was purified from hybridoma culture supernatant which was incubated with protein A sepharose beads ( GE Healthcare ) for 2 hours at room temperature while rocking . Beads were removed by column filtration , washed with PBS and antibody eluted with 100 mM glycine pH 2 . 8 . Eluate was dialyzed against PBS and concentrated using an Amicon YM-30 centriprep concentrator as described by the manufacturer . Purified antibody was filter sterilized and protein concentration determined by spectrophotometry . Anti- ( His ) 6 monoclonal antibody is a mouse monoclonal antibody that was developed in house and purified from hybridoma supernatants as described above . Chimpanzees were maintained in accordance with the Guide for the Care and Use of Laboratory Animals at the Southwest National Primate Research Center at the Texas Biomedical Research Institute ( Animal Welfare Assurance Number A3082-01 ) . The protocol ( #1070 PT0 ) was approved by the center's Institutional Animal Care and Use Committee . Healthy female chimpanzees were untreated or infused with 50 mg/kg or 250 mg/kg HCV1 antibody 30 minutes prior to inoculation with 32 chimp infectious doses ( 32 CID ) of H77 genotype 1a HCV serum [31] , [32] . Animals were followed for a total of 20 weeks following delivery of the HCV inoculum . Serum samples were obtained at varying time points to measure viral load using a TaqMan assay as well as HCV1 serum concentration using ELISA . In addition , clinical and safety laboratory testing was performed throughout the study ( 20 weeks ) . Chimpanzees were housed and maintained according to the Guide for the Care and Use of Laboratory Animals at the New Iberia Research Center ( Animal Welfare Assurance Number A3029-01 ) . The protocol ( #2011-8741-027 ) was approved by the center's Institutional Animal Care and Use Committee . IACUC approval restricted HCV1 dosing to 40 mg/kg in this study . Three female chimpanzees , chronically infected with H77 genotype 1a HCV , were infused intravenously with 40 mg/kg HCV1 over one hour . Chimpanzees were followed for 35 days with blood samples collected for safety labs ( hematology and serum chemistries ) . Viral load ( quantitative PCR – lower limit of quantification of 15 IU/ml ) and circulating antibody concentration ( ELISA ) to E2 amino acids 412–423 were measured at multiple time points during the five week study period . The RNeasy kit ( Qiagen ) was employed following the manufacturer's instructions to isolate RNA from 100 µl to 1 ml of serum and purified RNA was stored as frozen aliquots . RT-PCR was performed on 3 µl of isolated RNA using the Superscript III One-Step RT-PCR System with Platinum Taq ( Life Technologies ) as described by the manufacturer . The forward oligonucleotide consisted of a 5′ overhang and HindIII site followed by a sequence corresponding to a region upstream of the E1 coding region ( 5′– GCT TAG CAA GCT TCG CCG ACC TCA TGG GGT ACA TAC CGC TCG −3′ ) . The reverse oligonucleotide incorporated a 5′ overhang and XbaI site followed by a sequence complimentary to the terminal coding region of E2 ( 5′– CGC TTG CTC TAG ACG AGG TTC TCC AAA GCC GCC TCC GCT TGG −3′ ) . The resulting RT-PCR product , consisting of 1891 base pairs , was digested with HindIII and XbaI and ligated to pcDNA3 . 1 ( Invitrogen ) . DH5α Escherichia coli cells were transformed with the ligation reaction , plated to LB agar plates containing ampicillin , grown overnight at 37°C and colonies were assessed for E1/E2 PCR product incorporation . Positive clones ( 5–15 total ) were sequenced by the Sanger method and analyzed using Vector NTI software . Prototypical genotype 1a E1/E2 envelope glycoprotein genes were amplified from an H77 expression plasmid ( p90HCVconsensuslongpU ) obtained through the AIDS Research and Reference Program , Division of AIDS , NIAID , NIH from Dr . Charles M . Rice via Apath , LLC . The PCR product contained 5′ HindIII and 3′ XbaI restriction sites and were cloned into pcDNA3 . 1 containing a 3′ ( His ) 6 epitope tag . H77 soluble E2660 was cloned , expressed and purified as previously described [26] . To create expression constructs of E1/E2 isolated from chimpanzee serum , a forward primer ( E1F1 ) containing a 5′ HindIII site , Kozak sequence and start codon followed by sequence complimentary to 5′ end of the E1 gene was designed ( 5′– GAT GAG CAA AGC TTG CCG CCA CCA TGG CCA CCG GCA ACC TGC CCG GCT G −3′ ) . A reverse primer , E2R1 , containing a 5′ XbaI site followed by sequence complimentary to the 3′ end of the E2 gene was also synthesized ( 5′– GCA TTC ACT CTA GAC GCC TCC GCC TGG GAG ATC AGC −3′ ) . PCR with primers E1F1 and E2R1 using template consisting of cloned PCR products in pcDNA3 . 1 obtained during the sequencing of virus from chimpanzee serum was performed . Cloned E1/E2 was sequenced and confirmed . Mutagenesis of the full-length E1/E2 and E2660 expression plasmids was performed using the Quick Change II Site-Directed Mutagenesis kit ( Stratagene ) following the manufacturer's instructions . Following mutagenesis the E1/E2 encoding region was sequenced to confirm the introduction of the desired mutation as well as integrity of the entire coding sequence . RNA was extracted from HEK-293T/17 cells using the RNeasy kit ( Qiagen ) as described by the manufacturer . RT-PCR was employed to amplify nucleotides 342 to 600 of the CD81 gene which encodes for the CD81 large extracellular loop ( LEL ) . The amplicon was cloned into pGEX-6P ( GE Healthcare ) in frame with the N-terminal glutathione S-transferase ( GST ) and C-terminal myc epitope tag and expressed in BL-21 E . coli . Bacteria were lysed and the protein purified using glutathione sepharose chromatography . Purified CD81 LEL-GST was cleaved with PreScission protease to remove the GST tag and CD81 LEL was isolated by size-exclusion chromatohgraphy . Purified CD81 LEL ( 1 µg/ml ) was coated on ELISA plates overnight at 4°C and subsequently washed . E2660 containing a C-terminal ( His ) 6 tag ( 5 µg ) was incubated with varying concentrations of HCV1 or an irrelevant human antibody for 1 hour at room temperature . Complexes were added to the CD81 LEL-coated ELISA plate and incubated for 1 hour at room temperature . E2660 binding was detected using an anti- ( His ) 6 followed by goat anti-mouse IgG -alkaline phosphatase ( AP ) conjugate ( 1∶5000 , Jackson Immunoresearch ) and developed with p-nitrophenyl phosphate disodium salt ( PNPP ) at 1 mg/ml in 1 M diethanolamine . Absorbance ( 405 nm ) was analyzed using Molecular Devices Emax plate reader with the Softmax software . Chimpanzee serum and purified HCV1 was assessed for H77–E2660 ( mutant or wild-type ) binding using ELISA . 96-well microtiter plates were coated with 0 . 5–2 µg/ml of E2660 in PBS overnight at 4°C . 100 µl chimpanzee serum at varying dilutions was added to the wells and incubated at 22°C for 2 hours . Antibody binding was detected using an anti-human IgG-AP conjugate ( 1∶5000 , Jackson Immunoresearch ) followed by PNPP at 1 mg/ml in 1 M diethanolamine . Absorbance ( 405 nm ) was analyzed using Molecular Devices Emax plate reader with the Softmax software . Pseudovirus was generated employing an HIV backbone that contained a mutation to prevent HIV envelope glycoprotein expression and a luciferase gene to direct luciferase expression in target cells ( pNL4-3 . Luc . R-E- ) [55] , obtained through the AIDS Research and Reference Program , Division of AIDS , NIAID , NIH from Dr . Nathaniel Landau . HCV E1/E2 glycoproteins were provided in trans by co-transfection of HEK-293T/17 cells with pcDNA-E1/E2-H77-1a ( prototypical H77 sequence ) or pcDNA-chimp-E1/E2 ( E1/E2 from chimpanzee serum ) with pNL4-3 . Luc . R-E- . Supernatant containing virus particles was harvested 48 to 72 hrs post-transfection , concentrated using Centricon 70 concentrators , aliquoted and stored frozen at −80°C . Pseudovirus was pre-incubated with varying concentrations of antibody for 1 hr at room temperature before adding to Hep3B cells . After incubation for 72 hrs , infection was quantitated by luciferase detection with BrightGlo luciferase assay ( Promega ) and read in a Victor3 plate reader ( Perkin Elmer ) for light production . For HCVpp assays to measure chimpanzee serum inhibition of HCV1 neutralization the above procedure was employed with minor modifications . HCVpp were pre-incubated with 100 nM HCV1 or irrelevant human antibody in the presence or absence of varying concentrations of chimpanzee serum for 1 hr at room temperature before adding to Hep3B cells . Light output was quantified for both irrelevant antibody and HCV1 at each concentration of chimpanzee serum and the ratio of light output of HCV1 to irrelevant mAb samples was determined .
|
The majority of individuals infected with hepatitis C virus ( HCV ) become chronically infected and many go on to develop liver failure requiring liver transplantation . Unfortunately , the transplanted liver becomes infected with HCV in nearly 100% of transplant patients . Current treatments for HCV are poorly tolerated after liver transplantation and graft health is compromised by infection . We have developed a monoclonal antibody called HCV1 that blocks HCV from infecting liver cells in culture . Using chimpanzees as a model for HCV infection , we demonstrate that HCV1 has the ability to prevent HCV infection . We also show that HCV1 can treat chimpanzees chronically infected with HCV and reduce plasma viral load to below the level of detection for a period of 7 to 21 days . The virus that reemerges in the treated chimpanzees was resistant to HCV1 neutralization demonstrating target engagement . Given the ability of HCV1 to protect chimpanzees from HCV infection , we speculate that HCV1 may be beneficial in HCV- infected patients undergoing liver transplant .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biotechnology",
"animal",
"models",
"of",
"infection",
"antivirals",
"viral",
"evolution",
"mechanisms",
"of",
"resistance",
"and",
"susceptibility",
"viral",
"immune",
"evasion",
"virology",
"immunology",
"biology",
"microbiology",
"drug",
"discovery",
"immunoglobulins"
] |
2012
|
Human Monoclonal Antibody HCV1 Effectively Prevents and Treats HCV Infection in Chimpanzees
|
Schistosome eggs must traverse tissues of the intestine or bladder to escape the human host and further the life cycle . Escape from host tissues is facilitated by secretion of immuno-reactive molecules by eggs and the formation of an intense strong granulomatous response by the host which acts to exclude the egg into gut or bladder lumens . Schistosome eggs hatch on contact with freshwater , but the mechanisms of activation and hatching are poorly understood . In view of the lack of knowledge of the behaviour of egg hatching in schistosomes , we undertook a detailed dynamic and correlative study of the hatching biology of Schistosoma japonicum . Hatching eggs of S . japonicum were studied using correlative light and electron microscopy ( EM ) . The hatching behaviour was recorded by video microscopy . EM preparative methods incorporating high pressure freezing and cryo-substitution were used to investigate ultrastructural features of the miracidium and extra-embryonic envelopes in pre-activated and activated eggs , and immediately after eggshell rupture . Lectin cytochemistry was performed on egg tissues to investigate subcellular location of specific carbohydrate groups . The hatching of S . japonicum eggs is a striking phenomenon , whereby the larva is liberated explosively while still encapsulated within its sub-shell envelopes . The major alterations that occur in the egg during activation are scission of the outer envelope-eggshell boundary , autolysis of the cellular inner envelope , and likely hydration of abundant complex and simple polysaccharides in the lacunal space between the miracidial larva and surrounding envelopes . These observations on hatching provide insight into the dynamic activity of the eggs and the biology of schistosomes within the host .
The pathology associated with chronic schistosomiasis is related to host responsiveness to antigens released by schistosome eggs entrapped in tissues [1] , [2] . The primary source of the secreted antigens in developing eggs is a distinct extra-embryonal layer that surrounds the differentiating embryo ( miracidium ) [3] . This layer has been variously called the Reynold's layer or the outer envelope ( OE ) , and is derived in early development from cells that delaminate from the embryo [3]–[5] . Secreted antigens , released through preformed pores in the shell [3] , [4] , consist of a range of peptides and glycans , particularly those containing core mucins and fucose [3] , [6] , [7] . The schistosome eggshell is a highly cross-linked protein matrix derived from vitelline cells and fashioned in the ootype , an elaboration of the female reproductive system situated between ovary and uterus [8] , [9] . Eggshells are formed from precursor proteins , belonging to three families of tyrosine ( Tyr ) - and glycine-rich molecules [9] . The strong insoluble cross links of the shell are formed by the action of tyrosinases , which catalyze the hydroxylation of Tyr to dihydroxy-phenylalanine , which is subsequently oxidised to dopaquinone for cross linking [10] . An important inorganic component of the shell matrix is iron , which , along with eggshell precursors is stored in the vitelline cells [11] . The shell is largely resistant to degradation by immune effectors in the host or microbes in the environment . Schistosome eggs are fully embryonated as they escape the host , and as a consequence , miracidia can hatch immediately upon appropriate stimuli in the external environment . The primary cue for miracidial emergence is the osmotic changes that occur as the egg enters freshwater [12]–[14] . Although the physiological and molecular cascades leading to hatching are poorly understood , it has been , hypothesised that they involve , in addition to osmotic changes , calcium fluxes , and the activity of leucine aminopeptidases [3] , [13] , [14] . In view of the poor knowledge of the biology of egg hatching in schistosomes , we undertook a detailed dynamic and correlative study of hatching biology of Schistosoma japonicum . This species , we discovered , has an exquisite and dynamic hatching behaviour . The peculiar nature of its hatching biology enabled us to dissect and understand more clearly the internal cellular changes that occur during the process . In this study , use was made of the electron microscopy preparative technique of high pressure freezing ( HPF ) . Due to the impervious nature of the trematode eggshell , fixation of its enclosed embryonic structure is notoriously difficult [3] , [5] and the resultant ultrastructure of eggs and their content has been poor . An earlier investigation that used the cryo-preparative techniques of slam-freezing in liquid nitrogen to study egg ultrastructure of S . mansoni [4] , while informative , suffered from artifactual ice crystal damage in the tissues . Similarly , eggs studied after conventional fixation methods of glutaraldehyde fixation were applied [3] , [5] , suffered from artifact due to slow ingress of fixative into the eggs . With HPF procedures , samples are snap-frozen in liquid nitrogen at high hydrostatic pressure ( approximately 2000 bar ) [15] . Under high pressure , water in biological samples is inhibited from nucleating into ice crystals and , thus , is frozen in a vitreous state . Once frozen , samples can be fixed for electron microscopy at low temperature ( <−80 C ) using a cryo-substitution method which allows fixative to infiltrate samples in the presence of a resin-miscible solvent [15] . Thus , the procedure simultaneously preserves the ultrastructure of the sample and removes water to prevent ice nucleation during thawing . A major benefit of HPF is that specimens can be cryo-immobilized for observations during dynamic activities , methods not possible with conventional preparation methods . Here we describe the dynamic events of hatching of S . japonicum egg using HPF to immobilize miracidia for ultrastructural interpretation .
Schistosoma japonicum eggs of Chinese mainland ( Anhui Province ) origin were recovered from livers of infected mice after digestion in collagenase B following described methods [16] . Eggs were stored for 2 days in Petri dishes in sterile PBS at 4°C until use . The use of mice in this study was approved under Project P288 by the Animal Ethics Committee of the Queensland Institute of Medical Research . To study hatching behaviour of larvae , eggs were transferred to aged distilled water . Real-time video recordings were made at 100× magnification using a Panasonic Color CCTV Camera ( Model WV-CP610/G ) , connected to the video output of a Zeiss inverted microscope . Eggs and miracidia were fixed for transmission electron microscopy at three stages of the hatching process , as follows: Eggs from stages A and B were transferred to a solution of 20% ( w/v ) bovine serum albumen in PBS on a proprietary copper membrane and rapidly frozen in a Leica EM PACT2 High Pressure Freezer ( Leica , Wetzlar , Germany ) . Subsequently , the membranes and samples were transferred in cryo-tubes under liquid nitrogen to a Leica EM AFS freeze substitution apparatus for fixation and dehydration in 2% ( w/v ) osmium tetroxide and 0 . 5% uranyl acetate ( w/v ) in 100% anhydrous acetone . The tissues were cryo-substituted for 3 days , according to the following protocol . The temperature of the substitution chamber was increased from −160°C to −85°C over 2 h , and maintained at −85C for 48 h , after which the samples were brought to room temperature . The osmium-uranyl acetate-acetone solution was then replaced with anhydrous acetone at room temperature . After further changes of acetone , the samples were infiltrated with Epon resin ( ProSciTech , Townsville Australia ) . Final infiltration of resin was facilitated in a Pelco 34700 Biowave Microwave Oven ( Ted Pella Inc . , Redding , USA ) . Eggs from stage C were prepared by conventional EM processing as the miracidia had escaped the eggshell and were thus more easily fixed . For this , liver-derived eggs were incubated at room temperature in aged water in 24-well plates . As soon as a larva and its enveloping structures were observed to break from the shell , an aliquot of 3% glutaraldehyde ( v/v ) in 0 . 1 M phosphate buffer , pH7 . 4 was introduced into the well , thus immediately fixing and immobilizing the miracidia . The fixed miracidia were postfixed in 1% osmium tetroxide , then dehydrated in acetone at room temperature and embedded in Epon resin . Ultrathin sections were cut at 60 nm , mounted onto uncoated or formvar/carbon coated copper grids , and stained with uranyl acetate and Reynold's lead citrate . The sections were examined in a JEM1011 transmission electron microscope operated at 80 KV and photographed using a digital camera . EM observations were made from multiple specimens from each sample . Hatching experiments in water and praziquantel were repeated three times . Ultrathin sections of high pressure frozen and freeze substituted pre-activated eggs ( stage A ) , were subjected to lectin cytochemistry using biotinylated Concanavalin A , Wheat Germ Agglutinin , Peanut Agglutinin and Ulex Europaeus Agglutinin ( Vector Labs , Burlingame , USA ) . The carbohydrate specificity of the lectins is published [18] , [19] and presented here also in Table 1 . All lectins were diluted to 10 µg/ml in HEPES buffer , using 0 . 5% ( w/v ) gelatin as the blocking agent . Subsequently sections were labelled with rabbit anti-biotin antiserum ( Bethyl Laboratories , Montgomery , USA ) diluted 1∶200 in gelatin/PBS followed by protein-A conjugated to 10 nm gold particles ( Utrecht University , Netherlands ) in the same buffer . Sections were stained with uranyl acetate and lead citrate , and viewed in a JEM1011 transmission electron microscope as described above .
The hatching behaviour is shown in Videos S1 , S2 , S3 and S4 and Fig . 1 . Another version of the hatching biology in a movie ( made by MKJ ) has been deposited in the TDR website of the World Health Organization with the title The Great Escape ( currently file is at: http://www . who . int/tdr ) . After exposure to water , most miracidia do not display activity within the egg shells for at least twenty minutes , although some hatched spontaneously . Larvae continued to activate and hatch over a 24 hour period . Evidence of activation included arrhythmic twitching of the miracidium and rapid ciliary beating ( Videos S1 and S2 ) . These activities continued for a variable and sometimes prolonged period of time . The contents of the matrix within the shell became progressively clarified and large vacuoles in the lacuna surrounding the miracidium become more apparent . In most miracidia ( Video S3 , and “The Great Escape” ) , shell rupture occurred only after the miracidium swung its body from a longitudinal to a lateral orientation within the shell , so that the long axis was perpendicular to the shortest axis of the egg . This change in orientation did not occur for the miracidium filmed in Video S1 , but the egg in this specimen appeared smaller than surrounding eggs , suggesting that it is traction against the shell walls that was important rather than orientation , for this and other miracidia appear to use muscular activity to aid in shell rupture . The shell always ruptured along the long axis of the egg ( see fracture plane in empty shells in the background of Videos S1 , S2 , S3 and S4 ) . Following shell rupture , the larva emerged , enclosed within a sac formed by an extra embryonic “membrane” , the outer envelope ( OE; = Reynold's layer [4] ) . The OE underwent rapid expansion until it filled a volume at least twice that of the shell ( Videos S1 , S3 ) . Vacuoles within the OE also expanded at similar rates and continue to expand after the larva has escaped the OE ( Video S1 ) . The miracidium swam rapidly and erratically within the sac , stopping on occasions to push its anterior organ , the terebratorium , against the limiting OE . The medium in which the miracidium swam became increasingly viscous as the larva circled within it . Eventually , the miracidium stopped and forced its way through the OE , thereby liberating itself and the viscous contents of the sac . As assessed by fixation quality of internal structure of the egg , particularly mitochondria and endoplasmic reticulum , membranes , and the absence of ice crystal damage , the use of HPF produced superior ultrastructure of encased miracidia and investing envelopes . Different authors have used a variety of names to describe the ultrastructure of the schistosome egg [3]–[5] . The result is a confusing nomenclature . The system proposed by Swiderski [5] , which uses names commonly used for egg structures in the Platyhelminthes [20] , is adopted here , although alternative nomenclature is also given . Rapid freezing of samples in HPF ensures the simultaneous immobilization of all components of tissues , and is thus optimal for providing a snapshot of dynamic activity in cells and tissues [23] . We activated eggs in distilled water to induce hatching and immobilised eggs after 1 h to observe the structural changes that occurred in the membranes during hatching . Because individual miracidia respond to osmotic changes at different time-points and , thus hatch at different times throughout a 24 h time period , it is difficult to prepare multiple specimens by HPF . For this reason , eggs of S . japonicum were exposed to praziquantel to induce rapid and near synchronous hatching of all mature eggs [17] . There are some deleterious effects of praziquantel administration on miracidia . The major effect is on the musculature of the miracidium , which contracts markedly , so that the larvae appear pinched in their midline . The miracidia can still swim , however , through active motility of their cilia . Miracidia still move their bodies to lie laterally across the egg prior to shell rupture . A second effect is that some components of the lacuna do not separate completely from the cilia , so that some hatched larvae may drag sub-shell components through the aquatic milieu . Four changes were obvious in eggs that had been induced to hatch by praziquantel or water . Firstly , the basal lamina of the shell separated from the OE so that a distinct space was observed between the two envelopes ( Fig 5A–B ) . In shells ruptured by hatching , the shell material appeared completely separated from the OE ( Fig . 5C ) . Secondly , the lacuna surrounding the miracidium lost its flocculo-granular matrix and the matrix became clearer . Thirdly , lipoid bodies became distributed throughout the lacuna and were not aggregated anteriorly . Fourthly , the rosettes within the membrane sacs became less distinct and showed evidence of degradation ( Figs 5D , E ) . Immediately after eggshell rupture , the OE and vacuoles expanded rapidly ( Videos S1 , S3 ) . When miracidia and attendant envelopes were fixed during this phase of expansion , the only ultrastructurally identifiable features were the OE and the miracidium ( Figs 6A–D ) . The OE remains as an electron-opaque fibrous layer , although it was much thinner than in it was in the pre-activatedn egg ( Figs 6A–C ) . The branched fibrils of the OE separated from each other at the external water interface ( Fig . 5C ) . Other contents of the egg , including IE , were degraded and replaced by a coarsely granular material and numerous membrane profiles , some of which were reminiscent of endoplasmic reticulum . The vacuoles were represented by membranes surrounding a granular substance ( Fig . 5A–C ) . Lectin cytochemistry of the eggs ( Fig . 7; Table 1 ) revealed a variable distribution of carbohydrate moieties throughout the egg shell constituents . No lectin bound to egg shell , although there was weak lectin binding for UAE , WGA and PNA in the pores . The OE stained weakly with ConA . Con A labelled a range of structures , but particularly , the rosettes , matrix of the lacuna , regions of the miracidial tegument and vesicles of the penetration glands , lipoid bodies and lysosomes of the IE ( Fig . 7; Table 1 ) . PNA reactivity was observed only at the periphery of secretory bodies of the penetration glands and at the periphery of lipoid bodies . WGA reactivity was detected only within the lipoid bodies .
This study , the first detailed study of the ultrastructure of S . japonicum eggs , shows that the use of HPF as a means of immobilizing the eggs and egg contents provided a superior method of specimen preparation compared with other methods used previously for studies of S . mansoni eggs [3]–[5] . The structure of the eggshell and its sub-shell compartments were preserved with high quality and without freezing artifacts . Despite the use of osmium tetroxide as fixative , some cytochemical labelling using biotinylated lectins in conjunction with colloidal gold probes was possible , as has been demonstrated for studies on insects and nematodes [24] . The ability of the technique to immobilize eggs during hatching , allowing morphological and cytochemical observations , helps overcome problems brought about by the impervious shell [25] . The method provides a means of adding valuable data on the biological roles of drugs , such as praziquantel . Based on the vital , structural and cytochemical studies described here , the following conclusions can be made about the components of the S . japonicum egg . The shell is a dense homogenous layer and is completely devoid of carbohydrate residues . Numerous microspines are present at the surface and , along with the terminal egg spine , may provide traction for the egg to adhere to vascular endothelium in the mammalian host . For S . mansoni , the interaction of eggs and endothelium is intimate , as evidenced by the rapidity with which the host tissues overgrow freshly deposited eggs [26] . The shell itself is a dense homogenous layer , interrupted in places by cribriform pores . In all mature eggs observed here , the shell was of uniform density and thickness . Interestingly , our videos of hatching S . japonicum miracidia showed that the shell always ruptures along the long axis of the shell , indicating the presence of a pre-formed weakness in the shell along this axis . Such a line of weakness could arise from a different chemistry in certain regions of the shell , from the occurrence of a thinner band of eggshell material at the line of weakness , or by accumulation of pores in that area . None of these , however , were apparent in this study . The outer envelope of the schistosome egg consists of a filamentous material [4] , composed of numerous interlocked fibrillar proteins . The layer is closely applied to the eggshell , and likely bonds with the latter structure . The junction between the two layers was clearly broken in water- and praziquantel-induced hatching . After shell rupture , the OE thinned considerably as the mass it enclosed expanded rapidly . The filaments of the OE also appeared to unravel in the presence of water , suggesting that cross-links between filaments are water-soluble or readily hydrated . Secreted antigens produced by the egg must traverse the OE to escape the egg . Many secreted antigens are thought to arise in the IE [3] and there is substantial evidence that secreted antigens are generated by mature schistosome eggs [3] , [7] , as well as immature eggs , which lack the fibrous OE . Apart from the observation of small particles in regions of the OE subjacent to pores , there is little morphological or cytochemical evidence of antigens traversing the OE . The carbohydrate fucose , along with mucin-rich compounds , are abundant components of glycoproteins secreted by S . mansoni eggs [6] , but cytochemical analysis here did not reveal abundant carbohydrate components in the OE . Nevertheless , the mature OE is likely to be a water soluble matrix , as evidenced by the seemingly rapid transit of water through the layer and its behaviour in freshwater . Ashton and colleagues have proposed that the inner envelope is the primary source of secreted antigens in S . mansoni eggs [3] . The cellular layer is highly synthetic , and observations here indicate that among its possibly many functions can be listed the remodelling of the OE , secretion of proteins , and generation of substantial carbohydrate components present in the rosettes and lipoid bodies . Swiderski [5] described acid phosphatase reactivity in the IE of S . mansoni eggs just prior to hatching , and suggested that autolysis of this layer provided substantial nutritional benefit to the miracidium . In S . japonicum eggs , the IE degrades only after activation , suggesting that the IE remains extant prior to activation to maintain extraembryonic components of the egg . Structures in the lacuna of mature S . japonicum eggs bind lectins indicating the presence of carbohydrate moieties . Rosettes stain with ConA , which indiscriminately binds glucose and mannose . Similar structures in the related digeneans Fasciola hepatica and Echinostoma caproni also contain carbohydrate and are involved expansion of the viscous cushion , a structure that leads to rupture of the opercular boundary in their eggs [4] , [5] , [19] . The so-called lipoid bodies are also rich in carbohydrates , particularly glucose , mannose , galactosamine and N-acetyl-glucosamine , as is the general matrix of the lacuna . The process of egg hatching appears similar in S . japonicum and S . mansoni [13] , [14] . One of the clear differences between the two species , however , is that for S . japonicum , the OE emerges to form a sac around the escaping miracidium . Based on the accumulated evidence from a range of studies [13] , [14] , [17] , [19] , [27] , the following model of egg hatching can be postulated for S . japonicum . On contact with freshwater , osmosis induces the inflow of water into the egg , probably through the pores . Kusel [14] noted that S . mansoni eggs swell prior to hatching , indicating that the water inflow is abundant . Osmotic pressure appears to be a major factor in hatching of schistosome eggs [13] . In the related schistosome , Trichobilharzia regenti , a parasite of birds , it appears that other factors stimulate hatching [28] . In that species , larvae hatch within the nasal tissues of their avian hosts , stimulated by light or by slight osmotic change with inflow of water into the nasal cavities . The influence of other environmental parameters on hatching of eggs of Schistosoma species , especially light , is uncertain , as miracidia hatch in the dark [13] . Praziquantel , as noted here and by others [17] , [29] induces rapid hatching of schistosomes eggs in the absence of an osmotic gradient . There is mounting evidence that praziquantel modulates calcium transport in schistosomes [30] , [31] . Praziquantel causes hatching in PBS , implying that its effect is independent of osmotic pressure and likely involves the flux of calcium into the tissues of the eggs . Whether this modulation of calcium levels occurs in the inner envelope or miracidium to induce hatching remains to be determined . One of the immediate outcomes of egg activation in S . mansoni is the rapid release of the hydrolytic enzyme leucine aminopeptidase ( LAP ) , which has been recorded in hatching fluids upon release of miracidia from eggs [13] , [32] . Administration of praziquantel to S . mansoni eggs also induces an immediate release of LAP [29] . Blocking of LAP activity in eggs with bestatin , a general inhibitor of aminopeptidases , effectively blocks hatching activity [13] , [32] . LAPs are found in the gastrodermis and tegument of adult schistosomes [33] , but the site of expression in the eggs is unknown . One likely source of the enzyme is the IE itself , which in S . japonicum possesses lysosomes , and in S . mansoni contains compartments which stain for acid phosphatase , a marker of acidic compartments where LAPs are likely to function [5] . LAPs are terminal enzymes in catalytic cascades , liberating amino acids , particularly N-terminal leucines from peptides [33] , [34] . The natural substrate of LAP from eggs is uncertain , but potential roles for the enzyme include scission of the OE-shell boundary , autolysis of the IE itself , or perhaps to assist in degradation of proteins that occur in the carbohydrate-laden lacunae . Data on egg hatching in the parasitic Platyhelminthes is scant , but for those species that have been studied it appears that many use a form of extra-embryonic viscous cushion or vacuole that expands to induce pressure increases within the egg ( reviewed in [13] , [19] , [35] ) . It has been calculated that a pressure of 35 megaPascals ( 5000 psi ) is required to rupture schistosome eggshells [13] . Schmidt [19] used lectin cytochemistry in light microscopy studies to demonstrate that glycogen-like carbohydrate components were abundant in vesicles in the egg lacunae of F . hepatica and E . caproni . Schmidt argued that while polymerised forms of polysaccharides are generally water insoluble , their depolymerization will lead to greater water solubility and water binding potential . The lacuna of S . japonicum eggs appears to be enriched for carbohydrate polymers . Moreover , the potential for rapid expansion of the OE sac once liberated from the ruptured shell suggests a rapid osmosis-driven influx of water , mediated perhaps by the unravelling of previously hydrophobic molecules . The absence of rosettes and lipoid bodies in hatched eggs ( Fig . 6 ) suggest that these complex polysaccharides do , indeed , depolymerise . Based on the new information described here , we postulate that the eggs of schistosomes , and particularly , S . japonicum , proceeds as follows . Eggs exist within the mammalian host under isosmotic conditions . When transferred to freshwater , an inflow of water signals calcium fluxes in the external membranes of the IE of the miracidium , leading to activation of potentiating enzymes that initiate scission of the OE-eggshell boundary , autolysis of the IE and partial depolymerization of polysaccharides in the lacuna . The potential involvement of LAP in this process remains undetermined . With the dissolution of structures , and aided by mechanical activity of the miracidial cilia , the intra-ovular environment becomes free allowing for partial motility of the miracidium so that it can change its orientation in the egg . The increased pressure afforded by the expanding polysaccharides , together with muscular activity of the larva itself forces rupture of the shell . The freed OE allows the sudden influx of water , a flow previously limited by the physical constraints of the pores , and the OE sac expands rapidly . With more space to move , the miracidium swims rapidly , but stops intermittently to force a break in the OE , through which it finally escapes . It remains to explain the selective advantage of the swelling OE to ensure hatching and overall continuation of the life cycle of S . japonicum . The answer possibly and most simply is that the OE serves to push out a clear space in the faecal mass containing the egg to give the larva an opportunity to escape into surrounding water .
|
Adult schistosomes live within portal veins of their human hosts . Their offspring , laid as eggs within the venous system , escape by traversing the tissues between the blood vessel and the gut or bladder lumen . Eggs voided into the external environment hatch spontaneously on contact with freshwater , and the hatched larva escapes in search of a snail , which acts as intermediate host of the parasite . In this study , we used correlative microscopy techniques to examine hatching of the larvae of Schistosoma japonicum . This species has an exquisite hatching behaviour , which allows us to trace the cellular changes in the egg that lead to hatching . By using a correlative microscopy approach , incorporating video microscopy , electron microscopy of eggs prepared by high pressure freezing and lectin immunocytochemistry , we were able to describe the pre-hatching state of the eggs , and trace changes that occur during hatching . The insights gained from these direct biological studies will be of value in understanding the host–parasite interplay of schistosome eggs in their hosts .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"infectious",
"diseases/helminth",
"infections",
"developmental",
"biology/embryology",
"cell",
"biology/developmental",
"molecular",
"mechanisms"
] |
2008
|
Correlative and Dynamic Imaging of the Hatching Biology of Schistosoma japonicum from Eggs Prepared by High Pressure Freezing
|
The processing characteristics of neurons in the central auditory system are directly shaped by and reflect the statistics of natural acoustic environments , but the principles that govern the relationship between natural sound ensembles and observed responses in neurophysiological studies remain unclear . In particular , accumulating evidence suggests the presence of a code based on sustained neural firing rates , where central auditory neurons exhibit strong , persistent responses to their preferred stimuli . Such a strategy can indicate the presence of ongoing sounds , is involved in parsing complex auditory scenes , and may play a role in matching neural dynamics to varying time scales in acoustic signals . In this paper , we describe a computational framework for exploring the influence of a code based on sustained firing rates on the shape of the spectro-temporal receptive field ( STRF ) , a linear kernel that maps a spectro-temporal acoustic stimulus to the instantaneous firing rate of a central auditory neuron . We demonstrate the emergence of richly structured STRFs that capture the structure of natural sounds over a wide range of timescales , and show how the emergent ensembles resemble those commonly reported in physiological studies . Furthermore , we compare ensembles that optimize a sustained firing code with one that optimizes a sparse code , another widely considered coding strategy , and suggest how the resulting population responses are not mutually exclusive . Finally , we demonstrate how the emergent ensembles contour the high-energy spectro-temporal modulations of natural sounds , forming a discriminative representation that captures the full range of modulation statistics that characterize natural sound ensembles . These findings have direct implications for our understanding of how sensory systems encode the informative components of natural stimuli and potentially facilitate multi-sensory integration .
It is widely believed that sensory representations are optimized to process the stimuli to which they are exposed in natural environments [1] . Of particular interest is understanding the computational principles that underlie the generation of observed neural firing patterns . A popular hypothesis explored in recent years assumes that neural populations optimize a sparse code . This means that at any given time , only a small subset of a neural population fires to encode a given stimulus [2] . Such a representation is attractive for reasons of coding efficiency ( see , e . g . , [3] ) and conservation of physiological resources [4] . The sparse coding hypothesis has enjoyed particular success in studies of vision ( e . g . , [5] , [6] ) , and has also been supported more recently by both neurophysiological [7] , [8] and computational studies [9]–[11] of the auditory system . However , it has also been observed that some central auditory neurons , when driven by their preferred stimuli , exhibit sustained firing rates . Measuring from auditory thalamus and primary auditory cortex , Wang et al . observed that sustained responses were not simply phase-locked to the fast dynamics of the stimulus , suggesting that this rate-based code represented a meaningful , non-isomorphic transformation of the stimulus [12] , [13] . Indeed , such a code is particularly important for audition since it directly addresses the issue of how to indicate the continued presence of a sound in a complex acoustic environment . Results from Petkov et al . have also illustrated how sustained responses play a role in auditory scene analysis , forming part of the neural basis for the perceptual restoration of foreground sounds against a cluttered background [14] . Moreover , Wang has argued that a rate-based representation is critical for matching fast temporal modulations present in natural sounds to slower rates found in higher cortical areas [15] . Slower dynamics in acoustic signals are believed to be the main carrier of information in speech and music [16]; are commensurate with temporal dynamics of stream formation and auditory grouping [17]; and may play an important role in multi-modal sensory integration [15] . Related computational studies in vision have suggested how this principle may underlie the shapes of simple and complex cell receptive fields in primary visual cortex [18] , [19] . Importantly , a sustained firing rate , i . e . , one that is persistent and therefore slowly changing over time , is related to slow feature analysis , a well-known method for extracting invariances from sensory signals [20] ( see Discussion ) . To the best of our knowledge , however , there are no computational studies that explicitly consider the implications of a sustained firing-based code in central auditory areas . At first glance , the two coding schemes are seemingly at odds: on the one hand a sparse code seeks to minimize the activity of a neural population whereas a sustained firing-based code requires that neural responses persist over time but still form an efficient representation of the stimulus . However , it appears that central auditory responses can strike a balance between the two strategies , with a large , transient population response at the onset of a sound , and a sparse subset of preferentially driven neurons exhibiting a strong , sustained response throughout the sound's duration [15] , [21] . This picture suggests a mechanism for detecting and tracking target sounds in noisy acoustic environments and for generating a persistent signal that facilitates a stable perceptual representation . From a computational perspective , a better understanding of these mechanisms can inform models of auditory scene analysis as well as signal processing schemes for hearing prosthetics and automated sound processing systems . A general computational approach for exploring the effects of particular coding strategies in sensory systems is based on optimizing a statistical objective criterion that quantifies the principle governing the transformation between stimulus and internal representation . Upon convergence , one then compares the emergent representation to known properties of the sensory system being studied [1] . Here , we apply this framework to explore how optimizing a sustained firing criterion influences the shapes of model auditory spectro-temporal receptive fields ( STRFs ) when processing natural sounds , and we compare the emergent ensembles to those obtained by optimizing a sparse coding objective . STRFs describe the linear mapping between a spectro-temporal stimulus and an instantaneous firing rate [22] , and have proven useful not only for describing basic processing aspects of auditory neurons [23] , [24] , but also for shedding light on the nature of task-driven plasticity [25] . Figure 1 illustrates how a spectro-temporal stimulus is mapped to a set of instantaneous neural firing rates , whose ensemble response according to a desired coding strategy directly shapes the mapping . In this paper , we show how this framework allows us to not only explore how the timescales of natural sounds are captured by and reflected in an emergent sensory representation , but reveal key similarities between choice of a sustained versus sparse code . Moreover , we demonstrate how a sustained firing-based code suggests a mechanism for an emergent discriminative representation for ensembles of natural stimuli .
We optimized both the sustained objective and sparsity objective using an ensemble of natural stimuli comprising speech , animal vocalizations , and ambient outdoor sounds . Each ensemble of filters was initialized at random using zero-mean , unit variance Gaussian noise , and each STRF covered from 0–250 ms in time and 62 . 5–4000 Hz along the tonotopic axis . For the sustained objective , we considered a wide range of correlation intervals from very brief ( ) to very long ( ) . Examples of emergent STRFs for are shown in Figure 2A . For the spectro-temporal patches shown , red and blue colors indicate that the presence of energy in a particular spectro-temporal region yields excitatory and inhibitory responses , respectively . We observe a variety of STRFs that are highly localized , sensitive to narrowband spectral and temporal events , oriented , and some that are seemingly noise-like and not convergent to any particularly interesting shape . Importantly , such observations about these basic STRF classes align with those made in a number of previous physiological studies ( see , e . g . , [23] , [24] , [27] ) . Moreover , coverage of the STRFs appears to span the full time-frequency space . These results suggest that the sustained firing objective may underlie part of the coding strategy used by central auditory neurons . Shown in Figure 2B are examples of emergent STRFs obtained by optimizing the sparsity objective . Indeed , this particular objective yields STRFs that are highly localized and sparsely distributed , with sensitivity to bandlimited spectral and temporal events . While both objective criteria yield noisy STRFs , it is clear that the sparse ensemble is much more noisy , with a less extensive coverage of the basic sound classes as observed with the sustained ensemble . Since the information-bearing components of natural sounds vary concurrently across multiple timescales , it was expected that the structure of STRFs learned under the sustained objective would vary with the correlation interval . Indeed , inspection of the sustained ensembles for a range of suggested the presence of a number of latent classes whose membership varied smoothly from short to long correlation intervals . To quantify variations in population diversity over ecologically relevant timescales , we performed unsupervised clustering of the emergent STRFs and studied how class membership changed with objective function and correlation interval . We pooled STRFs from the sparse ensemble and from the sustained ensembles for 10 , 25 , 50 , 125 , 250 , 500 , 1000 , and 2000 ms , yielding a total of 3600 STRFs . We then applied normalized spectral clustering to discover latent classes among the pooled STRFs . In general , spectral clustering algorithms require an affinity matrix that specifies pairwise similarities between the objects being clustered . Viewing this affinity matrix as an undirected graph , spectral clustering finds a partition of the graph into groups whose elements have common similarity with one another . A natural measure of similarity between STRFs can be derived from the two-dimensional cross-correlation between pairs of spectro-temporal patches . Such a measure is similar to that considered by Woolley et al . [28] and is desirable since it does not depend on subjective choice of spectro-temporal features to use for clustering . In this work , we defined the measure of similarity between pairs of STRFs as the absolute value of the maximum value of the two-dimensional cross-correlation matrix; we used absolute value since we wished to group similar STRFs regardless of whether they were excitatory or inhibitory . Furthermore , as the STRFs tended to be distributed with a variety of phases in the input space , we considered cross-correlations for arbitrary time-frequency shifts ( see Methods for details ) . Results obtained using normalized spectral clustering of the emergent ensembles into nine classes are shown in Figure 3 . In the center panel of the figure , a stacked bar chart illustrates the the percentage of STRFs at a particular assigned to one of nine classes . Different segment colors correspond to each of the nine classes , and segment width is proportional to the number of STRFs assigned to that class . Surrounding the bar chart are examples from six classes that best illustrate how diversity varies with , namely noisy , localized , spectral , complex , temporal , and directional classes . These labels are qualitative descriptors of each class and not quantitative assessments of the time-frequency characteristics of each category . Inspection of the cluster groupings reveal rich structural variations over a wide range of correlation intervals . In particular , the STRFs labeled according to the noisy class are found to dominate the sparse ensemble , with a large presence in the sustained ensemble for . Membership in this class drops for between 10 and 125 ms , and begins to increase at 125 ms . We also observe that short correlation intervals ( 10 , 25 , and 50 ms ) have a large concentration of localized STRFs , with membership dropping with increasing . While the temporal class holds relatively steady across the sustained ensembles , we find that membership in the directional , complex , and spectral classes varied smoothly across . In general , we find that ensemble diversity is maximized for ( max . entropy of 3 . 08 bits ) , but the overall trends suggest rich ensemble structure between 10 and 250 ms , which is notably in the range of the timescales of natural sounds [29] , . This is further supported by the increasing presence of noisy STRFs for large correlation intervals ( 1000 and 2000 ms ) . In addition to studying structural variations in the shapes of the emergent STRFs , it is also of interest to examine the structure of the STRF outputs in response to natural sounds . In particular , we sought to address the extent to which enforcing sustained responses does indeed yield responses that persist over time . We defined the neuron to be significantly “active” when its firing rate exceeded 1 standard deviation over time . While this is not meant to be a precise measure of a neuron's activation ( since , for instance , the firing rate is not used to modulate a Poisson spike generation process ) , such a measure nevertheless quantifies and characterizes a strong versus weak ensemble response to natural stimuli . Shown in Figure 4A are the distribution of activation times for individual neurons for ensembles of 10 and 125 ms in response to a held-out set of natural stimuli . The neurons are shown sorted according to decreasing median activation time , and the interquartile ranges of activation time are indicated by the shaded regions . We observed that the most diversity in median activation times across ensembles occurred in approximately the top 10% of the most persistent neurons . To summarize these observations , we considered the distribution of median activation times of the top 10% of neurons with most persistent responses ( i . e . , the top 40 neurons ) ; these distributions are illustrated as boxplots in Figure 4B . As noted previously with the clustering results , shorter values favor mostly localized and noisy STRFs and consequently it was expected that activations would be brief . Interestingly , however , we observe that with increasing , median activations peak between 50 and 500 ms and fall off for large despite the STRFs being optimized to promote sustained responses over long intervals . This overall trend aligns with the previous clustering results that demonstrate how population diversity is maximized over intervals corresponding to timescales that predominate natural stimuli . The STRFs corresponding to the top 10% most persistent responses for are shown in Supplementary Figure 1 , and we find that they generally have a spectral tuning , but are fairly narrowband and localized . Additionally , we considered the responses of the top 40 most persistent responses obtained using the sparsity objective function; the distribution of median activations is in the first column of Figure 4B . We find that the sparse ensemble yields responses most similar to those for short . How do the emergent STRFs learned under the sustained firing objective compare to those observed in physiological studies ? Broadly speaking , we find that the emergent STRFs share many of the trends with biological receptive fields typically observed in animal models . We explored this issue by comparing our model ensembles with a set of 1586 STRFs recorded from awake , non-behaving ferret primary auditory cortex using TORC [31] and speech stimuli [27] , [32] ( see Methods for more details ) . Where applicable , we also compared our results with reported results from anesthetized ferrets by Depireux et al . [23] and cats by Miller et al . [24] in the literature . Illustrative examples of the types of STRFs found in the neural data are shown in Figure 5 . In particular , we find neural STRFs that are qualitatively similar those found in the localized , complex , noisy , and directional clusters shown earlier in Figure 3 . Because the temporal and spectral sampling rates used in our model are higher than those used in the physiological data , we did not find good matches with the temporal and spectral classes . To visualize the overlap between the spectro-temporal modulation coverage of the neural and model STRFs , we used the ensemble modulation transfer function ( eMTF ) . The eMTF is derived by averaging the magnitude of the 2D Fourier Transform of each neuron in a given ensemble , and jointly characterizes modulations in time ( rate , in Hz ) and in frequency ( scale , in cyc/oct ) . We first applied normalized spectral clustering to the neural STRFs to obtain nine clusters . Next , we computed the eMTF for each cluster , extracted isoline contours at the 65% level , and overlaid these curves on the eMTF of the model STRFs for . These results are shown in Figure 6 and illustrate the overlap between the model and neural data , particularly at the “edges” of the neural STRF modulations . While the overlap is not complete , it is clear that the modulation spectra of each ensemble are not disjoint . Moreover , the model eMTF suggests a general ensemble sensitivity to relatively fast modulations; this point is explored further in a later section ( “Emergent STRFs capture spectro-temporal modulation statistics of stimulus” ) . To better characterize the relationship between the neural and model data , we employed a statistical comparison of the distribution of the two datasets . If the models truly generated STRFs similar to those in physiological studies , then one might expect a nearest-neighbor ( NN ) similarity distribution akin to one derived from the neural ensemble we considered . We computed the symmetric KL-divergence between each of the model and within-physiology NN similarity distributions ( shown in Supplemental Figure 2 ) . We found that the sustained-response ( presented here ) and sustained-shape ( presented later in this paper ) distributions had KL divergences of 0 . 80 and 0 . 85 , respectively , whereas the sparse distribution had a KL distance of 1 . 05 . KL typically measures the expected number of bits required to code samples from one distribution using codes from the other . While these numbers are difficult to assess in absolute terms , they give a sense of how the different model optimizations and constraints compare to each other . These numbers reveal that the sustained ensembles are similarly comparable to the physiology , whereas the sparse ensemble has a somewhat worse match . Of course , caution must be taken with these numbers because the set of neural STRFs we analyzed represent only a subset of mappings that likely exist in central auditory areas . Next , we measured a variety of parameters from the neural and model STRFs ( for ) that more fully characterized the extent of spectro-temporal coverage and modulation sensitivity of the ensembles ( see Methods ) , the results of which are summarized in Figure 7 . Based on the distribution of directionality indices , shown in panel ( A ) , we observe that the model STRFs are largely symmetric , with the majority of neurons having no preference for upward or downward moving input stimuli ( mean0 ) . As indicated by the tails of this distribution , however , a subset of neurons have a strong directional preference . This agrees with the neural STRFs , and similar observations have been made in MGB and primary auditory cortex of cats by Miller et al . , as well as in measurements by Depireux et al . from primary auditory cortex of ferrets . Furthermore , panel ( B ) illustrates that a large number of model STRFs are fairly separable , with a peak in the separability index ( SPI ) distribution around 0 . 10 and an average value of 0 . 26 . This trend aligns with values reported in the literature by Depireux et al . in measurements from ferret auditory cortex ( mean of approx . 0 . 25 ) . However , it is worth noting that this low level of separability is not uniformly reported across physiological studies of receptive field of mammalian auditory cortex . For instance , the physiological data analyzed in the current study ( examples of which are shown in Figure 5 ) do yield a higher average SPI ( mean = 0 . 37 ) . The temporal modulation statistics of the model STRFs , as quantified by best rate ( BR ) , also align generally with results reported from mammalian thalamus and cortex . In panel ( C ) we observe a broad , bandpass distribution of best rates , with an average of 23 . 9 Hz . Reported physiological results from Miller et al . show similarly broad ranges of temporal tuning with preferences around 16 Hz and 30 Hz range for cortex and thalamus , respectively . The neural STRFs we analyzed show a somewhat slower tuning , with an average BR of 9 . 5 Hz . Furthermore , in panel ( D ) , we computed the normalized average rate profile from the model STRFs . We observe a peak at 7 . 8 Hz , with an upper 6-dB cutoff of 34 . 4 Hz . Here we find a close overlap with the rate profile computed from the neural STRFs as well as with average profile results as reported by Miller et al . ( peak at 12 . 8 Hz; upper 6-dB cutoff at 37 . 4 Hz ) . The spectral modulation statistics of the model STRFs , as quantified by best scale , are generally faster than those reported from studies of thalamic and cortical nuclei . The distribution of best scales shown in panel ( E ) is bandpass with a wide range of slow to fast spectral coverage , with an average tuning of 1 . 40 cyc/oct . The neural STRFs , in contrast , are tuned to much slower scales ( mean = 0 . 47 cyc/oct ) . Similarly , results from Miller et al . in MGB indicate a generally slower tuning ( 0 . 58 cyc/oct ) , whereas measurements from cortical neurons , while having a similarly wide range of tunings as with the model , indicate a slower average value of 0 . 46 cyc/oct and an upper cutoff of approx . 2 cyc/oct . Finally , the ensemble average scale profile , shown in panel ( F ) , is bandpass and exhibits a peak at 0 . 7 cyc/oct with an upper 6-dB cutoff of 2 . 9 cyc/oct . The neural STRFs , however , are much slower with peak at 0 . 2 cyc/oct and an upper cutoff of 1 . 9 cyc/oct . This is similar to observations from MGB by Miller et al . , where they reported that the ensemble average scale profile is generally low-pass , with average scale profile peaks and upper 6-dB cutoffs at 0 cyc/oct and 1 . 3 cyc/oct , respectively , with similar observations in cortex . In summary , while we cannot map the emergent STRFs to any exact synapse , they nevertheless reflect the general processing characteristics of various stations along in the central auditory pathway . There is good alignment with the neural STRFs and reported results in mammalian MGB and primary auditory cortex with respect to directional sensitivity and spectro-temporal separability . The temporal modulation statistics of the emergent sustained STRFs appear to be most similar to those measured from thalamus and cortex . Furthermore , the model STRFs are generally faster with regard to spectral modulations than those measured from thalamus and cortex . To explore the relationship between STRFs optimized to promote sustained responses and those that explicitly maximize population sparsity , we compared the average responses of the sustained ensemble for with the sparse ensemble . Specifically , we used the converged STRFs to analyze a held-out set of natural stimuli , computed a histogram of the population responses at each time , and computed the average histogram across the entire test input ( see Methods ) . Since the sparse ensemble was optimized to yield a highly kurtotic firing rate distribution , it was of interest to examine the shape of the distribution when promoting sustained responses . Results comparing the average histograms of sustained versus sparse responses is shown in Figure 8 , with log-probabilities shown on the vertical axis to emphasize differences between the tails of the distributions . The main observation is that both the sustained and sparse ensembles have distributions that have long tails and are are highly peaked around a firing rate of zero . For reference , we show the average histograms obtained by filtering the stimulus through the first 400 principal components of the stimulus ( see Supplemental Figure 3 ) as well as through a set of 400 random STRFs; a zero-mean , unit variance Gaussian distribution is also shown . Therefore , despite promoting temporally persistent responses , the sustained responses yield a population response that is not altogether different from an ensemble that explicitly maximizes kurtosis . Interestingly , this observation was also made by Berkes and Wiscott in the context of complex cell processing in primary visual cortex ( see Sec . 6 of [33] ) . Finally , we sought to explore the consequences of relaxing the constraint that the responses be mutually uncorrelated . Rather than directly constrain the responses , we considered constraints to the shapes of the model STRFs . This was achieved by solvingi . e . , we require the STRFs to form an orthonormal basis . So long as the stimuli are bounded , this set of constraints meets our requirements that ( 1 ) the output of the STRFs be bounded and ( 2 ) we minimize redundancy in the learned ensemble . As before , the optimization is described in the Methods . We consider an ensemble size of STRFs initialized at random . Examples of shape-constrained STRFs that optimize the sustained objective function for are shown in Figure 9 . Again , we observe STRFs that are bandpass , localized , oriented , and sensitive to a variety of spectral and temporal input . However , there was an apparent difference between the speed of the spectro-temporal modulations and those from STRFs learned subject to the response constraints . It is well known that natural sound ensembles are composed largely of slow spectro-temporal modulations [29] , [30] , [34] . However , the emergent STRFs learned subject to response constraints appear to be tuned to relatively fast spectral and temporal modulations , whereas the STRFs learned subject to shape constraints appear to have a broader tuning . To further examine how both sets of constraints jointly capture and are related to the spectro-temporal modulations observed in stimulus , we compared the average 2D modulation profile of the stimulus to the eMTFs derived from both sets of constraints . An interesting view of how the emergent STRFs capture the spectro-temporal modulations of the stimulus is illustrated in Figure 10 for . Shown is the average 2D modulation profile of the stimulus overlaid with a single isoline contour ( at the 65% level ) of the eMTFs learned subject to response ( thick red lines ) and shape constraints ( thick black lines ) . We also show the constellation of BR versus BS for each ensemble ( indicated by ‘’ and ‘’ for response and shape constraints , respectively ) . As implied by the contours , the response constraints yield STRFs that follow the spectro-temporal “edge” of the stimulus , while the shape constraints explicitly capture most of the “slowness” of the stimulus . As mentioned previously , the response constraints effectively force the temporal response of the sustained ensemble to be sparse , which consequently results in highly selective STRFs that tend to be tuned to fast modulations . Nevertheless , they implicitly capture the spectro-temporal extent of the stimulus . Moreover , since the shape constraints effectively force the STRFs to form a basis that spans the input space , this results in neurons that explicitly capture the slow modulations of the stimulus . Similar observations were made across the range of , and for each case it was clear that the spectro-temporal modulations of the stimulus are fully captured by the combination of both sets of constraints .
The combination of shape and response constraints on the sustained objective function yield STRF ensembles that appear to jointly capture the full range of spectro-temporal modulations in the stimulus . However , the distinct differences in MTF coverage illustrate the tradeoff between redundancy and efficiency in sensory representations . In particular , the shape constraints yield STRFs that are somewhat akin to the first few principal components of the stimulus ( see Supplemental Figure 3 ) . This is not surprising given that the objective function defines a notion of variance of linear projections , the component vectors of which are constrained to form an orthonormal basis . However , since the responses are not strictly enforced to be uncorrelated , orthonormality imposed on the filter shapes does not necessarily reduce redundancy in the resulting neural responses . In contrast , the response constraints yield STRFs that are highly selective to the input and are thus comparatively “fast” in the modulation domain . This representation can be thought of as more efficient since at any given time only a few neurons have a large response . However , while the shapes of individual STRFs fail to explicitly capture the slow spectro-temporal modulations predominant in natural sounds , it instead appears that the ensemble MTF of the response-constrained STRFs collectively forms a contour around the high-energy modulations of the stimulus that implicitly capture its spectro-temporal extent . Is this contouring of the average modulation spectrum of natural sounds something performed by the auditory system ? The neural STRFs we considered certainly had an eMTF that reflects a tuning to slower modulations near the MTF origin . However , there is some evidence that the auditory system uses an “edge”-sensitive , discriminative modulation profile for analyzing sound . Woolley et al . [36] , in an avian study , showed that the eMTF of neurons from Field L ( the avian A1 analog ) has a bandpass temporal modulation profile ( at low scales ) that facilitates a discriminative tuning of temporal modulations among classes of natural sounds . Nagel and Doupe [37] have also shown examples of avian Field L STRFs that orient themselves near the spectro-temporal “edge” of the stimulus space . Moreover , Rodriguez et al . [38] , in a study of mammalian IC neurons , showed that neural bandwidths can scale to better capture fast , but less frequently occurring , modulations . In light of these observations , the modulation profiles observed from the sustained STRFs for both response and shape constraints are consistent with the notion that the auditory system makes an explicit effort to capture all modulations present in natural sounds: fast , feature-selective , and consequently discriminative modulations , as well as frequently occurring slow modulations . The notion that sustained neural firings form part of the neural representation of sensory systems is not limited exclusively to the auditory modality . In fact , the sustained firing objective considered in this paper is related to a broad class of sensory coding strategies referred to collectively under the temporal slowness hypothesis . This concept proposes that the responses of sensory neurons reflect the time-course of the information-bearing components of the stimulus—which are often much slower with respect to the fast variations observed in the stimulus—and may therefore reflect invariant aspects of the sensory objects in the environment . Examples of early neural network models exploring slowness as a learning principle were considered by Földiák [39] , Mitchison [40] , and Becker [41] . More recently , a number of computational studies , particularly in vision , have established slowness as a general sensory coding strategy and have revealed relationships with a number of general machine learning techniques . Here we outline the connections between the sustained firing criterion considered in this study and previous work . Our definition of the sustained firing objective , , was adapted from a notion of temporal stability proposed by Hurri and Hyvärinen termed temporal response strength correlation ( TRSC ) [18] . This study considered modeling of simple cells in primary visual cortex , and their objective function was defined as ( 3 ) for a single fixed . By maximizing subject to the decorrelation constraints , they showed the emergence of spatial receptive fields similar to those observed in simple cells in primary visual cortex . It is clear that the objective functions and are equivalent for a single time step , but the main difference between the two is that we sought to enforce temporal stability over a time interval , rather than between two distinct times and . Interestingly , optimization of the TRSC objective was shown by Hyvärinen to yield a solution to the blind source separation problem [42] , suggesting perhaps that in the auditory domain , such a criterion may underlie separation of overlapping acoustic sources . The sustained firing objective is also related to a well-known model of temporal slowness known as slow feature analysis ( SFA ) [20] . The computational goal of SFA is to find a mapping of an input that extracts the slow , and presumably more invariant , information in the stimulus . Briefly , for an input , linear SFA finds mappings that minimize ( 4 ) subject to , , and . Note that the input is not necessarily the raw stimulus but could represent a non-linear expansion of the input , akin to applying a kernel function in a support vector machine [43] . Therefore , SFA finds a mapping of the input that varies little over time and whose outputs are bounded and mutually uncorrelated . In the visual domain , Berkes and Wiskott found that SFA could explain a variety of complex cell phenomena in primary visual cortex such as the emergence of Gabor-like receptive fields , phase invariance , various forms of inhibition , and directional sensitivity [33] . Similar to our study , they also found the emergence of a sparse population code based on SFA . More importantly , however , they established a link between SFA at the level of complex cells and , which in turn links to the sustained firing objective explored in our study . Specifically , they showed that when a complex cell output is expressed as a quadratic form [35] , [44] , the SFA objective could be written as ( 5 ) which is equivalent to maximizing ( and thus for a single time-step ) plus cross-correlation terms . As noted by Berkes and Wiskott , this relationship suggests that sustained firing rates at the level of simple cells are modulated as part of a hierarchical cortical processing scheme in primary visual cortex . Given the increasing understanding of such hierarchical circuits in the auditory system [45] , the possibility that sustained firing rates are varied as part of a higher-order processing strategy in primary auditory areas is an exciting prospect worth further exploration . Other important relationships exist between SFA and a number of general machine learning principles . Blaschke et al . [46] established a relationship between SFA and independent component analysis , a widely used method for blind source separation ( see , e . g . , [47] ) . Klampfl and Maass [48] showed that under certain slowness assumptions about the underlying class labels in observed data , SFA finds a discriminative projection of the input similar to Fisher's linear discriminant . Furthermore , SFA has links to methods for nonlinear dimensionality reduction: Creutzig and Sprekeler [49] described the link between SFA and the information bottleneck whereas Sprekeler [50] showed a connection between SFA and Laplacian eigenmaps . In summary , the temporal slowness hypothesis forms a sound basis for learning a representation from data with rich temporal structure . Slowness as a learning principle has also been shown to explain the emergence of simple and complex cell properties in primary visual cortex . As described above , the sustained firing principle considered in this paper has fundamental links to SFA , which in turn is related to a number of general machine learning strategies . To the best of our knowledge , ours is the first thorough study that establishes a link between the temporal slowness hypothesis and an emergent spectro-temporal representation of sound in central auditory areas . The ensemble modulation coverage results are particularly interesting since it is widely thought that “slow” spectro-temporal modulations carry much of the message-bearing information for human speech perception . Furthermore , it is known in the speech processing community that features that capture slow temporal [51] and joint spectro-temporal modulations [52] , [53] are important for noise-robust automatic speech recognition . The observed contouring effect resulting from the sustained firing criterion may thus reflect a mechanism to detect the spectro-temporal “edges” of the message-bearing components of the stimulus , and possibly contribute to a noise-robust representation of sound . We have recently considered this principle and have demonstrated that 2D bandpass filters derived from eMTF contours learned from a speech-only stimulus yield state-of-the-art noise-robust acoustic features for automatic speech recognition [54] . Moreover , it is possible that the contour level may be chosen adaptively as a function of ambient signal-to-noise ratio to better capture variations in the high-energy modulations of the stimulus . Also , since the emergent STRFs capture general spectro-temporal patterns that characterize the stimulus , it is possible that ensembles of STRFs could be learned in various speech-plus-noise scenarios to perhaps better characterize noise-corrupted acoustic environments . Such hypotheses can be readily verified experimentally and may have practical impact to automated sound processing systems in noisy acoustic environments . Finally , the framework considered in this paper can be extended in a number of ways . For instance , to address the linearity limitation of the STRF , it is worthwhile to consider a model based on a linear-nonlinear cascade [55] . As mentioned earlier , the auditory pathway is necessarily hierarchical , and warrants consideration of hierarchical computational models . Indeed , recent physiological evidence also indicates that the representation becomes increasingly complex and nonlinear as one moves from away thalamo-recipient layers in primary auditory cortex ( for a review , see [45] ) . Finally , a recent computational study in vision by Cadieu and Olshausen [56] proposes a hierarchical generative model that explicitly unifies notions of sparse coding and temporal stability . In particular , a two-layer network learns a sparse input representation whose activations vary smoothly over time , whereas a second layer modulates the plasticity of the first layer , resulting in a smooth time-varying basis for image sequences . One can imagine that such a framework could be extended to spectro-temporal acoustic stimuli .
An ensemble of natural sounds comprising segments of speech , animal vocalizations , and ambient outdoor noises was assembled for use as stimuli . Two sets were generated , one for training and one for evaluating the response characteristics of the STRFs . Phonetically balanced sentences read by male and female speakers were used [57] . Examples of animal vocalizations included barking dogs , bleating goats , and chattering monkeys [58] . The ambient sounds included , for example , babbling creeks and blowing wind , and other outdoor noises . The speech utterances were approximately three seconds each and comprised 50% of the stimulus . The animal vocalizations and ambient sounds formed the remaining 50% of the stimulus ( 25% each ) , were broken into three-second segments , and were windowed using a raised cosine window to avoid transient effects . Finally , segments from each class were downsampled to 8 kHz , standardized to be zero-mean and unit variance , and randomly concatenated to yield a waveform approximately three minutes in overall length , i . e . , 90 seconds of speech , 45 seconds of animal vocalizations , and 45 seconds of ambient outdoor noises . We used a computational model of peripheral processing to account for the transformation of a monaural acoustic stimulus to a joint time-frequency representation in the auditory midbrain; this representation is referred to as an auditory spectrogram [59] , [60] . The auditory spectrogram represents the time-varying spectral energy distribution on the ( logarithmic ) tonotopic axis , and accounts for the physiology of inner hair cell transduction and filtering on the auditory nerve , enhanced frequency selectivity in the cochlear nucleus via a lateral inhibitory network , and the loss of phase locking to stimuli observed in midbrain nuclei . The specific model details have been presented previously and as such we forego a detailed description here , except to note that we sampled the log-frequency axis over six octaves with ten equally spaced channels per octave , with a short-term integration interval of 5 ms , i . e . , we obtained a 60 channel spectral vector every 5 ms . An example auditory spectrogram is shown for a segment of speech in Figure 1A . To quantify the relationship between a spectro-temporal stimulus and its corresponding response in central auditory areas , we used the spectro-temporal receptive field . Such a functional characterization of a neuron is useful for identifying the components of the stimulus to which it is most sensitive . An STRF models the linear transformation of a time-varying spectro-temporal input to an instantaneous firing rate , i . e . , ( 6 ) where is an LTI filter that defines the STRF , is a spectro-temporal stimulus , and is the average firing rate . Without loss of generality , we assume . Observe that the mapping represents convolution in time and integration across all frequencies , and we can interpret the STRF as a matched filter that acts on the input auditory spectrogram . For discrete-time signals and filters , and assuming that has a finite impulse response , we can express Eq . 6 compactly in vector notation as ( 7 ) where are column vectors denoting the stimulus and filter , respectively [61] . Furthermore , to express the response of an ensemble of neurons , we concatenate the STRFs into a matrix and write ( 8 ) From the stimulus auditory spectrogram , we extracted 250 ms spectro-temporal segments once every 5 ms . Each segment was stacked columnwise into a vector where ( i . e . , 50 vectors/segment 60 channels ) . A total of 30 k spectro-temporal vectors were extracted from the stimulus . We subtracted the local mean from each segment and scaled each vector to be unit norm [18] , and note that this pre-processing was also applied to the test stimulus used for evaluating the STRF response characteristics . Finally , each spectro-temporal input patch was processed by the ensemble of STRFs to yield a population response . Figure 1B illustrates the procedure for obtaining stimulus vectors and response vector . To constrain the responses of the STRFs to have unit variance and be mutually uncorrelated , we first note that the individual constraints can be written aswhich can then be compactly expressed as an ensemble constraint ( 9 ) where denotes the sample covariance matrix and is the identity matrix . Since is real-symmetric , it is unitarily diagonalizable as , where is a matrix of ( columnwise ) eigenvectors with corresponding eigenvalues along the diagonal of . Substituting this decomposition into Eq . 9 , we obtainedwhere . By recasting the constraints , we can rewrite the original matrix of STRFs as and consequentlywhere corresponds to a whitening of the input acoustic data , i . e . , has a spherical covariance matrix . For computational efficiency , we reduced the dimensionality of the input using a subset of the principal components of the stimulus , i . e . , where and , , are the matrices of eigenvalues and eigenvectors , respectively , that captured 95% of the variance of the input . In this work , we found . Therefore , the core problem we wished to solve is: ( 10 ) where corresponded to either the sustained firing or sparse coding objective function . To optimize this nonlinear program , we used the gradient projection method due to Rosen , the basic idea of which is as follows [62] , [63] . Let denote the update to the matrix of ( rotated and scaled ) STRFs , let be a learning rate , and let be an integer used to adjust the learning rate . Assume is a matrix with orthonormal columns that is a feasible solution to the problem in Eq . 10 . We updated via gradient ascent as follows: ( 11 ) where is a projection of the gradient update so that satisfies the orthonormality constraint required in Eq . 10 . If the update was such that , we set and recomputed the projected gradient update , repeating until was non-decreasing . Finally , learning ceased when the relative change between and fell below a threshold or a maximum number of iterations were reached; in our experiments , we stopped learning for or a maximum number of 30 iterations . Upon convergence , the desired STRFs were obtained using . Note that for the case of the sustained firing objective , was formed from the sum of independent terms , allowing us to directly sort the emergent STRFs according to their contribution to the overall objective function; such a sorting was not possible for the sparsity objective . Of course , the above procedure required a suitable projection , and one was derived as follows [64] . In general , for a matrix , we wish to find a matrix with orthonormal columns that minimizesIntroducing a symmetric matrix of Lagrange multipliers , and recalling that , we sought to find a stationary point of the LagrangianComputing the ( elementwise ) partial derivative of w . r . t . and setting it to we obtained [65]Observing thatwe have thatAssuming had full column rank , then an optimal orthogonal matrix that minimized that can be used for the projection in Eq . 11 was found as ( 12 ) Finally , to optimize a given objective function subject to the STRFs being orthonormal , i . e . , , we solveHere we can again use Rosen's projected gradient method in Eq . 11 along with the projection defined in Eq . 12 , but the only difference from before is that it does not require pre-whitening of the stimulus . We first characterized the emergent STRFs based on parameters that described their individual spectro-temporal and modulation tuning . Next , we considered measures that characterized a variety of ensemble-based spectro-temporal and modulation properties . To summarize the spectro-temporal modulations present in the natural sound stimulus , we averaged the magnitude of the 2D Fourier transform of 250 ms patches ( non-overlapping ) of the auditory spectrogram . The optimization procedure resulted in a set of richly structured patterns that suggested the presence of a number of latent classes whose membership varied with both choice of objective function and correlation interval . To quantify these variations , we applied the normalized spectral clustering algorithm of Ng et al . [66] . We defined the similarity between a given pair of STRFs and by computing the normalized 2D cross-correlation matrix for arbitrary shifts in time and frequency and selecting the maximum of the absolute value of this matrix , i . e . , whereImportantly , the absolute value of the cross correlation was used here since we wished to group STRFs regardless of whether they were excitatory or inhibitory . Next , we pooled all STRFs we sought to cluster and constructed a pairwise similarity matrix . Viewing as a fully connected graph with edge weights specified by , spectral clustering finds a partitioning of the graph into groups such that edges between groups have low similarity whereas edges within a group have high similarity . Defining the degree matrix where and unnormalized graph Laplacian , the normalized spectral clustering algorithm is as follows: We clustered the STRFs initially into 12 groups . While this number was necessarily an arbitrary choice , it was found to sufficiently capture variations in population diversity with . However , we found that ( i ) three of the resulting clusters could be reasonably labeled as noisy , whereas ( ii ) two of the resulting clusters could be reliably labeled as localized; merely reducing the number of initial classes did not merge the clusters , but instead blurred distinctions among the other major categories we sought to study . We interpreted noisy patterns as those with no obvious spectro-temporal structure and not indicative of any subset of the stimulus . Merging of the initial 12 classes was achieved by computing the average of STRFs from the initial class labels and ranking the classes in descending order . Indeed , the three noisy classes had the highest average and consequently resulted in a group with average greater than 0 . 5 . Similarly , the localized STRFs were typically highly spherical and sorting the initial clusters by resulted in the two localized classes to be ranked highest . Consequently , we grouped these two clusters that had an average of greater than 0 . 69 . This resulted in a final cluster count of nine classes . We obtained ensembles of neural STRFs estimated using TORC [31] and speech stimuli [27] , [32] . There were 2145 TORC and 793 speech STRFs , and each STRF was pre-processed to cover 110 ms in time ( sampling rate = 100 Hz ) and span 5 octaves in frequency ( sampling rate = 5 cyc/oct ) . For the spectral clustering analysis , we subsampled the TORC set by randomly selecting 793 STRFs and combined them with the speech STRFs , yielding a total of 1586 STRFs in the neural data set . In this way , the neural data analysis was not biased towards one stimulus type or the other .
|
We explore a fundamental question with regard to the representation of sound in the auditory system , namely: what are the coding strategies that underlie observed neurophysiological responses in central auditory areas ? There has been debate in recent years as to whether neural ensembles explicitly minimize their propensity to fire ( the so-called sparse coding hypothesis ) or whether neurons exhibit strong , sustained firing rates when processing their preferred stimuli . Using computational modeling , we directly confront issues raised in this debate , and our results suggest that not only does a sustained firing strategy yield a sparse representation of sound , but the principle yields emergent neural ensembles that capture the rich structural variations present in natural stimuli . In particular , spectro-temporal receptive fields ( STRFs ) have been widely used to characterize the processing mechanisms of central auditory neurons and have revealed much about the nature of sound processing in central auditory areas . In our paper , we demonstrate how neurons that maximize a sustained firing objective yield STRFs akin to those commonly measured in physiological studies , capturing a wide range of aspects of natural sounds over a variety of timescales , suggesting that such a coding strategy underlies observed neural responses .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"computational",
"neuroscience",
"audio",
"signal",
"processing",
"signal",
"processing",
"biology",
"computational",
"biology",
"sensory",
"systems",
"neuroscience",
"engineering",
"coding",
"mechanisms"
] |
2013
|
Sustained Firing of Model Central Auditory Neurons Yields a Discriminative Spectro-temporal Representation for Natural Sounds
|
Several pathways control time to flowering in Arabidopsis thaliana through transcriptional and posttranscriptional gene regulation . In recent years , mRNA processing has gained interest as a critical regulator of flowering time control in plants . However , the molecular mechanisms linking RNA splicing to flowering time are not well understood . In a screen for Arabidopsis early flowering mutants we identified an allele of BRR2a . BRR2 proteins are components of the spliceosome and highly conserved in eukaryotes . Arabidopsis BRR2a is ubiquitously expressed in all analyzed tissues and involved in the processing of flowering time gene transcripts , most notably FLC . A missense mutation of threonine 895 in BRR2a caused defects in FLC splicing and greatly reduced FLC transcript levels . Reduced FLC expression increased transcription of FT and SOC1 leading to early flowering in both short and long days . Genome-wide experiments established that only a small set of introns was not correctly spliced in the brr2a mutant . Compared to control introns , retained introns were often shorter and GC-poor , had low H3K4me1 and CG methylation levels , and were often derived from genes with a high-H3K27me3-low-H3K36me3 signature . We propose that BRR2a is specifically needed for efficient splicing of a subset of introns characterized by a combination of factors including intron size , sequence and chromatin , and that FLC is most sensitive to splicing defects .
The switch from the vegetative to the reproductive phase is an important developmental transition in flowering plants . The timing of this transition is regulated by several factors , including endogenous and environmental signals . In Arabidopsis , the photoperiod , vernalization and autonomous pathways involved in flowering time control have been investigated in much detail . More recently , additional pathways such as the age or the ambient temperature mediated pathways have been described [1–4] . Most flowering time pathways converge at the activation of a common set of genes that promote flowering and that are known as floral integrators , namely SUPPRESSOR OF OVEREXPRESSION OF CONSTANS1 ( SOC1 ) , FLOWERING LOCUS T ( FT ) and LEAFY ( LFY ) , and at the repression of the major flowering repressor FLOWERING LOCUS C ( FLC ) [5] . The daily light duration is sensed by the photoperiod pathway . In temperate climates , Arabidopsis and many other species flower earlier in long day ( LD ) than short day ( SD ) conditions [5 , 6] . In the photoperiod pathway , CONSTANS ( CO ) activates FT expression in leaves . FT protein is a major mobile flowering-inducing signal and moves through the phloem into the shoot apical meristem ( SAM ) where it changes vegetative meristem identity to flowering [5] . Normal expression of CO in long day ( LD ) photoperiods requires the histone-binding protein MSI1 , and partial loss of MSI1 function such as in the partially complemented msi1 mutant line msi1-tap1 leads to reduced expression of CO , failure of FT and SOC1 activation and to delayed flowering [7 , 8] . FT and SOC1 are repressed by FLC in LD and SD [5] . Prolonged cold ( vernalisation ) inactivates FLC expression thus strongly shortening the time to flowering [9] . The autonomous pathway is known to promote flowering independently of environmental signals . Mutants in the autonomous pathway are extremely late flowering due to strong upregulation of FLC . The autonomous pathway genes belong to two main subfamilies: ( i ) chromatin modifiers such as FLOWERING LOCUS D ( FLD ) [10] , FVE [11 , 12] and RELATIVE OF EARLY FLOWERING 6 [13] and ( ii ) RNA binding proteins ( RBP ) such as FCA [14] , FPA [15] , FY [14] , FLOWERING LOCUS K ( FLK ) [16 , 17] and LUMINIDEPENDENS ( LD ) [18] . Together , the autonomous pathway genes form a group of partially independently acting genes rather than a classical linear genetic pathway [19] . Transcripts of many plant genes including most of the flowering-related genes contain several introns . Splicing removes the non-coding introns from pre-mRNAs to form mature mRNA ( for review see [20 , 21] . The spliceosome is a macromolecular complex consisting of five highly conserved small nuclear ribonucleoprotein particles ( snRNPs; U1 , U2 , U4 , U5 and U6 ) and a large number of stabilizing proteins [22] . The splicing reaction can be functionally divided into several steps , including spliceosome assembly , activation , catalysis , and disassembly of the spliceosomal machinery . During the activation step , DExD/H-box RNA helicases are known to play key roles [23–25] . DExD/H-box RNA helicases belong to a large , highly conserved protein family . These proteins play roles in many biological processes related to RNA metabolism , using energy from ATP hydrolysis [26] . RNA processing is much less studied in plants than in animals and yeast . However , during the last decade , the functional role of transcript processing in plants has received some attention ( for review see [20 , 21] . Several lines of evidence support a connection between RNA processing and flowering time control [27–30] . The proteins identified in pre-mRNA processing were involved in either 3’ end polyadenylation or 5’ end capping . However , little is known about the possible regulatory role of key proteins of the spliceosomal complex in control of flowering . Here we describe an early flowering allele of Arabidopsis BRR2a , which encodes a highly conserved spliceosome protein in eukaryotes . The single missense mutation of threonine at position 895 is associated with an early flowering phenotype . We demonstrate that defects in FLC splicing form the mechanism underlying the flowering phenotype . Genome-wide experiments established that full BRR2a activity was required only for a small group of introns . We propose that BRR2a is specifically needed for efficient splicing of a subset of introns characterized by a combination of risk factors in intron size , sequence and chromatin composition , and that FLC is most sensitive to splicing defects .
To understand the molecular mechanisms underlying the control of the floral transition by MSI1 [7 , 8] , a mutant screen for suppressors of the late flowering phenotype of msi1-tap1 plants was performed , resulting in six mutants with a shortened vegetative phase of msi1-tap1 [31 , 32] . Two suppressor mutants had been reported previously [31 , 32] . Here we describe the analysis of one of the remaining uncharacterized suppressor mutants , which was initially called chrottapösche ( cäö ) ( Swiss German for dandelion ) because of its increased leaf serration . To test whether the cäö early flowering phenotype was independent of the msi1-tap1 background , cäö was backcrossed to Columbia ( Col ) , and flowering time was measured . Under LD and SD conditions , msi1-tap1 flowered much later than Col , confirming earlier results [7 , 8] , while msi1-tap1 cäö flowered at a similar time to Col ( Fig 1 ) . In the Col background , cäö flowered earlier than both Col and msi1-tap1 cäö , demonstrating that the effect of cäö does not require the msi1-tap1 background . Therefore , cäö in the Col background was used in all subsequent experiments . In addition to early flowering , the cäö plants had other developmental defects ( Fig 2 ) . Leaves of cäö were small and serrated contributing to a smaller and more compact rosette ( Fig 2A–2C , S1A Fig ) . Siliques of cäö were shorter than WT ( Fig 2D ) , had reduced seed set and contained unfertilized ovules . In about 20% of cäö ovules female gametophyte development was delayed or completely absent ( Fig 2E , S1B Fig ) . Defective female gametophyte development had a sporophytic origin because it was mainly observed in homozygous cäö-/- plants and only occasionally in heterozygous cäö-/+ plants ( S1B Fig ) . Together , CÄÖ is important not only for normal flowering time but also other developmental programs including formation of female gametophytes . In order to isolate the causative mutation in cäö , a mapping population was established by crossing cäö in the Col background with Ler followed by next generation sequencing of F2 bulk segregants . A candidate region on the left arm of chromosome 1 ( S2 Fig ) contained only one mutation that was represented in all reads covering this region . The mutation was a G to A transition in the AT1G20960 gene and caused a T895I missense mutation ( Fig 3 ) . This mutation was subsequently confirmed by a specific dCAPS ( derived cleaved amplified polymorphic sequences ) molecular marker ( S3A Fig ) and Sanger sequencing ( S3B Fig ) . AT1G20960 , which was previously identified as EMBRYONIC LETHAL 1507 ( EMB1507 ) [33] , encodes an orthologue of yeast Brr2p ( Bad response to refrigeration 2 protein ) and is also called BRR2a [34] . Yeast Brr2p is a DEAD/DExH box ATP-dependent RNA helicase with a unique N-terminal domain and two consecutive helicase cassettes ( with a DExD/H domain and helicase superfamily C-terminal domain ) , each followed by a Sec63 domain ( Fig 3B ) [35 , 36] . Yeast and animal BRR2 proteins are integral components of the U5 small nuclear ribonucleoprotein ( snRNP ) and are essential for splicing through their contribution to the recruitment and activation of spliceosome complex components [24 , 37] . Because Brr2 is the original name given to these proteins , we refer to the mutant protein as BRR2a-T895I and to cäö as brr2a-2 . The mutated threonine 895 is located at the end of the first helicase domain and highly conserved in Brr2 proteins ( Fig 3C ) , suggesting that the T895I mutation interferes with BRR2a function . To confirm that the early flowering phenotype is caused by the disruption of BRR2a , an allelism test was performed . Heterozygous emb1507-4 null mutant plants were used to pollinate homozygous brr2a-2 plants . All F1 plants with the emb1507-4 allele but none of the plants without it displayed the brr2a-2 phenotype , establishing that the mutant BRR2a protein caused the cäö mutant phenotype . Considering the recessive nature of the brr2a-2 mutant , these data suggest that brr2a-2 is a hypomorphic rather than a neomorphic allele and that the cäö phenotype is caused by reduced activity of BRR2a . BRR2 sequences from different eukaryotes including yeast , metazoa , protozoa and plants were aligned and a phylogenetic tree was constructed ( S4 Fig ) . Arabidopsis BRR2a has two paralogues , BRR2b ( At2g42270 ) and BRR2c ( At5g61140 ) . BRR2a and BRR2b result from a recent gene duplication and are part of a clade with members in all green plants . In contrast , BRR2c belongs to a minor clade with only one protein from the fern Selaginella and one yeast protein . Furthermore , BRR2a shares 82% identity and 91% similarity with BRR2b but only 40% identity and 59% similarity with BRR2c ( S1 Table ) . Although BRR2a and BRR2b are conserved , the obvious phenotype of the brr2a mutant indicates that the genes do not have redundant functions . Data in the Arabidopsis eFP Browser [38] show that BRR2 genes are expressed in most tissues at variable levels ( S5 Fig ) , with BRR2a having the highest transcript levels . BRR2c was expressed considerably less than BRR2a , and BRR2b was expressed lowest of the three BRR2 homologues in most tissues . It is likely that the high expression of BRR2a accounts for the strong phenotypes of brr2a mutants . FLC is a MADS-box DNA binding protein and a major repressor of flowering time in Arabidopsis [39 , 40] . Many early flowering Arabidopsis mutants have reduced FLC expression whereas many late flowering mutants have increased FLC expression . We tested for possible changes of FLC transcript levels in 15 days old seedlings grown under SD conditions . Col plants containing an active FRIGIDA ( FRI ) allele , which express high levels of FLC , were included as control . As reported before , the expression levels of FLC were much higher in FRI than in Col [41] . In contrast , FLC levels were strongly ( >95% ) reduced in brr2a-2 ( Fig 4 ) . The MADS AFFECTING FLOWERING ( MAF ) genes are homologues of FLC and are often similarly regulated [42 , 43] . In brr2a-2 , expression levels of MAF1 and MAF4 mRNAs were about 50% reduced and MAF2 , MAF3 and MAF5 mRNA levels were also lower than in WT ( Fig 4B and 4C ) . Together , the results are consistent with the observed cäö early flowering phenotype and the effect of FLC on flowering time control [39] . FT and SOC1 promote flowering and both are repressed by FLC . We therefore tested if early flowering of brr2a-2 was associated with increased FT and SOC1 expression . The expression of FT was nearly three times higher in brr2a-2 than in Col ( Fig 4D ) and SOC1 had significantly increased expression as well ( Fig 4E ) . These results are consistent with the decreased FLC levels in brr2a-2 and indicates that early flowering is caused by increased expression of FT and SOC1 that have been released from the repression by FLC . We tested genetically whether brr2a-2 accelerates flowering via FLC by measuring flowering time of the double brr2a-2 flc mutant ( Fig 4F ) . Consistent with earlier reports , flc flowered earlier than Col . The brr2a-2 mutant flowered even earlier than flc , possibly because of reduced expression of other floral repressors such as the MAF genes . Importantly , the brr2a-2 flc double mutant did not show further acceleration of flowering revealing complete epistasis of BRR2a and FLC , which is fully consistent with the reduced FLC expression as the major cause of accelerated flowering in brr2a-2 . The reduced FLC transcript levels did not correlate with altered transcript levels of major FLC activators or repressors ( S6 Fig ) . Splicing of COOLAIR , an antisense transcript covering the FLC locus [44] , contributes to repression of FLC transcription and depends on a homolog of the yeast spliceosomal PRP8 protein [30] . COOLAIR splicing is strongly disrupted in brr2a-2 ( S7 Fig ) and reduced levels of COOLAIR and FLC transcripts are consistent with correlated FLC and COOLAIR production [44] . In contrast to a prp8 mutant in which defective COOLAIR splicing increased FLC transcription [30] , defective COOLAIR splicing in brr2a-2 did not increase FLC transcript levels . Therefore we tested whether the reduced FLC transcript levels in brr2a-2 were due to defects in FLC transcript splicing . Intron retention ( IR ) was tested by qPCR for randomly selected FLC introns 1 , 5 and 6 . For all three tested FLC introns , IR was about 8-fold higher in brr2a-2 than in Col ( Fig 5A ) . These results suggest that the BRR2a-T895I mutation resulted in an unproductive splicing complex that caused IR and reduced accumulation of correctly spliced FLC in brr2a-2 . Incorrectly spliced transcripts are subject to nonsense-mediated decay ( NMD ) mRNA quality control and have a much higher turnover rate than correctly spliced transcripts [45] , which could explain the reduced FLC transcript levels . The splicing defects of the FLC transcript can explain the early flowering phenotype but it was possible that in brr2a-2 transcripts of other genes were affected as well . We used PCR assays to test IR in transcripts of the three MADS-box genes MAF1 , AG and SEP3 , which have a similar intron-exon structure as FLC . For each of these genes , retention of the intron corresponding to FLC intron 1 , which showed strong retention in brr2a-2 , was tested . Although IR was increased in the transcripts of all three genes ( Fig 5B ) it never exceeded 10% and was thus much less than for FLC of which 90% of the transcripts retained intron sequences , suggesting that not all transcripts depend to the same extent on functional BRR2a for correct processing . To investigate other potential splicing defects in brr2a-2 we performed an RNA-seq experiment . For each of three wild-type and three mutant libraries between 23 and 40 million reads where generated of which 83–92% could be mapped to the Arabidopsis TAIR10 reference genome . About 90% of the mapped reads were from non-overlapping , annotated genes ( S2 Table ) and 4% of the mapped reads corresponded to intron sequences . Analysis of differentially expressed genes ( DEG ) identified 279 genes with increased and 103 genes with decreased transcript levels ( S3 and S4 Tables ) . The gene with the strongest transcript reduction was FLC ( 6 . 6 fold , p = 2 . 1E-15 ) and consistent with the early flowering phenotype , SOC1 expression was significantly upregulated ( 3 . 1 fold , p = 7 . 7E-10 ) . In addition , the transcript level of the BRR2a homolog BRR2b was significantly increased ( 2 . 8 fold , p = 1 . 6E-11 ) . Similarly , the gene encoding the predicted splicing factor AtPRP8b , which is thought to function together with BRR2 [34] , was upregulated in brr2a-2 ( 3 . 2 fold , p = 4 . 3E-08 ) . Upregulation of BRR2b and AtPRP8b expression could be caused by an autoregulatory loop responding to a reduced function of the BRR2a-T895I mutant protein . It is possible that the upregulation of BRR2b and AtPRP8b distort stoichiometry in the splicosome and enhance the defects in brr2a-2 . Among the upregulated DEGs , four gene ontology ( GO ) categories were significantly overrepresented ( p<0 . 05 , S5 Table ) . The GO categories "response to UV-B" and "response to salicylic acid stimulus" are consistent with earlier reports of connections between splicing and stress responsiveness [21] . Also the categories "regulation of transcription , DNA-dependent" and “response to karrikin” were enriched among the upregulated genes . Only one GO category ( "proteolysis" ) was significantly enriched ( p = 0 . 02 ) among the downregulated genes ( S6 Table ) . Next , the transcriptome data were searched for misexpression of genes involved in leaf development . Of 103 genes from GO category GO:0009965 ( leaf development ) for which transcripts were detected , three were significantly stronger expressed in brr2a-2 than in wild type: TEOSINTE BRANCHED 1 , CYCLOIDEA , AND PCF FAMILY 13 ( TCP13 ) , KIPRELATED PROTEIN 6 ( KRP6 ) and KRP1 ( S8A Fig ) . Two leaf development genes were less expressed in brr2a-2 than in wildtype: ASYMMETRIC LEAVES 1 ( AS1 ) and TCP24 . Reduced expression of both TCP24 and AS1 is consistent with earlier reports that TCP24 is an activator of AS1 [46] . Increased expression of KRP6 or KRP1 causes reduced leaf size and increased serration [47 , 48] and is consistent with the leaf phenotype of brr2a-2 plants . The ASTALAVISTA program suite [49] was used to collate alternative splicing ( AS ) events in the six libraries ( Fig 6 ) . Relative frequencies of AS events in WT were similar to those reported earlier [50] . In brr2a-2 , almost twice as many AS events were detected than in wild type ( Fig 6A ) . This increased AS frequency was caused mainly by IR and partly by more complex events such as double intron retention ( Fig 6A–6C ) . Exon skipping ( ES ) as well as use of alternative acceptors or donors did not differ between wild type and mutant . Therefore , we focused on IR events and used DESeq2 [51] with numbers of intron-derived reads to detect differentially retained introns ( DRIs ) . There were 914 DRIs with significantly increased but only 74 with decreased retention ( S7 and S8 Tables ) . Thus , the BRR2a-T895I mutation causes primarily increased intron retention . The low number of retained introns ( 1 . 2% of 74’581 introns with available read counts or 0 . 8% of 117’458 analyzed introns ) indicates that only a specific subset of introns is strongly affected in brr2a-2 . RT-PCR assays using independent RNA samples confirmed increased IR in brr2a-2 for 10 out of 10 tested genes , suggesting a low false positive rate of detected DRIs ( S9 Fig ) . Of the leaf development genes with altered expression in brr2a-2 ( S8A Fig ) only intron 1 of TCP13 was significantly differentially retained ( p = 1 . 1E-02; S8B Fig ) . The weaker retention of the other intron of TCP13 was not significant ( p>0 . 05 ) . The differentially expressed leaf development genes KRP1 , KRP6 , TRCP24 and AS1 all contain introns but did not show signs of increased IR in brr2a-2 plants ( S8B Fig ) . Thus , it is possible that the leaf development phenotype of brr2a-2 is related to splicing defects in the TCP13 transcription factor . It was possible that certain genes depend critically on BRR2a for splicing of several introns . However , although the genes with DRIs contain on average six introns , for most of them only a single intron was significantly retained ( Fig 6D ) . Thus , intron retention in brr2a-2 appears to be intron-specific rather than a gene or transcript property . It was possible that DRIs are characterized by specific sequence signatures . We used the R package motifRG to test whether DRIs contained enriched motifs using the sequences of unchanged introns as background . However , there were no sequence motifs enriched in DRIs over unchanged introns . Similarly , splice site sequences did not differ between DRIs and unchanged introns ( S10 Fig ) . Next , we tested whether DRIs differ in length from unchanged introns . Less efficiently spliced DRIs were significantly shorter than unchanged introns ( mean of 140 bp vs . 173 bp; p = 1 . 7E-13 , one-sided t-test ) ( Fig 6E ) . Conversely , more efficiently spliced DRIs where significantly longer than unchanged introns ( mean of 360 bp vs . 173 bp; p = 7 . 2E-6 , one-sided t-test ) . In addition to length , DRIs differed also in GC content from unaffected introns ( Fig 6F ) . Less efficiently spliced DRIs had a significantly lower GC content than unaffected introns ( 31 . 8% vs . 32 . 9%; p = 1 . 6E-13 , one-sided t-test ) ; and more efficiently spliced DRIs had a significantly higher GC content than unaffected introns ( 35% vs . 32 . 9%; p = 4 . 8E-5 , one-sided t-test ) . We also tested an effect of intron folding using but stability of the most likely secondary structure did not differ between retained and spliced introns ( see Materials and Methods for details ) . Thus , the BRR2a-T895I mutation appears to reduce splicing efficiency preferentially of short and GC-poor introns and shifts it to longer , GC-rich introns . Because chromatin can affect splicing [52] , we tested whether chromatin on DRIs differs from chromatin on other introns . To exclude confounding effects by transcription strength , DRIs were compared to two different control sets: ( i ) Non-differentially retained introns from the genes that have at least one DRI , and ( ii ) introns from a set of genes that do not contain DRIs and that have a similar median expression as the genes with DRIs . DRIs did not differ significantly from control introns regarding CHG methylation , CHH methylation , H3 density and H3K9me2 ( S11 Fig ) . In contrast , DRIs had less H3K4me1 and CG methylation ( mCG ) , and more H3K4me3 than control introns ( Fig 7 ) . H3K9ac was higher on DRIs than on the non-DRI introns in the same genes but similar to the level on introns of control genes . H3K27me3 and H3K36me3 did not differ between DRIs and non-DRI introns in the same genes but were higher and lower , respectively , on exons and introns of genes with DRIs than on control genes . In addition , exons on genes with DRIs had more H1 . 1 , H1 . 2 , and H3K27me3 and less H3K9ac than exons on control genes . Thus , BRR2a function is most important on genes generally rich in H3K27me3 and H1 , and low in H3K36me3 and H3K9ac . It is also most important for splicing of introns with low mCG and H3K4me1 , and high H3K4me3 . This shows that local chromatin properties can affect the outcome of a mutation in a spliceosomal subunit .
Yeast and human BRR2 proteins are evolutionary highly conserved spliceosome proteins of about 200 kDa [53 , 54] . They belong to the DEAD/DExH-box family of ATP-dependent RNA helicases with two putative RNA helicase domains , each with a highly conserved ATPase motif , followed by a SEC63 domain . Their ATPase and helicase activities are involved in rearrangements necessary for spliceosome activation through the unwinding of U4/U6 snRNP [24] . After the unwinding , BRR2 remains stably associated with the catalytic core of the spliceosome [55] and eventually promotes spliceosome disassembly by unwinding U2/U6 [56] . BRR2 also functions in promoting conformational rearrangements in the spliceosome during the first-to-second-step transition , which aid 3’ splice site positioning and formation of the second-step catalytic center [57] . BRR2 helicase activities are highly regulated to ensure the correct timing of spliceosome activation or disassembly [58] . Regulators of BRR2 functions include Prp8 [59 , 60] and the Snu114 GTPase [56] . Here , we identified a new allele of BRR2a containing a T895I mutation , which enabled us to demonstrate that Arabidopsis BRR2a functions in intron splicing . In BRR2a , threonine 895 and its neighboring amino acid sequences are highly conserved and located near the ATPase domain within the first conserved helicase I motif . A crystal structure and structural models for different yeast and human BRR2 helicase regions show a possible reorganization and pairing of these domains during the splicing process [61] . Furthermore , mutagenesis studies in the amino-terminal helicase domain revealed the role of this domain in splicing efficiency [23 , 62] . Together , the T895I mutation in brr2a-2 likely results in a partial loss of BRR2a function or impairs interaction of BRR2a with other proteins in the spliceosome . Plausible interactors are PRP8 and GAMETOPHYTE FACTOR 1 ( GFA-1 ) [34] , which are Arabidopsis homologues of the Brr2p regulatory yeast proteins Prp8 and Snu114 [63] , respectively . Indeed , GFA1 interacts with the carboxy-terminal domain of BRR2a and with PRP8a in yeast two-hybrid assays [34] and BRR2a and PRP8a copurify [64] . However , it is not known whether mutation of T895 in the amino-terminal domain affects this interaction . Other BRR2a candidate interactors are homologues of the serine/threonine protein kinase Prp4 and its substrates such as Prp1 . These proteins interact biochemically with fission yeast Brr2p . Future experiments will establish whether the T895I mutation affects BRR2a’s ability to interact with other proteins . An alternative explanation for reduced function of BRR2a-T895I could be that the T895I mutation weakens BRR2a interactions with U4/U6 , U5 or pre-mRNAs . This appears plausible because in vivo UV cross-linking and RNA sequencing have established that budding yeast Brr2p binds not only to the U4/U6 and U5 snRNA but also to pre-mRNAs [57] . It is likely that the polar T895 is surface-exposed , and if a change at this residue reduces BRR2a-RNA interactions , this could also reduce the activity of BRR2a-T895I in the splicing reaction . In Arabidopsis , the BRR2a-T895I mutation results in early flowering and altered leaf development . Splicing defects in the transcripts of the TCP13 transcription factor and altered transcript levels of TCP13 , TCP24 , AS1 , KRP1 and KRP6 are consistent with the leaf phenotype of brr2a-2 plants . KRPs repress cell proliferation by inhibiting cyclin-dependent kinases [65] / Sablowski , 2014 #13362] . Increased expression of KRP6 or KRP1 causes reduced leaf size and increased serration [47 , 48] consistent with the concomitant upregulation of these genes and altered leaf size and shape in brr2a-2 . TCPs are major transcriptional regulators that control cell proliferation in leaves ( for reviews see [66 , 67] ) . TCPs related to TCP13 ( CIN-like TCPs ) function highly redudnantly [68]; they promote the arrest of cell division , and overactivation can strongly reduce leaf size and increase serration . Because the CIN-like TCP13 homolog TCP4 is a direct activator or KIP1 [69] , it is possible that increased expression of TCP13 causes KRP upregulation , leading to premature arrest of cell proliferation and reduced leaf size in brr2a-2 . Retention of TCP13 intron 1 does not alter transcript coding potential as this intron 1 is located in the 5’ untranslated region; it could , however , affect translation efficiency . FLC is a major repressor of flowering that functions by repressing FT and SOC1 [39 , 40 , 70 , 71] . Expression and genetic data support the model that reduced FLC transcript levels in brr2a-2 allow untimely activation of FT and SOC1 , which together cause accelerated flowering . No strong changes in the expression of known regulators of FLC were found in brr2a-2 but strong defects in FLC splicing efficiency were evident from intron retention . Whereas the proportion of intron-containing FLC transcripts was increased , total amounts of FLC transcripts was decreased in brr2a-2 . This is similar to the pep-4 mutant [72] and possibly a consequence of RNA quality surveillance such as NMD , a co-transcriptional quality control pathway that degrades aberrant transcripts [45] . In addition to the NMD pathway , other RNA quality pathways operate in the nucleus and degrade transcripts with delayed processing independent of the presence of stop codons [73] . Although a functional correlation between transcription and transcript processing was previously reported in yeast and mammals [74] , it remains unknown to what extent RNA quality pathways can feed back to repress transcription . Complete loss of Arabidopsis BRR2a function is embryo lethal [33] , and the phenotype of the cäö mutant suggests that brr2a-2 is a hypomorphic allele and that BRR2a-T895I has reduced function . The pleiotropic morphological defects of brr2a-2 plants revealed that full BRR2a function is needed in flowering control and also other developmental programs . The developmental role of BRR2a is consistent with previous reports of requirements for spliceosome components in animal and plants development [34 , 75–79] . In cases of embryo lethality hypomorphic alleles such as brr2a-2 constitute unique and powerful resources to analyse gene functions during postembryonic life . The restricted phenotypic changes of brr2a-2 plants are also consistent with the result that splicing of only a group of introns was affected . Thus , introns differ in their dependency on BRR2a activity , and FLC is particularly sensitive to the BRR2a-T895I mutation . Our discovery of the cäö mutant has revealed that mutations in different spliceosome proteins can affect distinct sets of introns . We have shown that FLC splicing is greatly reduced in brr2a-2 while a recent study showed that FLC splicing was normal in a prp8 mutant [30] . Genome-wide sequencing data indicate that retained introns often came from genes with a high-H3K27me3-low-H3K36me3 signature . Retained introns differed also in their local chromatin composition and had more H3K4me3 , less H3K4me1 and less mCG than control introns . In addition , splicing efficiency in brr2a-2 plants was often shifted from short , GC-poor to long , GC-rich introns . However , the case of FLC shows that even long introns can have splicing defects in brr2a-2 . In addition , not all short introns were retained in brr2a-2 consistent with a combined effect of intron length , chromatin features and maybe other properties . This notion is supported by the observation that FLC , which contains a retained long intron , has the high-H3K27me3-low-H3K36me3 signature . Similar to other retained introns , intron 1 of FLC has low H3K4me1 . Only 0 . 3% , 31 . 6% and 14 . 1% of all introns have higher H3K27me3 , lower H3K36me3 or lower H3K4me1 , respectively , than intron 1 of FLC . In contrast , H3K4me3 and mCG on FLC intron 1 were close to the average of all introns . Our finding that retained introns have particular chromatin signatures is consistent with recent reports about effects of chromatin on splicing in animals ( for review see [52 , 80] . Although the underlying mechanisms are poorly understood , two main models have been developed: ( i ) In the kinetic coupling model , local chromatin affects the rate of RNA Polymerase II ( Pol II ) elongation . Slow elongation or pausing favors weak splice sites while fast elongation misses weak sites and favors downstream strong sites . ( ii ) In the recruitment coupling model , chromatin-binding adaptor proteins recruit specific splicing regulators to define splicing outcome [52 , 80] . Alternative splicing of the human fibroblast growth factor receptor 2 ( FGFR2 ) , for instance , is mediated by H3K36me3-based recruitment of the polypyrimidine tract–binding protein splicing factor [81] . Notably , one of the alternative spliced states of human FGFR2 shares the high-H3K27me3-low-H3K36me3 chromatin signature of Arabidopsis genes with IR in brr2a-2 [81] . More recently , alternative splicing of FGFR2 was found to be affected by a long non-coding antisense RNA that recruits Polycomb-group proteins and the H3K36 histone demethylase KDM2 to establish a high-H3K27me3-low-H3K36me3 chromatin signature [82] . We note that also at Arabidopsis FLC , a long antisense RNA affects H3K36me3 to establish a high-H3K27me3-low-H3K36me3 chromatin signature [83] . Future work needs to address whether regulation of FLC and FGFR2 splicing share a common mechanistic basis . In addition to histone modifications , also DNA methylation is associated with splicing [84] . In mammals , introns often have lower mCG than exons , and recruitment of CCCTC-binding factor and methyl-CpG binding protein 2 to mCG can affect splicing by modulating Pol II elongation rates [84] . In maize , it has been proposed that CHG methylation at splice acceptor sites may inhibit RNA splicing [85] and loss of DNA methylation at a splice acceptor site in the oil palm DEFICIENS gene was associated with splicing defects in somaclonal variation [86] . It is not known whether mCG affects splicing in Arabidopsis but our results suggest a mechanistic link between mCG and IR . Despite the similarities between chromatin features found related to splicing outcomes , ES is the most prevalent alternative splicing event in mammals and IR is most prevalent in plants . Therefore , more work is needed to establish whether the mechanisms that couple local chromatin properties to splicing are shared or differ between animals and plants . In summary , our suppressor mutant screen for accelerated flowering led to the discovery of the early flowering brr2a-2 mutant . Our data suggest a model in which BRR2a functions in the spliceosome with the T895I missense mutation leading to reduced splicing efficiency for transcripts of a selected group of genes , most importantly FLC . Reduced FLC expression allows unscheduled transcription of FT and SOC1 to accelerate flowering . Together , our work establishes correct splicing as an important mechanism for flowering time control and uncovers a complex relation between chromatin features and splicing outcomes in Arabidopsis .
The Arabidopsis thaliana wild-type and T-DNA insertion lines were in the Columbia-0 ( Col ) background . FRI in Col [87] , msi1-tap1 and flc-6 [7] were described before; emb1507-4 ( NASC ID: N16092 ) was obtained from the Nottingham Arabidopsis Seed Stock Centre . The EMS-mutated allele cäö in the msi1-tap1 background was isolated in a mutant screen that was described before [31] . For further characterization , cäö in the msi1-tap1 background was backcrossed into Col . Seeds were sown on 0 . 5× basal salts Murashige and Skoog ( MS ) medium ( Duchefa , Haarlem , The Netherlands ) , stratified at 4°C for 2–3 day , and allowed to germinate in growth chambers at 20°C for 10 days under LD ( 16 h light ) or SD ( 8 h light ) photoperiods . Plantlets were planted in soil and grown in growth chambers under the same conditions . Flower buds were emasculated at anthesis and the non-pollinated pistils were collected 2–4 days after emasculation . The samples were fixed with ethanol-acetic acid ( 9:1 ) , washed for 10 min in 90% ethanol , 10 min in 70% ethanol and cleared over-night in a chloralhydrate solution ( 66 . 7% chloralhydrate ( m/m ) , 8 . 3% glycerol ( m/m ) ) . Ovules were observed under differential interference contrast ( DIC ) optics using a Zeiss Axioplan microscope ( Zeiss , Jena , Germany ) . Images were recorded using DFC295 Leica camera ( Leica , Wetzlar , Germany ) . A mapping population was established by crossing cäö with the polymorphic ecotype Ler , and total genomic DNA was extracted from 150 F2 plants presenting the mutant phenotype using the Nucleon Phytopure genomic DNA extraction Kit ( Amersham Bioscience , Uppsala , Sweden ) . After library preparation using standard Illumina protocols , the DNA was loaded onto an Illumina Genome sequencer GA IIx and run for 36 cycles . The obtained short reads were mapped against the TAIR10 release of the Arabidopsis genome using Bowtie 2 [88] . Genome-wide SNP positions and pileup information were then collected and filtered as recommended in the Next-generation EMS mutation mapping software [89] . dCAPS primers were designed using dCAPS Finder 2 . 0 [90] ( S9 Table ) . The amplified fragments from genomic DNA of the wild type Col and mutant cäö were digested with HpaI ( Fermentas , Helsingborg , Sweden ) and loaded on a 2 . 5% agarose gel . The SNP in cäö was further validated by standard Sanger sequencing using primers LH1609: CTTGAAGGAAGATAGTGTAACTCGT and LH1324: CCGAATGTATCAGGTCAGCTCTT primers . RNA extraction and reverse transcription were performed as described previously [91] with minor modifications . The DNA-free RNA was reverse-transcribed using a RevertAid First Strand cDNA Synthesis Kit ( Fermentas , Helsingborg , Sweden ) according to manufacturer’s recommendations . Aliquots of the generated cDNA were used as template for PCR with gene-specific primers ( S9 Table ) . Quantitative PCR was performed using gene-specific primers ( S9 Table ) and SYBR green ( Fermentas , Helsingborg , Sweden ) on an IQ5 multicolor Real time PCR thermo cycler ( BIO-RAD , PA , USA ) . qPCR reactions were performed in triplicate; gene expression levels were normalized to a PP2A control gene , and results were analyzed as described [92] . Splicing efficiency was measured as described [30] where a primer in an exon was combined with a primer in a neighboring intron ( for the unspliced transcript ) or covering the splicing junction ( for the spliced transcript ) . For the location of primers for measuring COOLAIR splicing see Fig 3A in [30] . RT− controls were always included to confirm absence of genomic DNA contamination . Protein sequences of BRR2a homologues were obtained using PSI-BLAST searches , representative organisms from the different eukaryote kingdoms were selected , and their BRR2 amino acid sequences retrieved from protein databases at NCBI . Amino acid sequences were aligned using ClustalW implemented in MEGA5 [93] . Evolutionary analyses were conducted using MEGA5 , and a bootstrap Neighbor-Joining Tree was calculated for 1000 bootstrap trials . For RNA-seq , RNA was isolated from 15-day-old SD-grown Arabidopsis seedlings harvested at 1 h before darkness using the RNeasy Plant Mini Kit ( Qiagen ) . RNA was treated with DNAse I using TURBO DNA-Free Kit ( Ambion ) and ribosomal RNA was removed using the Ribo-Zero Magnetic kit Plant leaf ( cat# MRZPL116 , EpiCentre ) starting with 1 . 5μg total RNA . Sequencing libraries were generated from the rRNA depleted RNA using the ScriptSeq v2 RNA seq library prep kit ( cat# SSV21124 , EpiCentre ) . Sequencing was performed at an Illumina HiSeq2000 in 100 bp paired-end mode using v3 sequencing chemistry . FastQC v0 . 10 . 1 [94] was used to check read quality followed by removal of 10 bp adapter sequences in all samples with trimmomatic v0 . 32 [95] . Alignments against the Arabidopsis TAIR10 genome were performed with tophat v2 . 0 . 10 [96] using default parameters . De novo transcriptome assembly of mapped samples was performed using cufflinks v2 . 1 . 1 . [97] , the resulting gtf transcriptome files were merged using cuffmerge v2 . 1 . 1 . Splicing events were analyzed with ASTALAVISTA v3 . 0 [49] . RNAseq gene expression counts were generated with HTseq 0 . 6 . 1 [98] using the TAIR10 genome annotation . Before the analysis , samples were subjected to batch effect correction using the R package RUVseq v1 . 0 . 0 [99] with the empirical control option set to 5 , 000 genes . Differential gene expression analyses were performed with the R package DESeq2 v1 . 6 . 1 [51] using thresholds of p = 0 . 05 and fold change = 2 for DEG calling after multiple testing correction according to [100] . Intron counts for the first isoform of Arabidopsis transcripts were generated with HTseq 0 . 6 . 1 . Intron counts were corrected by gene expression using fold change . Differentially retained introns were selected with the DESeq2 package in R using thresholds of p = 0 . 05 and fold change = 2 after multiple testing correction according to [100] . Further analysis was performed with custom scripts in R v3 . 1 . 2 . Gene Ontology analysis was performed using GeneCodis [101] . The hypergeometric statistical test with Bonferroni correction was used with a filter requiring three genes as minimum category population . Data are available at GEO ( accession number GSE65287 ) . Data for chromatin properties were taken from the literature: H1 . 1 and H1 . 2 , [102]; H3K9me2 , [103]; H3K4me1 , [104]; H3K4me2 and H3K36me2 , [105]; H3K4me3 and H3K9ac , [106]; mCG , mCHG and mCHH , [107]; H3K27me3 [108]; H3 and H3K36me3 [109] ( H3K9me2 , CG , CHG and CHH: seedlings; H3K4me1 , H1 . 1 and H1 . 2: 3-weeks-old plants; H3K27me3 , H3K9ac , H3K4me3 , H3K9ac , H3 and H3K36me3: rosette leaves ) . It had been reported that budding yeast Brr2p has a particular role in splicing of highly structured introns [57] . Following the method described in [57] and [110] , secondary structures were predicted for intron sequences between branch site ( BS ) and the 3’ss using RNAfold from the Vienna package [111] . Branch sites were predicted as described in [110] . The free energies of the most stable predicted structure for each intron were compared between DRI and unchanged introns using a Wilcoxon signed-rank test . The difference was not significant ( p>0 . 05 ) .
|
Timing of flowering has a great effect on reproductive success and fitness . It is controlled by many external signals and internal states involving a large set of genes . Here we report that the Arabidopsis thaliana BRR2a gene is needed for normal flowering . BRR2 proteins are components of the spliceosome and highly conserved in eukaryotes . BRR2a is needed for splicing of a subset of introns , most noticeably in the transcript of the flowering repressor FLC . Reduced FLC expression increased transcription of key floral activators , leading to early flowering in both short and long days . Genome-wide experiments established that full BRR2a activity was required only for a small group of introns . We propose that uncompromised BRR2a activity is most important for efficient splicing of a subset of introns of particular size , sequence and chromatin composition , and that FLC is most sensitive to splicing defects .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"sequencing",
"techniques",
"rna",
"extraction",
"mutation",
"epigenetics",
"molecular",
"biology",
"techniques",
"plants",
"flowering",
"plants",
"extraction",
"techniques",
"chromatin",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"genome",
"complexity",
"sequence",
"alignment",
"chromosome",
"biology",
"gene",
"expression",
"molecular",
"biology",
"biochemistry",
"rna",
"spliceosomes",
"cell",
"biology",
"nucleic",
"acids",
"genetics",
"biology",
"and",
"life",
"sciences",
"genomics",
"computational",
"biology",
"introns",
"organisms"
] |
2016
|
BRR2a Affects Flowering Time via FLC Splicing
|
Multi-beam scanning electron microscopy ( mSEM ) enables high-throughput , nano-resolution imaging of macroscopic tissue samples , providing an unprecedented means for structure-function characterization of biological tissues and their cellular inhabitants , seamlessly across multiple length scales . Here we describe computational methods to reconstruct and navigate a multitude of high-resolution mSEM images of the human hip . We calculated cross-correlation shift vectors between overlapping images and used a mass-spring-damper model for optimal global registration . We utilized the Google Maps API to create an interactive map and provide open access to our reconstructed mSEM datasets to both the public and scientific communities via our website www . mechbio . org . The nano- to macro-scale map reveals the tissue’s biological and material constituents . Living inhabitants of the hip bone ( e . g . osteocytes ) are visible in their local extracellular matrix milieu ( comprising collagen and mineral ) and embedded in bone’s structural tissue architecture , i . e . the osteonal structures in which layers of mineralized tissue are organized in lamellae around a central blood vessel . Multi-beam SEM and our presented methodology enable an unprecedented , comprehensive understanding of health and disease from the molecular to organ length scale .
Cells and cellular connectivity undoubtedly play a substantial role in tissue and organ scale behavior , yet mechanisms by which higher order system characteristics emerge from local cellular and molecular phenomena remain a conundrum . Recent advances in coupled , multiscale imaging and modeling of biological systems promise to transform the fields of physiology and medicine [1 , 2] . Within that context , bridging across length scales presents a fundamental challenge [3–5] . Until recently , this challenge was addressed by studying large , complex systems using imaging modalities with increasing resolution , linking the macroscopic and nanoscopic worlds . This approach resulted in single , specific fields of interest with increasing resolution but also with the intrinsic risk of sampling error , which could lead to devastating consequences for a variety of medical applications ( e . g . biopsy ) . The previous lack of a suitable imaging approach that enables a seamless rendering of organ to subcellular structures provided the impetus to apply multi-beam scanning electron microscopy ( mSEM ) , a rapid throughput , high-resolution technology originally developed for quality control of silicon wafers [6] . Electron microscopy ( EM ) can resolve morphological details at the nanometer scale and is commonly used to characterize the structural and functional properties of biomaterials , biological tissues , and their cellular inhabitants [7 , 8] . The acquisition speed of EM , however , limits the capturing of high-resolution images within reasonable time frames and therefore typically is limited to areas within the micrometer range [9] . In contrast , multi-beam scanning electron microscopy ( mSEM ) circumvents typical throughput limitations inherent to conventional single-beam scanning electron microscopes ( sSEM ) [6] . Its novel design enables the analysis of nanoscale morphologies across macroscopic specimens by implementing parallel electron beams and a multi-channel detector [10 , 11] . Multi-beam SEM is capable of reducing acquisition time by more than one order of magnitude and , therefore , of imaging larger surface areas with remarkable resolution [11] , paving the path for seamless multiscale imaging of organ systems down to the cellular and even molecular scale [12] . As a consequence , this technology has drawn interest within the scientific community , particularly in areas related to brain connectomics [13 , 14] and cross-scale musculoskeletal mechanobiology [2 , 10] . Already , a few recent studies in connectomics have utilized mSEM to render volumetric image data from murine specimens and reconstruct neuronal circuits with single-synapse resolution [13 , 14] . Another recent study from our lab demonstrated the feasibility of using high resolution , navigable multiscale maps of human tissue , created for the first time with mSEM , to assess organ- to cell-scale health using epidemiological approaches [12] . Here we describe technical challenges and solutions for the creation of such maps for biomedical applications , using human sample datasets obtained with a mSEM prototype . Current multi-beam SEM acquisitions run at speeds up to 1 . 2 MPixel/s , and typical acquisitions often result in terabyte-size datasets comprised of thousands of individual image tiles that , once combined , form one large complete image ( Fig 1 ) . Initial coordinates of individual image tiles are recorded by the microscope stage . However , the precision of the stage can be worse than that of the nanometer resolution electron microscope . Additionally , the interactions between electron beams and specimen samples , as well as residual illumination aberrations exacerbated when imaging organ samples containing large , dense tissue areas , e . g . bone , affect the relative positions of beams during mSEM imaging . These effects lead to tile misalignment , which can then be addressed by a process commonly referred to as image stitching . Stitching algorithms align and reconstruct sets of overlapping image tiles into seamless photomosaics and are widely used in a variety of fields including microscopy , contemporary digital mapping and panoramic photography . However , upon implementation of stitching and compilation of the overarching image , another challenge remains that is the management , analysis , and dissemination of large data repositories necessary to harness the power intrinsic to mSEM in multiscale characterization of materials . Following the extraction of preliminary images of human hip tissue using an mSEM prototype , we developed a novel computational framework for the reconstruction of mSEM datasets . Stitching was accomplished by combining image pixel-based alignment with global registration accomplished using motion dynamics of mechanical systems . Using the Google Maps API , we created an interactive map of our dataset . Our computational framework potentiates the power of mSEM to enable seamless , multiscale study of organ systems comprised of tissues and their cellular inhabitants . Here we describe this computational framework , its inherent challenges , as well as potential directions for the future .
Human hip bone samples were collected by Dr Ulf Knothe , of the Orthopaedic and Rheumatologic Institute of Cleveland Clinic , in accordance with Institutional Review Board protocol #12–335 . This protocol involved collection of tissues normally discarded in the course of surgery . Due to this and the anonymization of all tissue samples prior to processing for specimen preparation and imaging , as well as later reporting of data , no consent was necessary . Femoral neck tissue samples were acquired from human patients ( age and gender not disclosed ) undergoing hip replacement and prepared according to techniques adapted from atomic force microscopy studies [15] . To facilitate chemical fixation , these samples were sectioned along the coronal and transverse planes , respectively . All specimen acquisitions were completed by the Department of Cleveland Clinic Surgical and Pathology units , per IRB protocol guidelines [16] . Undecalcified tissues were fixed in 2 . 5% glutaraldehyde , 4% formaldehyde , and 0 . 2M cacodylate buffer at 4°C . These tissues were then processed for bulk embedding in poly ( methyl methacrylate ) ( PMMA ) to promote gradual polymerization within a vacuum environment . Upon polymerization of the embedding medium , the specimens were polished , or precision CNC-milled , to achieve mirror-like planarity . Thereafter , samples were prepared for carbon coating and imaging . Selective etching took place , between imaging steps , using 0 . 02M HCl for 90s and/or 10% NaOCl for 11 min , per our previous atomic force microscopy protocols [15 , 17] . This enabled imaging of the respective organic or inorganic phase of the extracellular matrix from correlating tissues of the hip joint complex . One human sample was imaged with a 61-parallel-beam Zeiss MultiSEM 505 prototype , which operates with parallel electron beams arranged hexagonally to minimize electron-optical aberrations [10] , using a landing energy in the range of 1–3 keV , 100 ns of dwell time per beam , and a resolution of 10 nm . A surface area spanning 5 . 7 mm2 was imaged , resulting in 897 hexagonally shaped multi-beam fields of view ( mFOV ) , comprised of nearly 55 thousand high-resolution image tiles and a total of 75 thousand megapixels ( Table 1 ) . Each mFOV was composed of 61 rectangular , single-beam image tiles arranged in a flat , hexagonal pattern ( Fig 1 ) , with a frame size of 1288 x 1120 pixels for each tile . Tile overlap ranged from 2 . 4–55% . The stage used for this study operated with a precision of 2 μm . Image files were stored as bitmap files , accumulating circa 77 GB of storage space . Pixel coordinates for individual single-beam images were available from the microscope metadata , providing a first approximation for relative positioning . A Fourier-based direct alignment algorithm with a simple 2D planar motion model was considered to estimate the translational offsets between overlapping images [18] . Libraries from the registration toolkit TrakEM2 ( Fiji ) [19 , 20] were employed to calculate phase correlations [21] for each pair of overlapping image tiles , extracting a metric of pixel-to-pixel similarities , the correlation coefficients , and translation vectors that maximize pairwise alignment quality . The correlation coefficient and 2D translation vectors for the alignment of tiles i and j are here denoted , respectively , as Rij and pij . Due to the lower quality imaging of specific regions , some alignment parameters ( Rij and pij ) were corrected to reduce the undesirable contribution of image artifacts to overall alignment . Pairings were considered unsatisfactory if they met at least one of the following criteria: ( 1 ) had a correlation coefficient lower than R = 0 . 5 , ( 2 ) had an initial residual length larger than || r || = 300 pixels , ( 3 ) the resulting translation vector was larger than initial overlap dimensions , or ( 4 ) the resulting translation vectors would result in lack of overlap . The alignment parameters of these unsatisfactory pairings were corrected to R = 0 . 5 and p equal to an estimated translation vector . The estimated vector was calculated from the translation vectors with high-quality alignment , considering the type of overlap ( see S1A and S1B Fig ) . High-quality alignments were defined as all pairs ( m , n ) that fulfill Rmn > 0 . 9 ( S1C Fig ) . The estimated value of p of a specific alignment rejected by the aforementioned criteria was calculated as the geometric center of the set of vectors pmn . To globally register the entire dataset , pairwise residual errors were minimized using mass-spring-damper ( MSD ) system dynamics similar to previously proposed elastic registration models [22 , 23] . We chose this approach as opposed to other more common approaches , such as least square ( LS ) minimization , as a way of reducing computer memory requirements while maintaining feasible computation times , which become an issue for large acquisitions such as the one presented here . In a MATLAB simulation , each image tile was assigned a point mass concentrated at the image centroid , with the particles of each overlapping pair of images connected by a spring ( Fig 2A ) . The springs were configured to exert zero restoring force , i . e . to reach their equilibrium length , if the corresponding pair of overlapping images was positioned in a way that maximizes pairwise alignment quality ( Fig 2B ) . Image tiles ( mass particles ) have position , velocity and accumulated force at any given time instant . The dynamics of the system , derived from Newton’s Second Law , with Hooke’s law and a damping force term , were modeled with the following second-order ODE: mx . . +cx . +Fk=0 , ( 1 ) where x is the tile position vector , the over-dot denotes the time derivative , m is the mass of each particle , c is the damping coefficient of particle motion , and Fk corresponds to the net spring forces vector . This term corresponds to the sum of the forces applied by all springs j connected to a tile i Fk ( i ) =∑j kij rij ( 2 ) where kij=k ( Rij ) n ( 3 ) is the spring stiffness and rij is the residual vector between tiles i and j . The string stiffness expression was defined as a power law of R to favor the contribution of regions with high quality images . Its expression is defined by the maximum spring constant k , weighted by the pairwise correlation coefficient Rij , and an arbitrary power n . Pairwise residual vectors ( i . e . the relative positioning error between overlapping images ) can be calculated using the position vectors of tiles ( xi and xj ) and the relative image position vectors that best register the images as follows: rij= ( xi−xj ) −pij ( 4 ) The interconnected system of particles was configured with the mechanical parameters listed in Table 2 , arranged with initial estimated coordinates recorded by the microscope stage , and allowed to come to equilibrium , reaching a lower energy configuration and maximizing alignment ( S1 , S2 and S3 Videos ) . The center tile of an arbitrary mFOV was assumed as an anchor point , and therefore its displacements in both directions for all time instants were considered zero . Differential equations were solved with MATLAB intrinsic function ode45 , which implements an explicit Runge-Kutta method with a variable time step to perform time integration of the initial value problem . Optimal image tile montage was considered to be reached when the change in root-mean-square ( RMS ) of all residuals between consecutive iterations was less than 10−6 pixels . The integration step length was automatically chosen by the solver . All numerical simulations were calculated on an Apple Mac Pro with 3 . 5GHz ( 6 core ) Intel Xeon E5 processor in single-thread mode and 64GB of memory , running OS X 10 . 11 . 1 . Parametric studies ( not included in this manuscript ) found simulations with n = 5 to yield the lowest residual RMS and a damping ratio ζ=c/ ( 2mk ) ≈0 . 14 ( m = 1 kg , c = 0 . 25 Ns/m ) to converge to solution in shorter computational times . We compared our global optimization approach with both unweighted [21] and weighted least squares solutions using a collection of 60 mFOVs ( 3 , 660 tiles ) , excluding areas with widespread image artifacts . Least square solutions were calculated using the lsqnonlin function of Matlab , while the weighted optimization minimized transfer errors multiplied by Rn ( n = 5 , same weight function used in the stiffness of our springs ) . Once fully registered , the stitched dataset was imported into TrakEM2 . Variations in brightness amongst image tiles were minimized using non-linear blending [21] . Geographic information system ( GIS ) frameworks , such as Google Maps , frequently use pre-rendered , multi-resolution sets of images , referred to as a tiled pyramid structure . We adapted a TrakEM2-based CATMAID [23] exporter script [Beanshell script developed by Stephan Saalfield , https://github . com/axtimwalde/fiji-scripts/blob/master/TrakEM2/catmaid-export2 . bsh] to render the tiled pyramid ( S2 Fig ) structure consisting of 11 zoom levels ranging from 0–10 . The final , reconstructed mosaic of all single-beam images was partitioned in PNG-compressed 256 x 256 pixel tiles that collectively constitute the maximum zoom level ( highest resolution ) . Tiles of higher zoom levels were recursively rendered by grouping 'squares' of four tiles ( 512 x 512 pixels ) and downsampling each 'square' to a single , low-resolution 256 x 256 tile , doubling pixel size for each unitary decrement in zoom level . Using the Javascript application programming interface ( API ) of Google Maps , we created a custom map of a human femoral neck region , which was made freely accessible to the public on www . mechbio . org/ploscompbiol . The pre-rendered pyramid tile directories were uploaded to a web server with unique directory paths: ' ( maxzoom—zoom ) /y/y\_x\_ ( maxzoom—zoom ) . png' , where maxzoom represents the maximum zoom level , and x and y are the positions within the tile coordinate system , as specified in the Google Maps API custom map documentation [https://developers . google . com/maps/documentation/javascript/maptypes] . The MapType interface was utilized to create custom maps and specify the translation from screen to tile coordinate frames .
The histogram distribution of correlation coefficient values , calculated for more than 180k pairs of image overlaps , was negatively skewed , with overlapping image pairs concentrated at high correlation values ( close to 28% of all pairwise correlations were below R = 0 . 8 ) , and a median R value of 0 . 89 ( Fig 3A ) . Overlapping image pairs with low R values were spatially concentrated in specific regions ( Fig 3B ) that correspond to the presence of local imaging artifacts observed in the reconstructed dataset after stitching ( Fig 3C ) . Dark regions with disconnected signal patches were evident in the center of the scanned area and along the crack progressing diagonally upwards . These imaging artifacts are manifested as nonexistent , low , or blurred signals ( S3 Fig ) and were attributed to concavities of specimen topography that shielded the secondary electrons , generated in those regions , from the detector . Blurred areas , specifically , correspond to the regions with low R , arranged as vertical columns in Fig 3B . All three approaches ( unweighted and weighted least squares minimization and our relaxation model ) , calculated for a region of 60 mFOVs ( Fig 4A ) , showed similar ability to minimize registration errors , reducing the RMS of residuals by nearly 84% ( Fig 4B ) . Least squares approaches ( unweighted and weighted ) converged within slightly shorter CPU times ( ~48 minutes ) than our MSD approach ( ~54 minutes ) . However , LS approaches required over 10 times more RAM than our MSD approach ( 7270 MB and 710 MB respectively ) . In the complete dataset , consisting of 897 mFOVs ( 54 , 717 tiles ) , the length of residual vectors , || r || , was inversely related to local R values . A clear relationship exists between the spatial distribution patterns of || r || , calculated before alignment parameter correction described in Section 2 . 4 ( Fig 5A ) , and local correlation coefficients , R , shown in Fig 3A . Areas of large residual vector lengths were associated with regions that have poor pairwise alignment . Low R calculations can therefore lead to singularities on the large translation estimates , which are not representative of actual alignment corrections . The vast majority ( 97 . 3% ) of residual vector lengths greater than 300 pixels correspond to pairings with correlation coefficient R < 0 . 7 ( Fig 5B ) . This motivates the correction employed in translation parameters and the expression used for spring stiffness in Eq 3 , which guarantees lower forces applied in tile pairs with low coefficients ( and in most cases very large , nonrepresentative residuals ) when compared to regions with high alignment quality . Our mechanical system-based optimization algorithm was able to reduce the RMS of residuals by 76 . 6% and local || r || values , on average , by 72 . 6% . Fig 5C and 5D , respectively , show the spatial arrangement of || r || prior to and following stitching . Although there was an overall reduction in residual lengths , Fig 5D shows an increase in residual magnitudes across these artifact regions , in response to a lower stiffness of modeled springs assigned in these low R regions . Residual lengths , calculated after correcting alignment parameters prior to ( C ) and following stitching ( D ) , shifted on average from 42 to 8 pixels and overall RMS decreased from 47 to 11 pixels ( Fig 5E ) . Image pairs within the same mFOV had the lowest residual error , while the larger residuals were concentrated in regions of low R , as our algorithm favors alignment in regions with a high R . Global registration used around 1GB of RAM memory . The reconstructed maps reveal a detailed axial view of the composite nature of human cortical bone ( Fig 6 ) . Osteonal structural features are clearly visible with various bundles of blood vessels surrounded by lamellar bone , forming Haversian systems , distributed throughout the cross-section . Acid-etching these samples helps reveal the biological population of the tissue . Bone cells protrude from the mineralized matrix , revealing the morphological characteristics of osteocytes embedded within the mineral bone .
Recent advances in multi-beam SEM enable continuous nanoscale resolution imaging of macroscopic tissue samples and an overall increase in throughput greater than one order of magnitude compared to single beam electron microscopy . Here , we present computational methods to reconstruct a multitude of high-resolution images and develop interactive maps of mSEM datasets , allowing seamless navigation between length scales . The resulting maps reveal , with unprecedented detail across a range of length scales , the biological and material constituents and architecture of human bone . Below , we critically discuss the challenges of the presented methodology , with the goal of facilitating its routine use for structure-function characterization of biological tissues and their cellular inhabitants across length scales . Quantitative assessment is critical for the evaluation of novel imaging techniques . Considering the association between local imaging artifacts and corresponding low pairwise image alignment score , we can assign the correlation coefficient R as an indicator of image quality . Successful ( non-zero ) calculations having an R = 0 . 89 median , suggesting that a simple 2D translational motion model with Fourier-based alignment is an adequate approach to register mSEM images . Overall , low interimage correlations were constrained to the regions of the sample with topographic depressions and cracks . Surface anomalies reflected the challenge of preparing macroscopic hard tissue samples using methods designed for atomic force and electron microscopy , where residual stresses in macroscopic samples are released during sample preparation . Such topographical anomalies promote surface charging of electron beams and therefore compromise secondary electron detection , resulting in a low local signal-to-noise ratio . Low signal regions manifested as locally darkened and/or blurred areas in the raw images ( S3 Fig ) . Blood vessels , due to their smoothness and resulting lack of contrast in acquired images , also hindered pairwise alignment and registered with a low R . Efforts are currently underway to further develop sample preparation methods that avoid formation of such anomalies . Due to the extreme heterogeneity of properties amongst soft , organic inhabitant cells and their hard , inorganic composite environments [8] , bone is one of the most challenging tissues to image across various length scales . Previous studies highlight the specific challenges of sample preparation with regard to etching of such bone composite specimens [15 , 17] . This was the first time bone was tested in such technical context . With increased precision and experience in sample preparation methodologies , however , such surface anomalies can be avoided [10] . Therefore , the presented results should be interpreted with consideration of our sample characteristics , recognizing that the presence of low-quality alignments is markedly induced by local sample conditions . Hence , in context of all tissues that make up the human body , bone provides a robust testbed for the technology . Our results suggest that mass-spring-damper dynamics provide a rational and practical approach to perform global registration of mSEM acquisitions , which typically yield extremely large collections of tiles , reaching a solution at lower computational costs when compared to least squares minimization . Our MSD global registration algorithm corrected for intrinsic tracking discrepancies , requiring lower computational costs and similar final tile coordinates to LS approaches . For the 60 mFOVs highlighted in Fig 4A , the difference in final tile positioning between weighted LSM and the MSD model was on average 5 . 6 pixels . Yet , MSD outperformed LS in both computational time and memory usage . Both LSM and MSD approaches converged to a similar residual error , suggesting that some distortions cannot be overcome with rigid translation models . Future studies will compare , under consideration of computational cost , the accuracy of direct ( pixel-intensity matching ) alignment against other more accurate registration approaches . Feature-based methods , for instance , apply feature extraction ( e . g . MOPS [24] , SIFT [25] ) and global correspondence algorithms to estimate the geometric transformation model . These methods may apply a combination of translational and affine motion models to the dataset , which in the presence of sample-induced artifacts could improve pairwise registration in areas surrounding sample surface imperfections . Non-rigid models that account for aberrations induced by lens effects will also be taken into account in the alignment [26] . Additionally , we will also include a gain compensation step , to reduce intensity variation in overlapping regions , as seen in Fig 6A . Our resulting map shows the composite multiscale architecture of bone , revealing both its structural intricacies and biological milieu . This reconstructed dataset , composed of more than 54 , 000 megapixels , has a considerably large field of view of 5 . 7 mm2 , which compares to some of the largest nanometer resolution , electron microscopy fields of view in current literature [27] . Remarkably , the imaging of our specimen was performed in a practical amount of time , i . e . just under 3 . 5 hours . Furthermore , with the potential to scale mSEM technology according to higher beam counts , timeframe and throughput limitations will eventually become even less of a factor [28] . This combination of high throughput microscopy and image reconstruction with an online geographic information system ( GIS ) tool provides unparalleled , worldwide accessibility of human tissue images to the scientific community and public alike . The Google Maps platform , specifically , is a well-supported , familiar , user-friendly framework that allows for basic navigation of our dataset . Future developments will aim to share our maps in CATMAID [29] , which will allow for collaborative annotation and bookmarking of regions of interest . Top-down and bottom-up approaches look to explore the connectivity of biological and cellular components within certain physiological systems , to further model and analyze the progression of disease throughout the body [4] . The ability to image sub-cellular to tissue-organ scale structures is critical in providing an integrated understanding of physiological mechanisms [2 , 30 , 31] . A variety of different imaging modalities are usually required to bridge the gaps amongst various length scales . Nonetheless , when coupled with the image reconstruction method presented in this paper , multi-beam SEM enables high-resolution characterization of biological tissues across a wide range of length scales . Aside from the modeling and analysis of tissues and organs , biocompatibility at the interface of an implant can also be assessed . Bridging local ( e . g . , how bone cells adapt extracellular matrix to optimize structure for dynamic function ) to systemic perspectives , our method enables comprehensive characterization of multiscale biological phenomena including tissues , cells , and molecules . Such a step forward can be used in health diagnostics and to study health and disease etiology , enabling one to understand how tissue viability and cell connectivity relate to the disruption and failure of tissue and to the pathogenesis of diseases in organs . Already , a few recent studies in connectomics have utilized mSEM to render volumetric image data from murine specimens sectioned with a microtome and reconstruct neuronal circuits with single-synapse resolution [14] . Multiscale imaging of interfaces between musculoskeletal tissue compartments could reveal precise architectures of tight junctions that control functional barrier properties , which exert profound effects on human physiology [1 , 2 , 12 , 32] . In conclusion , our work provides the methodology to create large high-resolution images of biological tissues for structure-function characterization . Combining mSEM methodology with efficient stitching algorithms and GIS applications enables efficient navigation and dissemination of large collections of image data , as shown in this study , delivering a practical approach to assess materials over a wide range of scales . Open access of our reconstructed mSEM datasets , via the Google Maps platform , provides unparalleled , world-wide accessibility of human tissue images to the scientific community and public alike , in a well-supported , familiar , and user-friendly framework . This enabling step leads to a more complete understanding of health and disease , from the length scale of a single cell to the complex system of the human body .
|
Until recently , the assessment of organ and tissue health relied on site-sampling ( biopsy ) of micro-scale regions and was fraught with sampling errors . Overcoming these limitations requires a means for seamless imaging of organs , from their cellular inhabitants to whole organs , akin to charting a map of the organ and its resident cells . Map navigation necessitates the capacity to zoom in and out of regions of interest , with high precision , as well as to analyze relationships between cells , tissue degeneration and organ ( patho- ) physiology . Here we describe the process , in technical detail , based on a world-first case study of a human hip sample and its resident cell population . We acquired 55 , 000 nm-resolution images of the hip using multi-beam scanning electron microscopy ( mSEM ) . To reconstruct the entire dataset , we developed stitching algorithms to maximize map precision at smallest length scales , and rendered them using the Google Maps API . This enabled the exploration of the hip and its inhabitants in a seamless manner , from a global to a high-resolution local view of a single cell . The resulting navigable maps are available for research teams and the public alike to explore and to elucidate the cellular basis of tissue degeneration and organ failure ( mechbio . org ) .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"stiffness",
"mechanical",
"properties",
"medicine",
"and",
"health",
"sciences",
"diagnostic",
"radiology",
"engineering",
"and",
"technology",
"cardiovascular",
"anatomy",
"signal",
"processing",
"osteoblasts",
"microscopy",
"pelvis",
"connective",
"tissue",
"cells",
"materials",
"science",
"research",
"and",
"analysis",
"methods",
"musculoskeletal",
"system",
"imaging",
"techniques",
"animal",
"cells",
"blood",
"vessels",
"scanning",
"electron",
"microscopy",
"hip",
"connective",
"tissue",
"biological",
"tissue",
"bone",
"imaging",
"radiology",
"and",
"imaging",
"diagnostic",
"medicine",
"anatomy",
"cell",
"biology",
"image",
"processing",
"electron",
"microscopy",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"physical",
"sciences",
"material",
"properties",
"osteocytes"
] |
2016
|
Creating High-Resolution Multiscale Maps of Human Tissue Using Multi-beam SEM
|
Covalent modification of DNA distinguishes cellular identities and is crucial for regulating the pluripotency and differentiation of embryonic stem ( ES ) cells . The recent demonstration that 5-methylcytosine ( 5-mC ) may be further modified to 5-hydroxymethylcytosine ( 5-hmC ) in ES cells has revealed a novel regulatory paradigm to modulate the epigenetic landscape of pluripotency . To understand the role of 5-hmC in the epigenomic landscape of pluripotent cells , here we profile the genome-wide 5-hmC distribution and correlate it with the genomic profiles of 11 diverse histone modifications and six transcription factors in human ES cells . By integrating genomic 5-hmC signals with maps of histone enrichment , we link particular pluripotency-associated chromatin contexts with 5-hmC . Intriguingly , through additional correlations with defined chromatin signatures at promoter and enhancer subtypes , we show distinct enrichment of 5-hmC at enhancers marked with H3K4me1 and H3K27ac . These results suggest potential role ( s ) for 5-hmC in the regulation of specific promoters and enhancers . In addition , our results provide a detailed epigenomic map of 5-hmC from which to pursue future functional studies on the diverse regulatory roles associated with 5-hmC .
The potency and fate of a cell can be influenced strongly by the covalent modification of cytosine methylation at carbon five . This critical epigenetic mark influences cellular potency and differentiation by modulating DNA-protein interactions , which direct epigenomic states and transcriptional processes , allowing otherwise common genomes to be expressed as distinct cell types . DNA-methylation-mediated epigenomic processes include dosage compensation , control over aberrant retrotransposon expression , and regulation of centromeric and telomeric heterochromatin [1] . The importance of such processes is exemplified by the essential requirement for DNA methyltransferases ( DNMT1 , DNMT3A , and DNMT3B ) in embryonic and early mammalian development [2] , [3] . Coincident with critical roles for DNA methyltransferases in the regulation of pluripotency , Fe ( II ) /α-ketoglutarate-dependent hydroxylation of 5-mC to 5-hydroxymethylcytosine ( 5-hmC ) by Ten-eleven translocation ( Tet ) family proteins also contributes to the maintenance of pluripotency [4]–[6] . Discovery of this new epigenetic modification raises the possibility that 5-hmC could alter chromatin structure and thereby contribute to gene regulation . Recent functional studies have shown that Tet proteins , particularly Tet1 and Tet2 , are required for ES cell self-renewal and maintenance . However , despite the emergence of these important roles for Tet family proteins , and therefore 5-hmC-associated regulation in ES cells , the genomic- and chromatin-associated contexts of 5-hmC have gone unexplored in human embryonic stem cells . Although there are detailed chromatin state maps of histone modifications in human embryonic stem cells , much less is known about the distinction between 5-hmC and 5-mC localization , largely because of the inability of bisulfite sequencing to resolve the two marks [7] , [8] . Recent studies indicate distinct differences in the presence of stable 5-hmC and Tet1 in mouse ES cells , where strong promoter-proximal Tet1 binding is inversely correlated with the presence of both 5-mC and 5-hmC [9]–[13] , providing putative support for a Tet1-associated demethylation mechanism in the maintenance of unmethylated active promoters . Interestingly , these studies indicate that while Tet1 binding sites are highly enriched at transcription start sites ( TSSs ) in mouse ES cells , a significant fraction of detectable 5-hmC lies within gene bodies and other regulatory regions , which is also consistent with our previous study mapping 5-hmC genome-wide in mouse cerebellum [14] . Furthermore , at regions bound by both Polycomb ( PRC2 ) and Tet1 , the presence of 5-hmC is associated with a repressive state , indicating diverse regulatory roles for 5-hmC that depend at least in part on its chromatin context . Whether localization of 5-hmC with other distinct chromatin signatures results in diverse regulatory mechanisms remains to be explored . To unravel the biology of 5-hmC , we recently developed a selective chemical labeling method for 5-hmC by using T4 bacteriophage ß-glucosyltransferase to transfer an engineered glucose moiety containing an azide group onto the hydroxyl group of 5-hmC , which in turn can chemically incorporate a biotin group for detection , affinity enrichment , and sequencing . Here , to understand the role of 5-hmC in the epigenomic landscape of pluripotent cells , we profiled the genome-wide 5-hmC distribution and correlated it with the genomic profiles of 11 diverse histone modifications and six transcription factors in human ES cells . By integrating genomic 5-hmC signals with maps of histone enrichment , we link particular pluripotency-associated chromatin contexts with 5-hmC . Intriguingly , through additional correlations with defined chromatin signatures at promoter and enhancer subtypes , we found distinct enrichment of 5-hmC at enhancers marked with H3K4me1 and H3K27ac . These results suggest potential role ( s ) for 5-hmC in the regulation of specific promoters and enhancers . In addition , our results provide a detailed epigenomic map of 5-hmC from which to pursue future functional studies on the diverse regulatory roles associated with 5-hmC .
To assess the distribution and general chromatin context of 5-hmC in human embryonic stem ( ES ) cells , we first evaluated the cytogenetic localization of both 5-mC and 5-hmC by immunostaining metaphase chromosomes of human ES cells ( Figure S1 ) . Both 5-mC and 5-hmC were clearly present along the chromosomal arms ( Figure 1A–1D ) ; however , 5-mC displayed a distinctly strong signal at centromeric heterochromatin regions on all metaphase spreads examined ( Figure 1B , n>5 ) . Strikingly , at these same regions , 5-hmC appears completely depleted from 5-mC-enriched pericentromeric regions ( Figure 1A , 1E–1H ) . Given both the defined epigenetic architecture and distinct sequence content of relatively stable centromeric heterochromatic regions , these results may suggest an association of 5-hmC with more epigenetically dynamic loci , such as those throughout chromosome arms , and perhaps exclusion from more epigenetically stable heterochromatin , such as that present in metaphase centromeres . To further evaluate the epigenomic context of 5-hmC , we first established a genome-wide map of 5-hmC in human H1 ES cells by selectively enriching 5-hmC-containing fragments of DNA and subjecting them to high-throughput sequencing . We used a previously established approach to transfer a chemically modified glucose moiety , 6-N3-glucose , onto the hydroxyl group of 5-hmC , which in turns allows cycloaddition of biotin for affinity enrichment and deep sequencing . We prepared and sequenced libraries from 5-hmC-enriched as well as unenriched DNA from the same preparation and sequenced to a depth of >10 million unique , non-duplicate reads per condition . Analyses of chromosome-wide 5-hmC densities showed that , while unenriched input genomic reads were distributed amongst chromosomes close to randomly , as expected by chance , 5-hmC exhibited enrichment or depletion on specific chromosomes ( Figure 2A ) . To further localize regions of 5-hmC enrichment , we identified 5-hmC peaks genome-wide . In total , we identified 82 , 221 regions as significantly enriched for 5-hmC ( p-value threshold of 1e-8 , Table S1 ) . Association of 5-hmC-enriched regions with annotated genomic features indicated significant overrepresentation of 5-hmC within genes and depletion at intergenic regions ( Figure 2B ) , consistent with what has been observed previously in both mouse cerebellum and mouse ES cells [9] , [11]–[14] . Within genes , 5-hmC peaks were particularly enriched in exons ( Figure 2B , 6 . 14-fold over expected based on the genomic coverage of these regions ) , whereas we saw much lower frequency within intronic regions ( Figure 2B , 1 . 33-fold over expected ) , which is likely a result of the increased GC content within exons relative to introns . 5-hmC peaks were also significantly enriched within intragenic CpG islands ( CGIs ) ( 17 . 6-fold over expected ) and are more frequent than expected by chance at intergenic CGIs ( Figure 2B ) . Interestingly , we find significantly more 5-hmC peaks overlapping predicted enhancers than was expected ( 8 . 6-fold over expected , Figure 2B ) . These results indicate that in addition to gene body-associated regulatory roles , 5-hmC may also function within other genomic regions important for gene modulation . We also assessed the general sequence content of these peaks , including GC content and dinucleotide frequencies . We found that the frequency of CpG dinucleotides within 5-hmC-enriched regions was no greater than randomly chosen regions of the genome and significantly lower than CGIs , whereas CA , CC , and CT dinucleotides each exhibited an O/E >1 and enrichment relative to random genomic locations ( Figure 2C ) . Furthermore , GC content as a whole was significantly reduced compared with CGIs , and slightly increased relative to random genomic loci ( Figure 2D ) . These data suggest that 5-hmC-enriched loci occur most often in regions of the genome with moderate GC content and that it occurs less frequently within a high density of CpGs . In order to determine the specific chromatin contexts associated with 5-hmC in human embryonic stem cells , we obtained sequence data derived from immunoprecipitation of 5-mC ( MeDIP ) ( GSM456941 ) and 11 diverse histone modifications in H1 hES cells [15] . MeDIP , histone-ChIP , and unenriched input reads derived from the same experiments were binned genome-wide at 1 , 5 , and 10 kb . MeDIP and histone-specific signals were normalized to input values ( ChIP-Input ) . 5-hmC-enriched reads were binned genome-wide using identical parameters . Input-normalized 5-hmC signals were then subsequently correlated with input-normalized histone modification and 5-mC MeDIP values within the same genomic bin for all bins genome-wide in order to generalize the relative correlation between 5-hmC , 5-mC , and diverse histone modifications ( Figure 3 ) . We found that data binned at various sized intervals exhibited generally similar patterns on a genomic scale when comparing the relative correlations between 5-hmC and the various histone modifications tested . We find that in general , on a genomic scale , 5-hmC and 5-mC detected by MeDIP correlate better than any histone-specific mark tested ( Figure 3A , r2 = 0 . 448 ) , consistent with the fact that 5-hmC is derived from 5-mC and with previous reports showing a significant amount of overlap between the two marks in mouse ES cell genomes [9] , [11] , [13] . Although it is difficult to assess the ratio of 5-mC:5-hmC from genome-wide bisulfite sequencing data ( Methyl-Seq ) , we also determined the association between 5-hmC and 5-mC+5-hmC detected by Methyl-Seq ( Figure S2 ) . Within the CG context , regions with higher 5-hmC also tend to have a higher percentage of 5-mC+5-hmC , as would be expected . However , there are also a large number of regions with a high percentage of 5-mC+5-hmC that contain very low levels of 5-hmC and are therefore presumably dominated by 5-mC ( Figure S2A ) . These results are again consistent with the notion that 5-hmC is derived from 5-mC . We also compared 5-hmC signals to 5-mC+5-hmC within the non-CpG context , which occurs in human ES cells more frequently than in differentiated cell types [16] . 5-hmC has been reported to occur within non-CpG contexts in mouse ES cells as well [11] . Our analyses indicate that regions containing high levels of 5-hmC tend to harbor less non-CpG methylation ( Figure S2B and S2C ) . However , due to the low percentage of both CHG and CHH methylation throughout the genome , it is difficult to resolve the extent to which 5-hmC may occur at non-CpG sites and analyses do not exclude the possibility that 5-hmC occurs within a non-CpG context in human ES cells . Further resolution of single base pair 5-hmC will be required to conclusively establish the sequence contexts of hydroxymethylated cytosines . Correlations between 5-hmC and the 11 histone modifications tested were largely , with a few notable exceptions , in agreement with the previously observed associations between histone modifications and the percentage of overall DNA methylation ( 5-mC+5-hmC ) assessed by Methyl-Seq [15] . Consistent with the correlations between Methyl-Seq and histone modifications , we find a relatively strong association between 5-hmC and H3K4me1 ( r2 = 0 . 293 ) and H3K4me2 ( r2 = 0 . 152 ) compared with H3K4me3 ( r2 = 0 . 0518 ) ( Figure 3B–3D ) . The relatively strong correlations between 5-hmC , H3K4me1 , and H3K4me2 compared to H3K4me3 are also consistent with earlier observations showing enrichment of 5-hmC within active gene bodies , but depletion at TSSs . We also saw a relatively strong correlation between H3K18ac , a mark that directly regulated CBP/p300 enhancer complexes with transcriptional activation [17] , [18] , and 5-hmC ( Figure 3G , r2 = 0 . 324 ) . A significantly smaller albeit moderate correlation was found between 5-hmC and H3K27ac , H3K27me3 ( with H3K27ac > H3K27me3 ) , and H4K5ac ( Figure 3H , 3I and 3L ) . Both H3K9ac and H3K9me3 exhibited relatively low levels of correlation with 5-hmC ( Figure 3E and 3F ) . Surprisingly , we see a relatively weak correlation between 5-hmC and H3K36me3 ( Figure 3J ) . H3K36me3 is known to correlate well with gene expression levels and has been linked to transcriptional elongation in hES cells [19] , but is largely absent from TSSs . H3K36me3 is also one of the few histone marks for which there is a strong correlation with methylated DNA , as detected by bisulfite sequencing [15] . These results suggest the possible enrichment of H3K36me3 or 5-hmC on distinct groups of gene bodies in hES cells , which could depend on the level of gene expression . Together , the correlations between 5-hmC , 5-mC , and the 11 specific histone modifications tested indicate that , in addition to being generally associated with more euchromatic accessible chromatin , 5-hmC may be linked to diverse gene regulatory elements and transcriptional regulatory processes in human ES cells . Both cytogenetic localization of 5-hmC and genome-wide correlations with 11 diverse histone modifications indicate links between 5-hmC , more accessible euchromatic chromatin , and gene regulation . To test the dependence of gene-associated 5-hmC distributions on expression levels in human ES cells , we measured 5-hmC signals at genes with varying expression as measured by RNA-Seq RPKM [16] . Overall , 5-hmC displays a strong promoter-proximal bias in hES cells , while also being enriched within gene bodies , albeit to a lesser degree relative to the TSS ( Figure 4A–4E ) . Interestingly , we observed a distinct forking in the 5-hmC distribution around the TSS as expression levels rose , ultimately transitioning to a bimodal distribution at more highly expressed genes compared with genes expressed at lower levels ( Figure 4A–4E ) . However , the correlation between 5-hmC and both TSSs and gene bodies is not strictly linear . 5-hmC tends to be higher , both within the gene body and at the TSS , at genes expressed within the 25–75% range of all genes based on RNA-Seq RPKM ( Figure 4C and 4D ) , compared to the top 25% of expressed genes ( Figure 4A ) . Meanwhile , at genes within the bottom 25% , 5-hmC is mainly enriched directly over the TSS and only moderately enriched within the gene body . Thus , at genes exhibiting lower expression , 5-hmC is present directly at the TSS ( Figure 4E ) , whereas genes with intermediate expression display higher gene body 5-hmC and a distinct bimodal distribution ( Figure 4A ) at the TSS . At the most highly expressed genes , 5-hmC exhibits a similar distribution to that seen on intermediately expressed genes , but overall lower levels at both the TSS and gene body . These results are consistent with the observed dual function of 5-hmC in mouse ES cells , where the Polycomb complex PRC2 may act in combination with Tet1 to influence the distribution of 5-hmC at repressed genes , while at more highly expressed genes the presence of Tet1 , without PRC2 , results in loss of 5-hmC at the TSS and establishment of a bimodal distribution [9] , [10] . To further explore the enrichment of 5-hmC at gene bodies with intermediate levels of expression , we directly compared the distribution of 5-hmC to that of H3K36me3 in and around genes ranked by expression level ( Figure 4A ) . H3K36me3 is an intragenically enriched histone modification that also correlates well with gene expression levels [19] . We found that genes with the highest intragenic 5-hmC also had relatively low intragenic H3K36me3 ( Figure 4A ) , consistent with the relatively low genome-wide correlations between binned 5-hmC and H3K36me3 ( Figure 3J ) . The same genes were also those expressed at intermediate levels ( 25–75% range based on RNA-Seq RPKM ) . At the top 25% of expressed genes , H3K36me3 is highly enriched within gene bodies and transcription end sites ( TES ) , while 5-hmC tends to be lower at both TSSs and gene bodies compared to genes expressed at intermediate levels ( Figure 4A ) . These data suggest a complex relationship between 5-hmC , H3K36me3 , and gene expression levels in human ES cells . One possible explanation could be that 5-hmC functions to temper transcription at the genes that are not fully committed to a constitutive expression state . At genes expressed at the lowest levels , 5-hmC may play a role at the TSS to represses full-length transcription , while still maintaining the transcriptional potential of the marked genes . Such a role is consistent with the previously reported interaction between TSS 5-hmC and repression by Polycomb group complexes , which repress many developmentally regulated genes in ES cells [9] , [10] . At the genes with intermediate expression levels , 5-hmC may temper expression at both TSS and gene body . At genes with the highest expression , TSS- and gene body-associated 5-hmC may be , at least in part , replaced by H3K36me3 to allow full transcriptional potential . We note that such distributions of 5-hmC in ES cells is distinct from that observed in mouse brain , where 5-hmC is largely depleted from TSSs , enriched within gene bodies , and correlates well with gene expression levels ( Szulwach and Jin , unpublished observations and [14] ) . These differences may reflect stem cell-specific and brain-specific roles for 5-hmC-mediated gene regulation . Such differences may be accounted for by the relative enrichment of Tet1 in ES cells and/or yet-to-be-identified Tet-family co-factors compared to more differentiated cell types . In promoter-proximal regions of embryonic stem cells , 5-hmC exhibits a TSS-associated bias that is dependent on gene expression level ( Figure 4 ) . To further understand the relevance of this bias in terms of chromatin context , we examined the distribution of 5-hmC around 18 distinct promoter subtypes defined on the basis of their chromatin signatures [15] . Among 11 promoter subtypes with significant enrichment of the histone modifications tested in H1 hES cells , we found that 5-hmC distributions within the same regions could be classified into two groups . The first group reflected the distribution of 5-hmC at more highly expressed genes , with 5-hmC displaying a marked depletion directly over the TSS and a bimodal distribution around the TSS ( Figure 5A , 5B ) . This distribution corresponded to a strong H3K4me3 signal , consistent with an inverse correlation between 5-hmC and H3K4me3 ( Figure 5A–5C ) . Flanking the region of depletion were two peaks of 5-hmC , which overlapped with regions of H4K4me1 and H3K4me2 enrichment . A clear example of this could be seen at the well-characterized promoters of the DNMT3A locus , itself a highly expressed gene in ES cells ( Figure 5C ) . The bimodal distribution of 5-hmC , H4K4me1 , and H3K4me2 around TSSs might reflect paused promoters , at which divergent RNAPII is known to display pausing , and could suggest an influence of 5-hmC on transcription pausing at such promoters in hES cells . The second group of promoters displayed lower 5-hmC signal overall , but a more even distribution over the promoter regions , without a distinct region of depletion ( Figure 5D–5F ) , and reflected the distribution of 5-hmC at genes expressed at intermediate or low levels ( Figure 4 ) . Again , the distribution of 5-hmC correlated well with the presence of H3K4me1 and H3K4me2 , while H3K4me3 was also present ( Figure 5E and 5F ) . We also noted that this group of promoters displayed an overall weaker signal in each histone modification tested , relative to promoters exhibiting bimodal distributions of both 5-hmC and various histone modifications ( Figure 5B and 5E ) , which likely represents the expression status of this group of genes . Assessment of 5-hmC at an additional seven promoter types , which displayed low levels of modified histone enrichment in H1 hES cells , also displayed low levels of 5-hmC ( Figure S3 ) and less distinct distribution patterns , consistent with a link between defined histone modifications and 5-hmC at TSSs . Association of 5-hmC-enriched regions with annotated genomic features suggested that , in addition to playing important roles within gene bodies and gene proximal regions , 5-hmC might also function at distinct regulatory elements , including enhancers ( Figure 2B ) . To address the potential role of 5-hmC at enhancers as well as the distinct chromatin contexts associated with each , we determined the distribution of 5-hmC at 12 different sets of predicted enhancers defined on the basis of chromatin signature [15] . Strikingly , we found that 5-hmC marked each of five enhancer subtypes displaying enrichment of H3K4me1 , H3K18ac , H4K5ac , and H3K27ac in H1 hES cells , while enhancer subtypes exhibiting less enrichment of these marks also tended to be less enriched for 5-hmC ( Figure 6A and 6B ) . A clear example of a 5-hmC-associated enhancer occurred upstream of the ES-specific gene PRDM14 , where a 5-hmC peak was identified as directly overlapping an E8 type enhancer ( Figure 6C ) . PRDM14 has been reported as an integral factor contributing to pluripotency via interactions with the core transcriptional circuitry in ES cells [20] , [21] . This may suggest a functional role for 5-hmC , in combination with at least H3K4me1 , at this upstream enhancer in maintaining expression of PRDM14 and contributing to the pluripotency of human ES cells . In combination with the general enrichment of 5-hmC peaks at predicted hES cell enhancers ( Figure 2B ) , these data demonstrate distinct marking of ES cell enhancers with 5-hmC and defined chromatin signatures . We further tested the distribution of 5-hmC around a set of 12 ChIP-rich regions that were previously identified as exhibiting enrichment of specific histone modifications , but that lay outside of defined promoters or predicted enhancer regions ( Figure S4 ) [15] . In general , 5-hmC signals were significantly lower at such regions , and few patterns were apparent . However , we did find that ChIP-rich regions with H3K36me3 displayed markedly lower levels of 5-hmC and that regions enriched for K3K9me3 actually exhibited depletion of 5-hmC ( Figure S2 ) , consistent with the lower genome-wide correlations we found between 5-hmC and these two histone modifications ( Figure 4E and 4I ) . DNA methylation has been implicated in regulating transcription factor binding dynamics and has been found to differentially mark sites of core pluripotency-associated transcription factors in ES cells [16] . We therefore asked whether or not 5-hmC marked sites bound by six transcription factors mapped genome-wide in H1 hES cells , including the pluripotency-associated transcription factors NANOG , OCT4 , and SOX2 , as well as more general factors , such as p300 and TAF1 ( Figure 7 ) . At sites of all types we could detect a slight enrichment of 5-hmC and direct overlap between subsets of 5-hmC peaks and transcription factor binding sites , consistent with previous observations in mouse ES cells detecting 5-hmC at transcription factor binding sites [9] , [11] , [13] . However , signals varied across factors . Among pluripotency-related factors , we find distinct marking and enrichment of 5-hmC at of only NANOG sites ( Figure 7A ) . An example of 5-hmC enrichment at a NANOG binding site was seen directly upstream of DNMT3B ( Figure 7B ) , a gene expressed a high levels in ES cells . Consistent with a lack of 5-hmC at many TSSs , we also observe depletion of 5-hmC at TAF1 interaction sites ( Figure 7A ) . Although we observed good correlation between histone modifications demarcating enhancers and enrichment of 5-hmC at specific subtypes of enhancers defined by chromatin signature , we did not observe distinct 5-hmC marking at p300 sites ( Figure 7A ) . We further addressed this by asking what the overlap was between the 82 , 221 identified 5-hmC enriched regions ( Table S1 ) , predicted enhancers [15] and p300 sites [16] . As expected a large proportion of p300 sites ( 1795 of 3094 , 58% ) overlap predicted enhancers ( Figure 7C ) . However , the fraction of predicted enhancers explained by p300 binding remained quite low ( 1 , 795 of 58 , 023 , 3 . 1% ) , suggesting a significant amount of enhancer regulation by p300 independent mechanisms . Interestingly , we find that while only a small fraction of p300 sites ( 166 of 3094 , 5 . 4% ) overlap 5-hmC enriched regions , a significant percentage of predicted enhancers ( 19 , 973 of 58 , 023 , 34 . 4% ) overlap with 5-hmC enriched regions ( Figure 7C ) . Furthermore , sites that were enriched for 5-hmC , bound by p300 , and predicted as enhancers were quite rare , occurring only 25 times . These data suggest that significant portion of predicted enhancers are also enriched in 5-hmC , but lack p300 binding , and may indicate a role for 5-hmC in regulating p300 independent enhancers . Together these results indicate that 5-hmC may also influence the chromatin states at protein-DNA interaction sites , thereby modulating the function of key transcription factors and diverse enhancer subtypes .
Recent studies have shown that Tet family proteins can catalyze 5-methylcytosince ( 5-mC ) conversion to 5-hydroxymethylcytosine ( 5-hmC ) and play important roles in self-renewal and cell lineage specification in embryonic stem ( ES ) cells [4]–[6] , [11] , [22] . These findings suggest a potential role for 5-hmC-mediated epigenetic regulation in modulating the pluripotency of ES cells . To unveil this new regulatory paradigm in human ES cells , here we used a selective 5-hmC chemical labeling approach coupled with affinity purification and deep sequencing that we developed before to establish the genome-wide distribution of 5-hmC in human ES cells . Integration of 5-hmC distributions with genome-wide histone profiles led us to identify the pluripotency-linked chromatin contexts associated with 5-hmC . Through association with genomic features defined on the basis of chromatin signatures , we find 5-hmC-mediated marking of not only specific promoters and gene bodies , but also distinct enhancer subtypes , including those marked with H3K4me1 and H3K27Ac . Lastly , we find 5-hmC is associated with the binding sites of specific core pluripotency transcription factors and a lack of 5-hmC at others . Our results suggest that 5-hmC is an important epigenetic modification associated with the pluripotent state that could play role ( s ) in a subset of promoters and enhancers with defined chromatin signatures in ES cells . By correlating genome-wide distributions of 5-hmC with those of 11 diverse histone marks , we found that 5-hmC displayed relatively strong correlations with H3K4me1 and H3K4me2 versus H3K4me3 , which , as expected , is consistent with previous correlations between DNA methylation detected by Methyl-Seq and histone modifications [15] . 5-hmC also exhibited a strong correlation with H3K18ac , a mark regulated by CBP/p300 at enhancers that is associated with transcriptional activation . We also found more modest correlations with H3K27ac , H3K27me3 , and H4K5ac , and very low correlations with H3K9ac and H3K9me3 . However , our data suggested that 5-hmC was not strongly correlated with H3K36me3 , a histone modification previously linked to DNA methylation detected by Methyl-Seq . This intriguing difference suggested differential marking of gene bodies by 5-hmC and H3K36me3 in pluripotent cells . Direct comparisons of genic 5-hmC and H3K36me3 indeed revealed that genes with the highest levels of TSS and gene body 5-hmC tend to exhibit intermediate levels of expression and harbor less intragenic H3K36me3 , compared to genes with the highest levels of expression . Although a number of intriguing explanations might account for these observations , one possibility is that 5-hmC may function to temper transcription at both the TSS and gene body of intermediately expressed genes , while maintaining their potential to be more fully expressed when needed . Upon full activation , 5-hmC may be at least partially removed as the transcriptional unit acquires H3K36me3 and commits to a more fully active state . Restriction of 5-hmC at the TSS of repressed genes and its presence at both TSSs and gene bodies of intermediately expressed genes may also indicate distinct regulation of 5-hmC at these locations . At TSSs of genes that are repressed or expressed at low levels , Polycomb group complex , PRC2 , may interact with 5-hmC to repress but maintain the potential for expression of targeted genes , as has been previously suggested [9] , [10] . However , such distributions are distinct from those observed in mouse cerebellum [14] , where 5-hmC is significantly enriched compared to ES cells , largely absent from TSSs , and high within gene-bodies , positively correlating gene-expression . Thus , distinction of mechanisms differentially influencing the state and regulation of 5-hmC within genes bodies in the context of gene expression outcomes will be important towards understanding the role of 5-hmC in both brain and ES cells . Our genome-wide analyses of 5-hmC also revealed a general promoter-proximal bias of 5-hmC around RefSeq transcripts in human ES cells , which is consistent with the recently published work on mapping 5-hmC in mouse ES cells [9] , [11]–[13] . This TSS-associated bias was also dependent on gene expression levels , with 5-hmC transitioning from a position directly over the TSS at repressed genes to a bimodal distribution at more highly expressed genes , likely reflecting the observed dual function of 5-hmC in mouse ES cells [9]–[13] , although this correlation was not strictly linear . Interestingly , we find that the bimodal distribution of 5-hmC is also strongly correlated with the distributions of H3K4me1 and H3K4me2 , but inversely correlated with H3K4me3 . The bimodal distribution of 5-hmC , H4K4me1 , and H3K4me2 around TSSs might reflect the establishment of divergent paused RNAPII , which is known to play a critical regulatory role at developmentally regulated transcripts in ES cells [23] , [24] . This could thereby point to an influence of 5-hmC on transcription pausing at such promoters in hES cells . We also noted that such a promoter-proximal bias of 5-hmC in ES cells is distinct from that observed in mouse brain , where 5-hmC is largely depleted from TSSs and enriched within gene bodies ( Szulwach and Jin , unpublished observations and [14] ) , where it also correlates well with gene expression . This could suggest that such a bias reflects a stem cell-specific role for 5-hmC-mediated gene regulation at and around certain TSSs . Such differences may be accounted for by the enrichment of Tet1 , or yet-to-be-identified co-factors of Tet1 , in ES cells relative to more differentiated cell types . Analyses of 5-hmC-enriched peaks and their correlation with enhancer-associated specific histone modifications , such as H3K4me1 , H3K18ac , and H3K27ac , suggested that , in addition to being present at promoters , 5-hmC could also mark other diverse regulatory elements in the genome , such as enhancers . Interestingly , assessment of 5-hmC distributions at the predicted enhancers in H1 hES cells demonstrated the enrichment of the epigenetic mark at specific enhancer subtypes , including those enriched for K3K4me1 , H3K27ac , H3K18ac , and H4K5ac . Despite a good correlation between 5-hmC and histone marks demarcating enhancers , we found that only small fraction of regions bound by p300 were also enriched for 5-hmC . Finally , we examined the correlation of 5-hmC distributions with the genome-wide binding sites of six transcription factors that have been linked to maintaining the pluripotency of ES cells [16] . We find that 5-hmC can also mark NANOG binding sites , while being depleted at TAF1 sites . These results further suggest diverse roles for 5-hmC in regulating the accessibility of transcription factors in defined chromatin contexts , including those regulating pluripotency in ES cells . In summary , here we present the genome-wide distribution of 5-hmC and its correlation with 11 diverse histone modifications and six transcription factors in human ES cells . By integrating genomic 5-hmC signals with maps of different histone marks , we link particular pluripotency-associated chromatin contexts with 5-hmC . Our study suggests that 5-hmC could play diverse roles in regulating specific promoters , gene bodies , and enhancers in ES cells , thereby providing a detailed epigenomic map of 5-hmC from which to study its contribution to pluripotency .
H1 human ES cells were maintained on mitomycin C-treated STO cells in ES medium consisting of DMEM/F12 medium ( Invitrogen ) supplemented with 20% serum replacement ( SR; Invitrogen ) , 1 mM L-glutamine ( Invitrogen ) , 100 µM nonessential amino acids ( Invitrogen ) , 0 . 1 mM ß-mercaptoethanol ( Sigma ) , 1X Antibiotics-Antimycotic ( Invitrogen ) , and 4 ng/mL bFGF ( Invitrogen ) . The fully grown H1 cells were mechanically isolated and transferred into a prepared dish with fresh feeder cells . Prior to the isolation of genomic DNA , cells were treated with dispase ( 2 mg/ml in DMEM/F12 ) to detach human ES cells from feeder cells . Metaphase chromosomes were prepared by standard protocols as described previously [25] . The slides with hES metaphase chromosome spreads were washed with PBS for 5 min . The slides were immersed in 1N HCl and incubated at 37°C for 30 min . After HCl treatment , the slides were washed with PBS for 15 min followed by blocking with 3% goat serum/0 . 4 Triton X-100 in PBS for 1 h . The samples were incubated with primary antibodies at 4°C overnight . The following primary antibodies were used: rabbit anti 5-hydroxymehtylcytosine ( 1∶10 , 000 , #39769 , Active Motif ) , mouse anti-5-methylcytosine ( 1∶1000 , Eurogentec , BI-MECY-0100 ) . On the second day , the slides were washed with PBS and then incubated with secondary antibodies: goat anti-rabbit Alexa488 ( 1∶500 , #A11008 , Invitrogen ) and goat anti-mouse Alexa568 ( 1∶500 , #A11031 , Invitrogen ) . The slides were counter-stained with the fluorescent nuclear dye 4′ , 6-diamidino-2-phenylindole ( #B2261 , Sigma ) . The slides were examined using a Zeiss AX10 microscope , and images were processed with Photoshop software . More than 5 metaphase spreads were examined for 5-mC and 5-hmC . Genomic DNA was isolated by cell lysis in digestion buffer ( 100 mM Tris-HCl , pH 8 . 5 , 5 mM EDTA , 0 . 2% SDS , 200 mM NaCl ) , Proteinase K treatment ( 0 . 667 ug/ul , 55°C overnight ) . The second day , an equal volume of Phenol:Chloroform:Isoamyl Alcohol ( 25∶24∶1 Saturated with 10 mM Tris , pH 8 . 0 , 1 mM EDTA ) ( P-3803 , Sigma ) was added to samples , mixed completely , and centrifuged for 5 min at 14 , 000 rpm . The aqueous layer solution was transferred into a new Eppendorf tube and precipitated with 2 volumes 100% ethanol and 1/10 volume 3 M NaOAc . The genomic DNA was recovered and dissolved with 10 mM Tris-HCl , pH 8 . 0 . Genomic DNA samples were further sonicated into ∼500 bp by Misonix sonicator 3000 ( using microtip , 4 pulses of 27 s each , with 1 min of rest and a power output level of 2; the sonication was performed always on ice ) . The fragment size of sonicated DNA was verified by agarose gel electrophoresis . The DNA concentration was determined with NANO-DROP 1000 ( Thermo Scientific ) . The dot blot was performed on a Bio-Dot Apparatus ( #170-6545 , BIO-RAD ) . Briefly , the serially diluted C , 5-mC , or 5-hmC only standard DNA samples ( Zymo research ) were mixed with 2N NaOH and 10 mM Tris·Cl , pH 8 . 5 , and loaded onto 6X SSC rinsed Hybond-N+ membrane ( Amersham Biosciences , #RPN303B ) . The completely dried membrane was baked for 30 min at 80°C and then blocked with PBS containing 5% dry milk and 0 . 1% Triton X-100 for 1 h at room temperature . The primary rabbit anti-5-hydroxymethylcytosine antibody ( 1∶10 , 000 , #39769 , Active Motif ) or ( 1∶1 , 000 , mouse monoclonal anti-5-methylcytosine , BI-MECY-0100 , Anaspec ) was applied to the membrane and incubated overnight at 4°C . The second day , the membrane was rinsed with PBS and the signal was developed after incubation with HRP-conjugated secondary antibody for 30 min . 5-hmC enrichment was performed using a previously described procedure with an improved selective chemical labelling method [14] . Briefly , the 5-hmC labelling reactions were performed in a 100-µL solution containing 50 mM HEPES buffer ( pH 7 . 9 ) , 25 mM MgCl2 , 300 ng/µL sonicated genomic DNA ( 100–500 bp ) , 250 µM UDP-6-N3-Glu , and 2 . 25 µM wild-type β-GT . The reactions were incubated for 1 h at 37°C . After the reaction , the DNA substrates were purified via Qiagen DNA purification kit or by phenol-chloroform precipitation and reconstituted in H2O . The click chemistry was performed with the addition of 150 µM dibenzocyclooctyne-modified biotin into the DNA solution , and the reaction mixture was incubated for 2 h at 37°C . The DNA samples were then purified by Pierce Monomeric Avidin Kit ( Thermo ) following the manufacturer's recommendations . After elution , the biotin-5-N3-gmC-containing DNA was concentrated by 10 K Amicon Ultra-0 . 5 mL Centrifugal Filters ( Millipore ) and purified by Qiagen DNA purification kit . DNA libraries were generated following the Illumina protocol for “Preparing Samples for ChIP Sequencing of DNA” ( Part# 111257047 Rev . A ) . We used 25 ng of input genomic DNA or 5-hmC-captured DNA to initiate the protocol . DNA fragments of ∼150–300 bp were gel-purified after the adapter ligation step . PCR-amplified DNA libraries were quantified on an Agilent 2100 Bioanalyzer and diluted to 6-8 pM for cluster generation and sequencing . We performed 38-cycle single-end sequencing using Version 4 Cluster Generation and Sequencing Kits ( Part #15002739 and #15005236 respectively ) and Version 7 . 0 recipes . Image processing and sequence extraction were done using the standard Illumina Pipeline . FASTQ sequence files were aligned to the Human reference ( NCBI36 , hg18 ) using Bowtie 0 . 12 . 6 , retaining only unique , non-duplicate genomic matches with no more than 2 mismatches within the first 25 bp . Unique , non-duplicate reads from non-enriched input genomic DNA and each 5-hmC-enriched sequence set were counted in 1000- , 5000- , and 10 , 000-bp bins genome-wide and subsequently normalized to the total number of non-duplicate reads in millions . We find that bins of varying size produce largely similar patterns genome wide and have reported values within a bin size of 10 kb within all figures . Input-normalized values were subtracted from 5-hmC-enriched values per bin to generate normalized 5-hmC signals . Summary of sequence output: H1 , 5-hmC enriched = 10038770 non-duplicate reads , H1 , Unenriched input = 20656172 non-duplicate reads . Chromosome-wide densities were determined as reads per chromosome divided by the total number of reads in millions . Expected values were determined by dividing 106 by the total NCBI36/hg18 length , and multiplying by chromosomal length . Expected values were divided by 2 for chromosomes X and Y . For MeDIP/histone modification correlations , unique , non-duplicate reads from non-enriched input genomic DNA and 5-hmC-enriched DNA were counted in 10 , 000-bp bins genome-wide and subsequently normalized to the total number of non-duplicate reads in millions . Input-normalized values were subtracted from 5-hmC/histone-enriched values per bin to generate normalized 5-hmC/histone signals . All histone ChIP-Seq data was acquired from Sequence Read Archive ( SRA ) , accession SRP000941 , [15] . MeDIP data were obtained from NCBI GEO Accession GSM456941 . All histone ChIP-Seq data and MeDIP were mapped and processed with the identical parameters used for 5-hmC reads described above . 5-hmC peaks were identified using MACS [26] with the following parameters: effective genome size = 2 . 7e+09; Tag size = 38; Bandwidth = 200; P-value cutoff = 1 . 00e-08; ranges for calculating regional lambda are: peak_region , 200 , 1000 . Association of 5-hmC peaks with genomic features was performed by overlapping peak locations with known genomic features obtained from UCSC Tables for NCBI36/hg18: RefSeq Whole Gene , 5′UTR , Exon , Intron , 3′UTR , +/−500 bp of RefSeq TSS , RefSeq Intergenic ( complement of Whole Gene ) , CpG Islands ( +/−2 kb of CGI , Intergenic/Intragenic/TSS based on RefSeq Whole Gene ) . Predicted enhancer locations were obtained from [15] . Peaks were assigned to a given genomic feature if overlapping ≥1 bp . Expected values were determined based on the percent base coverage of each defined genomic feature in NCBI36/hg18 . All histone ChIP-Seq data were acquired from Sequence Read Archive ( SRA ) , accession SRP000941 . All histone ChIP-Seq data were mapped and processed with the identical parameters used for 5-hmC reads described above . Chromatin signatures for promoters , enhancers , and ChIP-rich regions were acquired from [15] . 5-hmC reads were counted in 100-bp bins , in the 4 kb directly surrounding identified binding sites , as well as 4 kb upstream and downstream of the immediate 4-kb region . Read counts were normalized to the total number of aligned reads in millions and input reads counted and normalized in the same manner were subsequently subtracted to determine 5-hmC enrichment . RNA-Seq RPKM values and transcription factor binding sites for KLF4 , NANOG , OCT4 , p300 , SOX2 , and TAF1 in H1 ES cells were described previously [16] . For correlations between 5-hmC and gene expression , 5-hmC reads were counted in 100-bp bins , in the ±5 kb directly surrounding TSSs and TESs . Read counts were normalized to the total number of aligned reads in millions . For transcription factor binding sites , 5-hmC reads were counted in 40 equally sized portions within , upstream , and downstream of the binding sites . Read counts were normalized to the total number of aligned reads in millions and input reads counted and normalized in the same manner were subtracted to determine 5-hmC enrichment . Methyl cytosine counts in the CG , CHG , or CHH context were obtained directly from [16] and the percent methylation in each 10 kb bin genome-wide was determined as the weighted sum of methylated cytosine detected at each position .
|
Recent studies revealed the oxygenase-catalyzed production of 5-hydroxymethylcytosine ( 5-hmC ) as a modification to mammalian DNA . 5-hmC is known to play important roles in self-renewal and cell lineage specification in embryonic stem ( ES ) cells , suggesting a potential role for 5-hmC–mediated epigenetic regulation in modulating the pluripotency of ES cells . To unveil this new regulatory paradigm in human ES cells , here we use a 5-hmC–specific chemical labeling approach to capture 5-hmC and profile its genome-wide distribution in human ES cells . We show that 5-hmC is an important epigenetic modification associated with the pluripotent state that could play role ( s ) in a subset of promoters and enhancers with defined chromatin signatures in ES cells .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genome",
"sequencing",
"genomics",
"chromosome",
"biology",
"genetics",
"epigenetics",
"biology",
"human",
"genetics",
"genetics",
"and",
"genomics"
] |
2011
|
Integrating 5-Hydroxymethylcytosine into the Epigenomic Landscape of Human Embryonic Stem Cells
|
To determine the sensitivity , specificity , and field utility of the Cepheid GeneXpert Chlamydia trachomatis ( CT ) Assay ( GeneXpert ) for ocular chlamydia infection compared to Roche Amplicor CT assay ( Amplicor ) . In a trachoma-endemic community in Kongwa Tanzania , 144 children ages 0 to 9 were surveyed to assess clinical trachoma and had two ocular swabs taken . One swab was processed at Johns Hopkins University , Baltimore MD , using Amplicor , ( Roche Molecular Diagnostics ) and the other swab was processed at a field station in Kongwa using the GeneXpert Chlamydia trachomatis/Neisseria gonorrhoeae assay ( Cepheid ) . The sensitivity and specificity of GeneXpert was compared to the Amplicor assay . Of the 144 swabs taken the prevalence of follicular trachoma by clinical exam was 43 . 7% , and by evidence of infection according to Amplicor was 28 . 5% . A total of 17 specimens ( 11 . 8% ) could not be processed by GeneXpert in the field due to lack of sample volume , other specimen issues or electricity failure . The sensitivity of GeneXpert when compared to Amplicor was 100% and the specificity was 95% . The GeneXpert test identified more positives in individuals with clinical trachoma than Amplicor , 55% versus 52% . The GeneXpert test for C . trachomatis performed with high sensitivity and specificity and demonstrated excellent promise as a field test for trachoma control .
Trachoma , a chronic conjunctivitis caused by repeated infection with Chlamydia trachomatis ( CT ) , is currently the leading cause of infectious blindness [1] . In recognition of the public health problem of trachoma as a Neglected Tropical Disease ( NTD ) , the World Health Assembly passed a resolution in 1998 calling for the elimination of blinding trachoma by the year 2020 [2] . The World Health Organization ( WHO ) has recommended the implementation of a multi-faceted control strategy , with the acronym of SAFE , by National Trachoma or NTD Control programs . SAFE stands for surgery ( to correct trichiasis ) , antibiotics ( particularly azithromycin ) to reduce the community pool of infection , face-washing programs to reduce transmission in children and environmental change to keep transmission low . With the free donation of azithromycin to endemic countries , trachoma control programs have scaled up the attempt to achieve the goal of eradication by 2020; 280 million doses of azithromycin have been provided to endemic countries since 1999 [3] . National Trachoma Control programs monitor efforts of implementing the SAFE strategy by measuring the prevalence of follicular trachoma in children ages 0 to 9 years . However , follicular trachoma can require a long time to resolve , and while there may be a rapid decline in infection following SAFE , there is often a less rapid decline in clinical disease . First reported after Mass Drug Administration ( MDA ) in the Azithromycin in Control of Trachoma Trial [4] , infection declined at one year in Tanzania , from 20% to 7% , but the decline was less marked for clinical trachoma , which declined from 64% to 42% . This finding is not unexpected , as research in animals has reported a longer time for resolution of clinical signs following the clearance of infection [5] Several investigators working in trachoma endemic countries have reported that between 40–60% of clinical follicular trachoma seen in children may not have infection [6] , [7] , [8] . The proportion is even higher with successive rounds of MDA [7] . In addition , there are cases of infection that may be either sub-clinical or have not yet manifested disease in these communities [9] . Therefore , data on the prevalence of infection may be a useful adjunct to the prevalence of clinical disease in understanding the impact of programs on trachoma over time . As was seen in The Gambia , there may be instances where infection has been eliminated and only residual clinical disease is present [10] . Existing nucleic acid amplification tests , considered a gold standard for a laboratory test of infection [11] , [12] , [13] , involves instrumentation that is expensive and requires developed laboratories not widely available in many trachoma endemic countries . Thus , there is a need for a simple , inexpensive rapid test for CT that can be performed in the field in trachoma endemic areas . The first attempt at a field test for CT was initially promising [14] but proved to not be robust under field conditions [15] . A study in the International Chlamydia Research Laboratory at Johns Hopkins compared ocular swabs tested with the Amplicor CT assay ( Roche Molecular Diagnostics , Indianapolis , IN ) , Abbott m2000 RealTime CT Assay ( m2000 ) ( Abbott Molecular Diagnostics , Des Plains , IL ) and a new test , Cepheid GeneXpert CT/NG Assay ( GeneXpert ) ( Cepheid , Sunnyvale , CA ) , and found excellent concordance [16] . The GeneXpert has already been shown to be sensitive and specific in identifying genital chlamydia infections [17] , and to be easy to use in a field setting . The GeneXpert platform is already in place in many developing country settings for use in diagnosing tuberculosis [18] , [19] . The ease of testing specimens using the GeneXpert system without the need for expensive equipment or extensive training , along with the extremely low likelihood of contamination as experienced by the tuberculosis program , suggested that this might be an ideal field test if it could be shown to maintain high sensitivity and specificity when testing ocular chlamydia samples in the field [18] , [19] . The purpose of this study was to compare the sensitivity and specificity of the GeneXpert test , as conducted in a trachoma field station in Kongwa Tanzania , against the Amplicor test carried out at the Johns Hopkins ( JHU ) International Chlamydia Laboratory in Baltimore , MD on specimens from the same eye of the same children . We also report on the experience of using GeneXpert in the field .
The study received ethical approval from the Tanzania National Institute for Medical Research and the Johns Hopkins Institutional Review Board . Written informed consent was also obtained by the parent/guardian for the inclusion of each child . Both eyes were graded for trachoma using the WHO simplified grading scheme [20] by an experienced trachoma grader using 2 . 5 loupes . Trachoma was assessed as follicular trachoma ( TF ) , the presence of at least 5 follicles size 0 . 5 mm on the conjunctiva and inflammatory trachoma ( TI ) , which is the presence of severe inflammation that obscures 50% or more of the deep tarsal vessels . Ocular swabs were collected from the left upper eyelid of each index child using identical methods . A Dacron swab ( Fisher HealthCare , Houston , TX ) was rotated and swiped across the upper conjunctiva three times and placed dry in a vial . Vials were placed in a cooler in the field . The vial containing the swab for GenXpert testing was immediately transferred to a minus 20 degree Celsius freezer at Kongwa Trachoma Project offices at the end of the day and stored until processing . The swab for Amplicor testing was stored cold until shipped frozen within 30 days of collection to the International Chlamydia Laboratory at Johns Hopkins University ( JHU ) . All testing occurred throughout September 2012 . The ocular specimens sent to JHU were processed for the detection of CT using the AMPLICOR CT/NG test ( Roche Molecular Diagnostics , Indianapolis , IN ) according to manufacturer [16] instructions . DNA extraction of the specimens was performed on the Roche MagNA Pure LC extraction robot with 200 uL of sample resulting in 100 uL of DNA elute using the MagNA Pure LC DNA Isolation Kit I ( Roche Diagnostics ) . Extracted DNA was processed using Amplicor PCR according to manufacturer's instructions; positive and negative controls were included in all DNA extractions and PCRs . It has been previously shown that automated extraction works as well as manual extraction for CT DNA in swab specimens [21] . Samples with an optical density ( OD ) of over 0 . 8 were recorded as positive for CT , and samples with ODs of under 0 . 2 were recorded as negative , samples with an OD between 0 . 2 and 0 . 799 were considered equivocal . Equivocal specimens were retested in duplicate; samples that retested equivocal on two occasions were considered negative . Specimens were processed at both sites without knowledge of the results of the other site , or the trachoma status of the child . Dry ocular swabs were rehydrated with 1–1 . 2 mL of sterile molecular grade diethylpirocarbonate ( DEPC ) ( Quality Biological Inc . Gaithersburg , MD ) water prior to testing in the field . Each specimen was vortexed for 30 seconds prior to its addition to the GeneXpert CT/NG Research Use Only Assay cartridge ( Cepheid Inc; Sunnyvale , CA Version 1 . 0 cartridge ( Lot # 01905 ) ) ; along with one tube ( 3 . 0 mL ) of binding reagent added to the binding reagent opening of the cartridge . The cartridge was loaded onto the GeneXpert module and analyzed using the GeneXpert CT/NG Assay Version 1 . 0 software . The assay run time was 1 hour and 45 minutes . The GeneXpert System is a closed , self-contained , automated platform that has minimal risk of contamination . It combines on-board sample preparation with real-time PCR to deliver answers directly from unprocessed samples . Results were reported by the computer as positive or negative for chlamydia or indeterminate ( Invalid , Error , or No Result ) . If the initial GeneXpert result was indeterminate , the specimen was re-tested one time using a new aliquot of specimen , if available , and a new GeneXpert cartridge . The assay produces an adequacy control result and an amplification control result . When either of these failed , the test was also indeterminate , and repeated . Both controls need to be amplified for a valid test result . A cohort of 144 children aged 0–9 years , from the Chilangalizi community in Kongwa , Tanzania , were enrolled in the study during January 2012 . We expected , given the trachoma rate in that community of about 50% , to have 30% positive by Amplicor Testing . We were testing a field usable potential test , and hoped to achieve at least 90% sensitivity using the Amplicor results as the ‘gold standard’ . Selecting a 95% significance level and allowing for +/−10% precision , the estimated sample size was 117 children . With our effective sample size of 127 we improved the level of precision . A total of 144 paired ocular swab specimens for this study were collected from the same eye during the survey . Of those 144 eyes , a third swab from 68 eyes was collected for another study where GeneXpert testing was to also be done in the laboratory at JHU . Specimens for testing by GeneXpert were prepared and tested the same way as described in the field and at JHU ( Figure 1 ) . Data from the third swab was only used in this study to provide more information on any discordant pairs . All results were sent to the data coordinating center at JHU . The prevalence of trachoma and ocular CT infection was determined for the 144 eye samples . The sensitivity and specificity of the GeneXpert test was compared to Amplicor and the ease of use of GeneXpert in the field was determined by examining reasons for failure to return results and anecdotal comments from the field laboratory technician ( AJ ) . Cepheid and Amplicor test performers were blind to the corresponding test result .
At Kongwa Trachoma Project ( KTP ) , GeneXpert testing for 144 samples returned results for 127 or 88 . 2% of the samples . Reasons for invalid test results included insufficient sample ( error code 5007 ) in 9/144 ( 6 . 3% ) , other material in sample ( error code 2008 ) in 4/144 ( 2 . 8% ) and a sudden loss of electricity , which resulted in loss of 4/144 ( 2 . 8% ) specimens . The characteristics of the children who had results from both GeneXpert and Amplicor as compared to children who had results from Amplicor only are shown in Table 1 . There was no significant difference between those in the final analyses and those who did not have results from GeneXpert . The same proportion of positive specimens was found by GeneXpert and Amplicor in children with TF , 39% and slightly more with GeneXpert in children with TI and in children without signs of trachoma ( Table 2 ) . All 35 positives by Amplicor were also positive by GeneXpert , and five additional positives were found by GeneXpert that were Amplicor negative ( Table 3 ) . When compared to Amplicor the sensitivity of GeneXpert was 100% and specificity was 95% . The 5 samples that were GeneXpert positive at KTP , but negative by Amplicor were re-tested at Johns Hopkins using GeneXpert , and 2 of the 5 were positive . The field technician in Kongwa noted that GeneXpert was easy to run according to the protocol . It was also observed that due to the extreme dryness of the environment , more water was needed to ensure the volume of the specimen required for optimal performance in the Cepheid machine . Once the adjustment was made by adding 1 . 2 mL of DEPC water to the sample , instead of 1 . 0 ml , there were no more failures due to insufficient sample . It was also noted that the generator ( Robin [Subaru , Japan Model RBG5000CLE , 4 . 5Kwatts , Freq 50 Hz ) was insufficient to run the GeneXpert module during electricity failure; a generator with greater power would likely have helped avoid the 4 sample losses .
In 144 paired ocular samples , we evaluated the sensitivity and specificity of GeneXpert test for C . trachomatis , as carried out under field conditions , against the Amplicor CT PCR test , as carried out at the International Chlamydia Laboratory at Johns Hopkins University . Sensitivity and specificity were high , and the result of further testing of the GeneXpert positive/Amplicor negative specimens increased the likelihood that some of these discordant samples might have been true positives . There is an added advantage of the GenXpert assay over other laboratory tests for CT , as the GeneXpert system includes a sample adequacy control test ( SAC ) , which insures that there is human DNA in the sample or the test will be reported out as “indeterminate” . It also includes a specimen processing control ( SPC ) to indicate that amplification takes place for the SPC control , indicating that there are no PCR inhibitors present . It is theoretically possible that the negative Amplicor specimens and the positive GeneXpert specimen discordance was in part due to inadequate sampling for the Amplicor specimen . This is unlikely because all the specimens were collected the same way and none of the GeneXpert samples were indeterminate , either in the field or at JHU . Since there was no order to the samples , the chances that all the indeterminate samples were sent for Amplicor processing is low . For both tests , only one freeze/thaw cycle occurred , which could result in reduction of positive results [16] . The volume of sample used by Amplicor is only 50 µl of extracted DNA from a starting volume of 200 µl of original sample . At low prevalences , there is concern that the aliquot taken for testing may not contain chlamydia . Since the extraction and processing is all internal in GeneXpert , we do not know what volume is used for testing; only that 1 . 0 ml is used at the beginning . However , this was not a low prevalence community with 49% trachoma and 28% infection in the children . Thus , while a theoretical possibility , it is unlikely an explanation for the greater number of positive samples with GeneXpert . We do not think that the loss of 17 samples during GeneXpert testing affected our results . The loss was higher among children ages 3 to 5 years , but was not statistically significant . Moreover , there was no reason to suspect bias as a result of loss of electricity , or insufficient volume of sample as these are unlikely to be related to any characteristic of a child but rather the time of the day/environment of the testing . All that was required for field testing by GeneXpert was a freezer for the samples , a computer for the GeneXpert platform to process the assay and return results , a vortexer , DEPC water , disposable pipettes , and sterile gloves for working . At KTP a desk was set up next to the freezer for the GeneXpert equipment to provide a workspace . With GeneXpert's 4 cartridge module , about 24 samples could be processed in an eight hour work day . We estimate about 86 samples can be run in two eight hour days , but other activities can be undertaken during that time . Unlike the many steps required by laboratory personnel to run a standard PCR test , the GeneXpert requires the simple addition of water to the sample , vortexing , and removal of sample as the only steps open to contamination . The use of disposable pipettes and attention to details minimizes greatly the chance for contamination using the GeneXpert platform . For future testing a dedicated generator that could supply ample power to the platform module would provide backup power to allow for a smooth transition to generator power from state supplied electricity in the case of a power outage . The requirement for a minimum of 1 ml of reconstituted sample limited how much left over sample could be retained without significant dilution and in our case when the test was lost due to sudden shut down in electricity or error in sample processing; it was only possible to perform a single retest of the failed specimen . In fact , we found that adding 1 . 2 ml eliminated errors of not enough sample , due to processing in a dry climate . The potential cost of the test kits may dictate the use of a pooling strategy . Even at an assumed low price of $10/test , 100 tests would cost $1 , 000 . A study to determine if pooling could be accomplished on the GeneXpert could potentially decrease costs . In addition , a university trained American researcher with computer and lab experience performed the field test in Tanzania , and while the GeneXpert is simpler than alternatives , GeneXpert remains untested in field settings by Tanzanian workers . However , the GeneXpert platform is being rolled out all over Africa as part of testing for tuberculosis , so there is no reason to suppose that positive experience would be any less so when testing for CT . Finally , the GeneXpert CT/NG assay was a research use only assay at the time of this study , but has now been approved by the Federal Drug Administration . The low cost of the processing platform , the ease of processing with readily available materials , plus our results showing the high sensitivity and specificity , suggest this approach may be ideal for a field test for trachoma control programs .
|
Trachoma , an eye infection caused by C . trachomatis , is the leading cause of infectious blindness worldwide , affecting the developing world . The current standard for trachoma treatment involves mass drug administration ( MDA ) of an antibiotic that is given to a community to reduce transmission . A field test for the presence of infection would be a useful adjunct in measuring MDA impact . However , the current standard for measuring infection involves expensive , delicate instrumentation that is often only in laboratories in developed countries or capital cities , and eye swab specimens are mostly shipped to the developed world for analysis . This study compared a standard method for infection analysis , Roche Amplicor , in the United States , with a new test , the Cepheid GeneXpert , in the field in Tanzania . We collected 144 duplicate eye swabs in children ages 0–9 years . 12% of specimens could not be analyzed by GeneXpert due to correctable technical difficulties . Of those analyzed , 100% of samples negative by Amplicor were also negative by GeneXpert , and 95% of samples positive by GeneXpert were also positive by Amplicor . The GeneXpert was easy to use with minimal opportunities for contamination , and is a promising new test for field-testing infection in trachoma control efforts .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"neglected",
"tropical",
"diseases",
"infectious",
"diseases",
"trachoma"
] |
2013
|
Field Evaluation of the Cepheid GeneXpert Chlamydia trachomatis Assay for Detection of Infection in a Trachoma Endemic Community in Tanzania
|
Elg1 , the major subunit of a Replication Factor C-like complex , is critical to ensure genomic stability during DNA replication , and is implicated in controlling chromatin structure . We investigated the consequences of Elg1 loss for the dynamics of chromatin re-formation following DNA replication . Measurement of Okazaki fragment length and the micrococcal nuclease sensitivity of newly replicated DNA revealed a defect in nucleosome organization in the absence of Elg1 . Using a proteomic approach to identify Elg1 binding partners , we discovered that Elg1 interacts with Rtt106 , a histone chaperone implicated in replication-coupled nucleosome assembly that also regulates transcription . A central role for Elg1 is the unloading of PCNA from chromatin following DNA replication , so we examined the relative importance of Rtt106 and PCNA unloading for chromatin reassembly following DNA replication . We find that the major cause of the chromatin organization defects of an ELG1 mutant is PCNA retention on DNA following replication , with Rtt106-Elg1 interaction potentially playing a contributory role .
The genetic material in eukaryotes is packaged into chromatin , composed mainly of DNA and nucleosomes . During DNA replication , DNA helicases separate the two parental strands of DNA and nucleosomes are removed from the DNA . Once the nascent DNA strands have been synthesized , the nucleosomal structure must be reassembled to restore the chromatin and permit reinstatement of epigenetic information . Defective chromatin re-assembly leads to improper chromatin formation and loss of epigenetic marks carried on the parental histones , resulting in genomic instability [1] . Various replication-associated factors play a key role in ensuring all the genetic and epigenetic information is efficiently duplicated . A critical component of the replication machinery is PCNA , which serves as the processivity factor for DNA polymerases . Apart from acting as an accessory factor for DNA polymerase , PCNA coordinates replication-associated processes including chromatin re-assembly , cohesion establishment , DNA repair and the damage response [2] . PCNA is loaded onto chromatin during replication by the Replication Factor C ( RFC ) , a pentameric complex consisting of Rfc1-5 [3 , 4] . During the initiation of each Okazaki fragment , RFC loads PCNA prior to polymerase δ recruitment . On completion of each Okazaki fragment , PCNA must then be unloaded , which requires the Elg1 RFC-Like Complex ( also called Elg1-RLC; [5 , 6] . The Elg1-RLC contains the same Rfc2-5 subunits as RFC , but the largest subunit Rfc1 is replaced by Elg1 . Timely removal of PCNA is important , and PCNA accumulation in the absence of Elg1 contributes to genomic instability phenotypes such as elongated telomeres , telomeric silencing , chromosomal rearrangements , cohesion defects , and increased sister chromatin recombination [7–11] . Histone chaperones are crucial auxiliary components of the replication machinery [12 , 13] , which ensure the proper coupling of DNA replication with re-assembly into nucleosomes [14] . FACT complex of the budding yeast S . cerevisiae contains subunits Spt16 and Pob3 , and can bind both H2A-H2B and H3-H4 . FACT associates with components of the replication machinery including the MCM complex and DNA polymerase δ [15 , 16] and acts in parental histone recycling and placement on the newly replicated DNA , as well as being implicated in transcription-coupled chromatin control [17 , 18] . In S . cerevisiae newly synthesized histone H3-H4 dimers are bound by the histone chaperone Asf1 , with new histone H3 preferentially acetylated at H3K56 . Asf1 binding and H3K56Ac modification promote the interaction of new H3-H4 with further histone chaperones including CAF-1 and Rtt106 [19] , and Asf1 additionally interacts with RFC [20] . CAF-1 is a three subunit complex consisting in yeast of subunits Cac1 , Cac2 , and Cac3 . Two CAF-1 complexes associate to assemble an H3-H4 tetrasome in the initial step of nucleosome re-assembly [21] . CAF-1 promotes nucleosome assembly at replication forks through interaction with PCNA and by binding to DNA directly [22–24] . Rtt106 is also implicated in nucleosome reassembly following DNA replication . Containing two Pleckstrin Homology Domains that mediate its preference for K56-acetylated H3 [21] , Rtt106 has been shown to dimerize to mediate assembly of an H3-H4 tetrasome [25 , 26] . Deletion of RTT106 when combined with deletion of CAC1 showed a defect in deposition of H3K56Ac , which is marker of newly deposited histone in yeast [19 , 27] . Rtt106 is also involved in heterochromatin formation: rtt106Δ mutant cells exhibit loss of silencing at mating type loci and telomeres [19 , 28] . In addition , Rtt106 is proposed to be important for nucleosome assembly during transcription at highly transcribed genes [29] and in regulation of histone gene expression [30 , 31] . However , it remains unknown how Rtt106 is recruited to required sites of nucleosome assembly . Because of the links between PCNA and nucleosome assembly , and the effects on chromatin and genome stability caused by ELG1 deletion [9] , we were prompted to investigate whether the PCNA unloading factor Elg1 has a role also in the chromatin re-assembly process . Here we show that Elg1 activity is critical for timely nucleosome organization on nascent DNA . We moreover discovered that Elg1 interacts with histone chaperones , in particular Rtt106 and the FACT complex , with the interaction of Elg1 and Rtt106 not dependent on PCNA . We find however that the most significant cause of defective post-replication nucleosome organization in an elg1Δ mutant is delayed unloading of PCNA , with Elg1-Rtt106 interaction potentially playing a contributory role .
The process of DNA replication and nucleosome re-assembly are tightly coupled . Because it acts at replication forks in PCNA unloading , we examined if Elg1 also affects nucleosome deposition onto newly replicated DNA . Initially , we examined Okazaki fragment length in strains lacking Elg1 . Okazaki fragment length can be used as a proxy for nucleosome deposition , since fragment length tends to be determined by the newly deposited nucleosome on the immediately preceding fragment [13 , 32] . To permit the visualization of Okazaki fragments , we used a strain background with an Auxin-Inducible Degron ( AID ) -tagged copy of the DNA ligase gene CDC9 , which accumulates unligated Okazaki fragments during S phase in the presence of auxin . Cells were synchronized in G1 then released into S phase for 55 min , and then Okazaki fragments visualized by 3' end-labelling and gel electrophoresis as described [6 , 13] ( Fig 1A & 1B ) . In normal cells , Okazaki fragment lengths tend to cluster around 180 bases and 360 bases corresponding to mono- and di-nucleosomal sizes . As previously described , Okazaki fragments are somewhat extended in the mutant cac2Δ which lacks the CAF-1 chromatin assembly factor ( Fig 1C ) [13 , 32] . This lengthening is believed to reflect aberrant and delayed nucleosome repositioning , which causes continued nick translation and Okazaki fragment lengthening by DNA polymerase δ , since it does not encounter a nucleosome on the previously synthesized DNA that would stimulate its disengagement . In an elg1Δ mutant , we found that Okazaki fragment lengths also differed from wild-type , showing a generally broader distribution with a higher proportion of fragments extended in length when compared to wild-type ( Fig 1C & S1 Fig ) . This Okazaki fragment lengthening suggests that the elg1Δ mutation may cause a nucleosome assembly defect . The lengthened Okazaki fragment phenotype was not shared by a ctf18Δ mutant , which lacks the Ctf18-RLC complex that is involved in establishment of cohesion [33 , 34] . The effect of Elg1 in limiting Okazaki fragment length therefore appears specific to Elg1-RLC . Since Cdc9 depletion is intrinsic to the Okazaki fragment detection assay , we cannot exclude the possibility that lack of Cdc9 contributes to this Okazaki fragment lengthening effect in the elg1Δ mutant . To examine chromatin re-assembly in elg1Δ using a different approach , we next tested the sensitivity of chromatin to digestion by Micrococcal Nuclease ( MNase ) , since defective chromatin re-assembly can result in increased accessibility to digestion by this nuclease . There was no evident abnormality in MNase sensitivity of bulk chromatin in an elg1Δ mutant . However , defects in replication-coupled chromatin re-assembly tend to be transient and quickly restored following replication by redundantly acting histone chaperones and/or replication-independent histone turnover [35] . To test nucleosome deposition onto newly replicated DNA , we used cultures synchronized by release from α factor into S phase and examined the MNase sensitivity of nascent DNA labelled with the thymidine analog 5-Bromo 2-deoxyuridine ( BrdU ) ( Fig 2A & 2B ) . These experiments used strains genetically modified to incorporate BrdU . After Southern blot transfer of MNase-digested DNA to membrane , nascent DNA was specifically visualized by probing the DNA on the membrane with anti-BrdU antibody . Validating the assay , nascent DNA in a cac1Δ mutant ( S2C Fig ) was more sensitive than wild-type to MNase digestion , due to delayed chromatin re-assembly [35] . We found that nascent DNA in the elg1Δ mutant ( Fig 2C ) was also more sensitive to MNase than wild-type , as evidenced by an increased proportion of mononucleosomal compared to disomal digested fragments ( Fig 2C lower panel , compare proportion of disome and monosome bands and signal traces of 45 min samples of nascent DNA in Fig 2D ) . This increased sensitivity to MNase digestion in elg1Δ was reproducible , as illustrated by the additional gels shown in S2A & S2B Fig . The magnitude of the effect did vary between experiments: the proportion of mono-nucleosomal to total nascent DNA was increased 1 . 7-fold in elg1Δ relative to wild-type in Fig 2C , 1 . 2-fold in S2A Fig , and 2 . 6-fold in S2B Fig . Such variation is to be expected given the semi-quantitative nature of such experiments , but overall the elevated accessibility of nascent DNA to MNase digestion is indicative of defective or delayed nucleosome assembly . The differences in sensitivity to MNase are not caused by different rates of progression through S phase of WT and elg1Δ cells ( S3 Fig ) . To summarize , our observation of extended Okazaki fragments and increased sensitivity to micrococcal nuclease in the elg1Δ mutant suggest a role for Elg1 in replication-coupled nucleosome re-organization . The results presented above prompted us to investigate effects of the elg1Δ mutation on nucleosome assembly genome-wide . We used thymidine analog 5-ethynyl-2’-deoxy-uridine ( EdU ) to label newly replicated DNA in G1-arrested cells released into S phase . Following MNase digestion , EdU-labelled nascent DNA was isolated by affinity purification ( Fig 3A ) . After deep sequencing [35] , nucleosomal reads were then aligned with respect to origins of replication ( Fig 3B & 3C ) or transcription start sites ( TSS ) of all genes ( S4 Fig ) . While no difference in the organization of nucleosomes either upstream or downstream of origins was observed in G1 control samples , a clear defect in organization of nucleosomes is observed in elg1Δ ( Fig 3B ) at early time points after release ( 27 min , 30 min , 33 min ) when compared to WT . As cells reach the end of S phase ( 60 min ) the nucleosomal pattern in the elg1Δ mutant becomes more organized and similar to WT , consistent with recovery of normal nucleosome distribution as previously described [35] . Defective nucleosome organization in elg1Δ mutant is somewhat similar to that seen in a cac1Δ mutant ( Fig 3C & S4B Fig ) although the cac1Δ mutant shows an increased spacing of nucleosomes on nascent DNA that is not obviously shared by elg1Δ . To identify interaction partners of Elg1 potentially connected to nucleosome assembly , we used SILAC-based mass spectrometry to identify co-precipitating proteins . Strains expressing untagged or FLAG-tagged versions of Elg1 were differentially labelled with isotopically light or heavy lysine and arginine , and immunoprecipitated proteins ( Fig 4A ) were analyzed by mass spectrometry . As expected , the Elg1-FLAG samples showed strong enrichment of Elg1 and Rfc2-5 ( the other Elg1-RLC subunits ) and also of PCNA . Strikingly , the histone chaperone Rtt106 was also enriched at levels similar to the Rfc2-5 subunits ( Fig 4B & 4C ) . Also enriched were Spt16 and Pob3 , two subunits of the FACT complex . Both Rtt106 and FACT complex are implicated in replication-coupled nucleosome assembly: while FACT appears to mediate recycling of parental histones , Rtt106 is involved in depositing newly synthesized histones [18 , 19 , 36] . The interactions suggest that these histone chaperones , particularly Rtt106 , could potentially mediate the nucleosome assembly role of Elg1 . We carried out further co-immunoprecipitation experiments with Rtt106 to confirm and investigate the Elg1-Rtt106 interaction . Immunoprecipitation of Elg1-FLAG pulled down HA-tagged Rtt106 ( Fig 5A ) . Pulldown of Elg1 truncation mutants showed that both the Elg1 N-terminal and C-terminal regions are important for the interaction with Rtt106 ( S6 Fig ) . These regions are unique to Elg1 , having only very limited sequence similarity with Rfc1 or Ctf18 . Consistently , neither Rfc1 nor Ctf18 showed interaction with Rtt106 in co-immunoprecipitation experiments ( Fig 5B ) , suggesting interaction with Rtt106 is a property specific for Elg1 amongst the major subunits of RFC and its related complexes . Immunoprecipitation of Elg1-FLAG pulled down not only Rtt106 but also PCNA , reflecting the function of Elg1-RLC as the major PCNA unloader . Co-immunoprecipitation experiments in the presence of increasing salt concentrations showed that interaction with PCNA was lost at a concentration where Rtt106-Elg1 interaction was retained ( Fig 5C , 250mM potassium acetate & S7 Fig ) , indicating that the Elg1-Rtt106 interaction is not mediated through PCNA . Note that a band appearing in Western analysis slightly below full-length Elg1 ( Fig 5A , 5B & 5C ) appears to represent a degradation product whose appearance is stimulated by increased salt concentration . To summarize , our results indicate that robust interaction occurs between Elg1 and Rtt106 , specific to Elg1 amongst the RFC-related complexes . Since Elg1 is important for nucleosome deposition and interacts with Rtt106 , we reasoned that , during DNA replication on the lagging strand , Elg1 might concomitantly recruit Rtt106 as it unloads PCNA , thereby coupling PCNA unloading and chromatin re-assembly . Alternatively , Rtt106 might participate in the PCNA unloading function of Elg1 . Examining the accumulation of PCNA on chromatin in the absence of Rtt106 ( S8 Fig ) did not show clear evidence for a role for Rtt106 in PCNA unloading . We therefore followed up the possibility that Elg1 interaction is important to recruit Rtt106 for chromatin re-assembly , by investigating whether recruitment of Rtt106 to replicating regions is dependent on Elg1 . We carried out ChIP-seq analysis of HA-tagged Rtt106 on cells released into hydroxyurea from a G1 arrest . However contrary to our expectation , we did not consistently observe association of Rtt106 with newly replicated regions at early origins ( e . g . ARS306 , ARS510 , ARS310 , S9B Fig ) . Nor did we observe convincing Rtt106 recruitment to replicating chromatin in a similar experiment carried out under unperturbed conditions ( i . e . in WT cells with no HU treatment ) . Our ChIP experiments did effectively identify Rtt106 binding as we did observe Rtt106 localization at the promoter HTA1-HTB1 promoter ( S9A Fig ) , as previously described [37] . Rtt106 recruitment to the HTA1-HTB1 promoter was not affected in the absence of Elg1 ( S9A Fig ) . We did notice Rtt106 association with the promoters of some genes encoding putative drug exporters , that in some cases appeared Elg1-dependent . This promoter association does not appear replication-linked , since it was observed at some late-replicating regions that forks will not reach under the HU block conditions of the experiment . The importance of Rtt106 promoter binding will be described elsewhere . Given the effect of Elg1 on chromatin re-assembly and its interaction with Rtt106 , we tested whether the two proteins act in chromatin re-assembly in the same pathway . Specifically , we examined whether the elg1Δ and rtt106Δ mutations have similar effects on the length of Okazaki fragments . We found that rtt106Δ causes only mild lengthening of Okazaki fragments , the degree of lengthening much less than observed for elg1Δ . Moreover , the effect of elg1Δ rtt106Δ double mutation on Okazaki fragments appeared to be additive rather than epistatic when compared to the single mutations ( Fig 6A ) . These effects suggest that Elg1 acts in a distinct pathway from Rtt106 . Hence , we considered other mechanisms through which elg1Δ might affect chromatin re-assembly . The absence of Elg1 results in prolonged accumulation of PCNA on chromatin [11] , which could potentially interfere with nucleosome deposition causing defective chromatin re-organization . To investigate this possibility , we made use of trimer instability mutations in PCNA . These mutations cause the PCNA ring to be disassembly-prone , falling off DNA spontaneously even in the absence of Elg1 and thereby suppressing the PCNA accumulation phenotype of the elg1Δ mutant [11] . Okazaki fragment length assays were performed in double mutants where elg1Δ was combined with two different trimer instability PCNA mutants , pol30-R14E ( Paul Solomon Devakumar et al . in revision ) and pol30-D150E [11] . We observed that in these double mutants , Okazaki fragments were restored to normal length , when compared to the elongated Okazaki fragments of the elg1Δ single mutant ( Fig 6B & S10 Fig ) . Based on this observation , we propose that when normal PCNA unloading fails due to absence of Elg1 , aberrant PCNA accumulation on the newly replicated DNA leads to defective nucleosome deposition .
In this investigation , we show that Elg1 contributes to proper nucleosome assembly across the genome after DNA replication , as evidenced by Okazaki fragment lengthening ( Fig 1 ) and elevated sensitivity of nascent DNA to micrococcal nuclease digestion ( Figs 2 & 3 ) in an elg1Δ mutant . Okazaki fragment length has previously been examined in several studies as a proxy for nucleosome deposition [32] . This assay could raise the concern that the DNA ligase-deficient background required to visualize Okazaki fragments might itself impact on fragment length or nucleosome re-assembly , but a different study [38] obtained consistent results , also finding that nucleosome position determines S . cerevisiae Okazaki fragment positioning , using a completely different approach that analyzed mutations inserted by an error-prone polymerase α prone to ribonucleotide insertion . Moreover , in assays that measure the micrococcal nuclease sensitivity of nascent DNA ( in cells where DNA ligase activity is intact ) we confirmed that nucleosome deposition is affected by the elg1Δ mutation . Therefore , the Okazaki fragment lengthening phenotype indeed reflects a nascent strand chromatin re-assembly defect . To understand interactions that may contribute to the chromatin re-assembly effect of Elg1 , we examined the proteins that co-precipitate with Elg1 in pull-down experiments , and identified novel interactions of Elg1 with histone chaperones , in particular Rtt106 and the FACT complex . Interestingly , Rtt106 appears to bind the Elg1-RLC in almost stoichiometric amounts , in an interaction that does not depend on PCNA . Rtt106 does not interact with either Rfc1 or Ctf18 . Consistently , we found that both the N-terminal and C-terminal regions that are unique to Elg1 are needed for Rtt106 interaction ( S6 Fig ) . To examine the extent to which Rtt106-Elg1 interaction versus the Elg1 PCNA unloading function are important for chromatin re-assembly , we made use of disassembly-prone mutants of PCNA which do not accumulate on chromatin even in the absence of Elg1 . Using these mutations to relieve PCNA accumulation on chromatin in an elg1Δ background restored Okazaki fragments to normal length , indicating that prompt and effective PCNA unloading is absolutely essential for normal nucleosome deposition in the wake of replication forks . How might PCNA accumulation result in defective nucleosome assembly and associated Okazaki fragment lengthening ? Okazaki fragment length is proposed to be regulated by nucleosome deposition on the previously synthesized section of DNA [13 , 38] as illustrated in Fig 7 . The newly deposited nucleosome on the last piece of DNA synthesized is believed to form an obstacle to progression of polymerase δ as it carries out strand displacement synthesis , prior to completing synthesis of each Okazaki fragment . Encounter of pol δ with the nucleosome is suggested to favour pol δ disengagement and dissociation , allowing PCNA to recruit DNA ligase [39] with ligation of the completed Okazaki fragment to the nascent lagging strand determining the final Okazaki fragment length ( Fig 7 , Model i ) . We propose that in the absence of Elg1 , accumulated PCNA in the wake of the replisome obstructs normal placement and spacing of nucleosome deposition , so that the nucleosomal barrier to pol δ synthesis is not present , resulting in longer Okazaki fragments being synthesized prior to their eventual completion and ligation ( Fig 7 , Model ii ) . Combining the elg1Δ mutation with a PCNA trimer-unstable mutant prevents the accumulation of PCNA , relieving the block to nucleosome deposition and restoring the normal mechanism of Okazaki fragment length determination ( Fig 7 , Model iii ) . Our findings support the suggestion that nucleosome deposition is a very early event that precedes and stimulates pol δ dissociation , the polymerase in turn allowing DNA ligase recruitment by PCNA [39] and subsequent Okazaki fragment ligation , which is necessary for PCNA unloading by the Elg1-RLC . Our results are therefore consistent with the previously identified dependence of PCNA unloading on Okazaki fragment ligation [6] . A very recent study by [40] provides an interesting illustration of the consequences of disrupting PCNA removal by Elg1-RLC and nucleosome deposition . Janke et al used an assay that measures heterochromatin disruption , by testing for failure to silence expression of a Cre recombinase gene . Their finding that silencing is disrupted by an elg1Δ mutation ( or by histone chaperone mutations ) implies that normal replication-coupled chromatin assembly is needed to preserve silencing at a specific heterochromatic locus . Our study generalizes the conclusion that Elg1 activity is needed for normal chromatin inheritance , with the discovery that nucleosome deposition problems caused by failure to unload PCNA extend genome-wide . Since delayed PCNA removal appears to be the main cause of the chromatin re-assembly defect observed in elg1Δ , what is the significance of Elg1 interaction with histone chaperones , in particular Rtt106 and FACT complex ? Identification of these interactions raises the suggestion that Elg1 might recruit histone chaperones to assist in chromatin reassembly , with Elg1 thereby contributing to chromatin re-configuration or maturation . However , our ChIP analysis failed to identify a clear role for Elg1 in localizing Rtt106 to newly synthesized DNA . We did find that Elg1 has effects on Rtt106 chromatin association at the promoters of a number of genes , particularly genes involved in cellular transport and drug resistance . However , this effect is unlikely to be coupled to DNA replication since we observed Rtt106 association with several such sites in G1 phase samples . Slight sensitivity of an elg1Δ mutant to HU [8] would be consistent with a need for Elg1 in controlling the expression of genes required for drug response and export . The possibility of a non-replication-associated role for Elg1 in regulating gene expression through histone chaperone recruitment is the subject of ongoing study . While PCNA accumulation appears to be the immediate cause of delayed nucleosome deposition in the elg1Δ mutant ( Fig 7 ) , our results do not exclude the possibility of a role for Rtt106-Elg1 in replication-coupled chromatin re-establishment , especially since presence of multiple , redundant histone chaperones activities in yeast complicates analysis of chromatin re-assembly . However , we could not obtain unambiguous , reproducible evidence of a role for Elg1-Rtt106 interaction following replication . One possibility is that Elg1 does contribute to coordination of chromatin re-assembly , operating through Rtt106 and/or other histone chaperones , in a pathway acting at a later stage of chromatin maturation operating after histone deposition and Okazaki fragment ligation . The role of Elg1 appears to be conserved , since its mammalian homolog , called ATAD5 , also appears to mediate PCNA unloading [41] . Mammalian cells lacking ATAD5 show PCNA accumulation on chromatin similar to that observed in yeast , and it seems likely that such PCNA retention may impact chromatin re-assembly . The major phenotype of mice lacking ATAD5 is cancer predisposition , and indeed ATAD5 mutations are also proposed to contribute to human ovarian cancers [42 , 43] . Defects in genomic function caused by derailed chromatin re-assembly following replication might therefore contribute significantly to human cancer development or progression .
All yeast strains used in this study are listed in S1 Table . Gene disruptions and epitope tags were introduced by standard PCR based methods [44 , 45] . Okazaki fragment purification and detection was performed as described previously in [13] . Yeast cells were grown to OD600 of 0 . 2 at 30°C in 60ml YPD media and then alpha factor was added to arrest cells in G1 phase . 400μg/ml BrdU was added to the culture and incubated for 30 minutes for cells to take up BrdU . Cells were then released into S phase by resuspending in fresh YPD containing 400μg/ml BrdU . Then 20 ml samples were collected at desired time points into formaldehyde ( 1% final concentration ) and incubated with rotation for 15 minutes at room temperature . 125mM glycine was then added to neutralise formaldehyde . Cells were washed twice in 10 ml of ice cold 1X PBS , then with 2 ml of spheroplasting buffer ( 1M sorbitol , 1mM beta-mercaptoethanol ) before resuspending in 1ml of spheroplasting buffer with 300μg/ml 100-T Zymolase then incubated at 30°C for 20 minutes . Spheroplasts were washed in 1ml of spheroplasting buffer and resuspended in 600μl of Digestion buffer ( 1M sorbitol , 50mM Nacl , 10mM Tris-HCl pH7 . 4 , 5mM MgCl2 , 5mM CaCl2 , 0 . 075% Nonidet P-40 , 1mM beta-mercaptoethanol , 0 . 5mM spermidine ) . 200μl aliquots were subjected to micrococcal nuclease ( NEB , M0247S ) digestion ( 200 or 600 gel units ) for 5 minutes at 37°C . Digestions were stopped by adding 1/10 volume of stop solution ( 250mM EDTA , 5% SDS ) . 5μl of 20 mg/ml Proteinase K was added and incubated overnight at 55°C . Following phenol-chloroform extraction , DNA was precipitated using 1/10 volume of 3M sodium acetate and 2 volumes of 100% ethanol . The air-dried DNA pellet was resuspended in 20μl of TE buffer with RNase A ( 1mg/ml ) and incubated for 2 hours at 37°C . DNA samples were electrophoresed on a 1 . 4% agarose gel , which was incubated in denaturing buffer ( 0 . 5M NaoH , 1M NaCl ) twice for 15 minutes followed by incubation in neutralization buffer ( 0 . 5M Tris-Hcl , 3M Nacl ) for 30 minutes . The DNA was then transferred to Amersham Hybond N+ membrane by Southern blotting . DNA was cross-linked to the membrane with UV light ( 1200J ) . The membrane was then incubated in 5% milk in TBS-tween for 60 minutes and probed with anti-BrdU antibody ( ab12219 , abcam ) . Whole cell extract preparation , western blotting and co-immunoprecipitation experiments were performed as described previously [5 , 6] . Antibodies used were: anti-BrdU ( ab12219 , abcam ) , anti-FLAG ( F1804 , Sigma ) , anti-HA ( HA . 11 clone 16B12 , Covance ) , anti-PCNA ( ab70472 , abcam ) . SILAC Quantitative proteomic analysis was performed as described previously [46] . Yeast strains were grown to an OD600 of 0 . 25 in YPD . Alpha factor was added to arrest cells in G1 and released into YPD containing 0 . 2M hydroxyurea at 23°C for 60 minutes . Formaldehyde ( 1% final concentration ) was added and incubated with rotation first at room temperature for 20 minutes and then at 4°C overnight . Cells were washed 3 times with ice-cold 1X Phosphate buffered saline . Cells were pelleted and frozen at -80°C . Rtt106 ChIP using anti-HA ( HA . 11 clone 16B12 , Covance ) and data analysis were performed as described previously [6] . ChIP-Seq data are uploaded to Array Express under accession number: E-MTAB-6985
|
DNA replication is the central process that duplicates the genetic information during cell multiplication . Many cellular factors play important roles in the efficient and accurate duplication of DNA , critical for faithful transmission of genetic information . One such factor is Elg1 . Elg1 acts to unload PCNA , the ring-shaped processivity factor that holds DNA polymerases on DNA for replication . In this work , we identify an additional role for Elg1 during replication . We show that lack of Elg1 leads to defects in packaging of DNA into chromatin after DNA replication . In addition , we found that Elg1 interacts with histone chaperones , factors which play key role in chromatin formation . Examining causes of the chromatin re-assembly defect , we show that accumulation of PCNA on DNA is the main cause of defective chromatin formation in the absence of Elg1 . By uncovering a new route through which Elg1 ensures chromosomes are perfectly copied , our findings advance understanding of how Elg1 contributes to the stability of the genome through its key roles in DNA replication .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"nucleases",
"enzymes",
"cell",
"cycle",
"and",
"cell",
"division",
"cell",
"processes",
"dna-binding",
"proteins",
"enzymology",
"fungi",
"dna",
"replication",
"immunoprecipitation",
"epigenetics",
"dna",
"synthesis",
"phase",
"co-immunoprecipitation",
"chromatin",
"research",
"and",
"analysis",
"methods",
"chromosome",
"biology",
"proteins",
"gene",
"expression",
"histones",
"nucleosomes",
"precipitation",
"techniques",
"yeast",
"biochemistry",
"hydrolases",
"eukaryota",
"cell",
"biology",
"nucleic",
"acids",
"genetics",
"biology",
"and",
"life",
"sciences",
"organisms"
] |
2018
|
Identification of Elg1 interaction partners and effects on post-replication chromatin re-formation
|
Systemic lupus erythematosus ( SLE ) , a complex polygenic autoimmune disease , is associated with increased complement activation . Variants of genes encoding complement regulator factor H ( CFH ) and five CFH-related proteins ( CFHR1-CFHR5 ) within the chromosome 1q32 locus linked to SLE , have been associated with multiple human diseases and may contribute to dysregulated complement activation predisposing to SLE . We assessed 60 SNPs covering the CFH-CFHRs region for association with SLE in 15 , 864 case-control subjects derived from four ethnic groups . Significant allelic associations with SLE were detected in European Americans ( EA ) and African Americans ( AA ) , which could be attributed to an intronic CFH SNP ( rs6677604 , in intron 11 , Pmeta = 6 . 6×10−8 , OR = 1 . 18 ) and an intergenic SNP between CFHR1 and CFHR4 ( rs16840639 , Pmeta = 2 . 9×10−7 , OR = 1 . 17 ) rather than to previously identified disease-associated CFH exonic SNPs , including I62V , Y402H , A474A , and D936E . In addition , allelic association of rs6677604 with SLE was subsequently confirmed in Asians ( AS ) . Haplotype analysis revealed that the underlying causal variant , tagged by rs6677604 and rs16840639 , was localized to a ∼146 kb block extending from intron 9 of CFH to downstream of CFHR1 . Within this block , the deletion of CFHR3 and CFHR1 ( CFHR3-1Δ ) , a likely causal variant measured using multiplex ligation-dependent probe amplification , was tagged by rs6677604 in EA and AS and rs16840639 in AA , respectively . Deduced from genotypic associations of tag SNPs in EA , AA , and AS , homozygous deletion of CFHR3-1Δ ( Pmeta = 3 . 2×10−7 , OR = 1 . 47 ) conferred a higher risk of SLE than heterozygous deletion ( Pmeta = 3 . 5×10−4 , OR = 1 . 14 ) . These results suggested that the CFHR3-1Δ deletion within the SLE-associated block , but not the previously described exonic SNPs of CFH , might contribute to the development of SLE in EA , AA , and AS , providing new insights into the role of complement regulators in the pathogenesis of SLE .
SLE ( OMIM 152700 ) is a debilitating autoimmune disease with strong genetic and environmental components , characterized by the production of autoantibodies resulting in tissue injury of multiple organs [1] . In SLE patients , aberrant complement activation leads to inflammatory injury [2] , and fluctuation of serum C3 is a commonly used clinical biomarker of SLE disease activity [3] . In addition , a hereditary deficiency of C1q , C1r , C1s , C4 or C2 of the classical complement pathway impairs the clearance of immune complexes and debris from apoptotic cells , which strongly predisposes to SLE susceptibility [2] . Common variants of C3 and C4 have also been associated with risk of SLE [4] , [5] , [6] . Collectively , these findings indicate the important role of complement in the development of SLE . Complement factor H ( CFH ) , a key regulator of the alternative complement pathway , modulates the innate immune responses to microorganisms , controls C3 activation and prevents inflammatory injury to self tissue [7] , [8] . CFH inhibits complement activation by preventing the formation and accelerating the decay of C3 convertase and acting as a cofactor for factor I-mediated degradation of C3b , both in plasma and on cell surfaces . Structurally , CFH contains 20 short consensus repeats ( SCRs ) . SCR1-4 in the N-terminus mediate the cofactor/decay accelerating activity and SCR19-20 in the C-terminus are essential for cell surface regulation of CFH . In addition , CFH contains specific binding sites for polyanion ( heparin or sialic acid ) , C-reactive protein ( CRP ) and microorganisms . CFH has five related proteins ( CFHR1-5 ) , all of which are also composed of SCRs [9] . SCRs in the N-terminus and C-terminus of CFHRs are highly homologous to SCR6-9 and SCR19-20 of CFH , respectively , suggesting that CFHRs and CFH may compete for binding to ligands . CFHRs lack SCRs homologous to SCR1-4 of CFH , and consequently do not exhibit cofactor/decay accelerating activity . Distinct from CFH , CFHR1 can inhibit C5 convertase activity and the formation of terminal membrane attack complex ( MAC ) [10] . A recent study has shown that CFH deficiency accelerates the development of lupus nephritis in lupus-prone mice MRL-lpr [11] . However , the role of CFHRs in the pathogenesis of SLE is still unknown . CFH , CFHR3 , CFHR1 , CFHR4 , CFHR2 and CFHR5 , that present in tandem as a gene cluster located in human chromosome 1q32 , are positional candidate genes within the 1q31-32 genomic region linked to SLE [12] , [13] . In recent years , multiple exonic SNPs in CFH , such as I62V , Y402H , D936E and A473A , have been specifically associated with various human diseases including age-related macular degeneration ( AMD ) [14] , [15] , atypical hemolytic uremic syndrome ( aHUS ) [16] and membranoproliferative glomerulonephritis type II ( MPGN II ) [16] , [17] as well as host susceptibility to meningococcal disease [18] . In addition , a common deletion of CFHR3 and CFHR1 ( CFHR3-1Δ ) has been associated with increased risk of aHUS [19] and decreased risk of AMD [20] . Taken together , these data prompted us to test whether genetic variants in CFH and CFHRs predisposed to SLE susceptibility . Although recent genome wide association studies ( GWAS ) have b`n successfully used to identify SLE susceptibility genes [21] , they still may be underpowered for specific genomic regions due to many factors such as sample size , marker density , ethnicity of subjects and over-stringent significance threshold . In these cases , a well-designed candidate gene-based association study can be used as a complementary approach to GWAS to identify genetic variants with modest effect size . In this study , we fine mapped the CFH-CFHRs region using 60 SNPs and assessed their association with SLE susceptibility in a collection of 15 , 864 subjects ( 8 , 372 cases vs . 7 , 492 controls ) from four ethnic groups . In addition , we assessed the association of CFHR3-1Δ with SLE by using tag SNPs .
To assess the association of CFH and CFHRs genes with SLE , we genotyped 60 tag SNPs covering the ∼360 kb CFH-CFHRs region in unrelated case-control subjects derived from four ethnic groups including European Americans ( EA ) , African Americans ( AA ) , Asians ( AS ) , and Hispanics enriched for the Amerindian-European admixture ( HS ) ( Figure 1A ) ( Table S1 ) . According to the latest Hapmap CEU dataset ( release 28 ) , within the CFH-CFHRs region , 203 of 224 ( 90% ) common SNPs ( frequency>5% ) could be captured by SNPs used in this study with r2>0 . 70 . Within the most-studied gene CFH , previously identified disease-associated exonic SNPs including I62V ( rs800292 , typed ) , Y402H ( tagged by rs7529589 ) , D936E ( tagged by rs10489456 ) and A474A ( tagged by rs1410996 ) were evaluated for the association with SLE . In the largest dataset ( 3 , 936 EA cases vs . 3 , 491 EA controls ) , after removing those failing the Hardy-Weinberg equilibrium ( HWE ) testing or showing low genotyping quality , fourteen SNPs were significantly associated with SLE ( allelic P<0 . 05 ) ( Table 1 ) , of which rs6677604 , located in intron 11 of CFH , exhibited the strongest association signal ( minor allele frequency [MAF]: 23 . 0% in case vs . 20 . 1% in control , P = 2 . 4×10−5 , OR[95%CI] = 1 . 19[1 . 10–1 . 28] ) . In the second largest dataset ( 1 , 679 AA cases vs . 1 , 934 AA controls ) , four SNPs were significantly associated with SLE ( Table 1 ) , all of which confirmed the association detected in EA , with rs16840639 , located in the intergenic region between CFHR1 and CFHR4 , showing the strongest association signal with a similar effect size ( MAF: 37 . 5% vs . 33 . 7% , P = 6 . 6×10−4 , OR[95%CI] = 1 . 18[1 . 07–1 . 31] ) . After Bonferroni correction for multiple comparisons , the association of rs6677604 and rs16840639 with SLE remained significant in both EA and AA ( Table 1 ) . However , in the two smaller datasets ( 1 , 265 AS cases vs . 1 , 260 AS controls and 1 , 492 HS cases vs . 807 HS controls ) , we failed to detect significant association of these SNPs with SLE ( Table S1 ) . Of note , we did not detect significant association of I62V , Y402H and D936E with SLE in any of the four datasets ( Table S1 ) . A474A was associated with risk of SLE in EA ( P = 0 . 015 before correction , OR[95%CI] = 1 . 09[1 . 02–1 . 16] ) , but it was not confirmed in the other three ethnic groups ( Table S1 ) . To localize the underlying causal variant , we compared all SLE-associated SNPs ( P<0 . 05 ) identified in EA and AA and carried out linkage equilibrium ( LD ) analysis . Fourteen SNPs , spanning from intron 6 of CFH to the 3′ region downstream of CFHR5 , were associated with SLE in EA . However , only 4 of 14 SNPs , spanning from intron 6 of CFH to the 5′ region upstream of CFHR4 , showed consistent association with SLE in AA , suggesting a smaller SLE risk region . Of interest , within the risk region , rs6677604 and rs16840639 exhibited the strongest association with SLE in EA and AA , respectively . We found that rs6677604 and rs16840639 were in strong LD with each other in both EA ( r2 = 0 . 96 ) and AA ( r2 = 0 . 77 ) . Haplotype analysis showed that rs6677604 and rs16840639 could be defined into a ∼171 kb block in EA and a smaller ∼146 kb block in AA , respectively ( Figure 1B ) . The minor allele of rs6677604 or rs16840639 perfectly tagged two SLE risk haplotypes in EA ( H1: 16 . 1% vs . 14 . 1% , P = 0 . 0010; H2: 6 . 7% vs . 5 . 7% , P = 0 . 014 ) , and the minor allele of rs16840639 perfectly tagged the only risk haplotype in AA ( H1: 35 . 5% vs . 32 . 2% , P = 0 . 0028 ) ( Figure 2 ) . Using the conditional haplotype-based association test , we showed that after conditioning on rs6677604 or rs16840639 significant associations of all other SNPs were eliminated in both EA and AA ( Table 1 ) , which suggested that rs6677604 and rs16840639 could account for all association signals in the CFH-CFHRs region . Due to the strong LD between rs6677604 and rs16840639 , the conditional test could not be applied to further distinguish their association signals . To compare between rs6677604 and rs16840639 , we combined their ORs detected in EA and AA to generate a meta-analysis P value . The combined P value of rs6677604 ( Pmeta = 6 . 6×10−8 , OR[95%CI] = 1 . 18[1 . 11–1 . 26] ) was stronger than that of rs16840639 ( Pmeta = 2 . 9×10−7 , OR[95%CI] = 1 . 17[1 . 10–1 . 2] ) . Taken together , these data suggested that the underlying causal variant of SLE was captured by two strongly SLE-associated SNPs rs6677604 and rs16840639 in this study , which might reside in a ∼146 kb block . Neither rs6677604 nor rs16840639 are located in genomic regions with known biological function , which prompted us to seek other likely causal variants within the SLE-associated block . CFHR3-1Δ is a likely functional variant within the ∼146 kb SLE-associated block ( as shown in Figure 1A and 1B ) , which results in the deletion of CFHR3 and CFHR1 and has been associated with AMD and aHUS [19] , [20] . Because co-segregation of the CFHR3-1Δ deletion with the minor allele of rs6677604 in subjects with European Ancestry was observed in a previous study of AMD [20] , we hypothesized that the association of CFHR3-1Δ with SLE was captured by SNPs in this study . Using multiplex ligation-dependent probe amplification ( MLPA ) ( location of MLPA markers were shown in Figure 1A ) , we genotyped CFHR3-1Δ in 275 EA , 106 AA , 282 AS and 196 HS subjects , and then measured its LD with rs6677604 . We found that CFHR3-1Δ and rs6677604 were in complete LD in EA ( r2 = 1 . 00 ) and AS ( r2 = 1 . 00 ) , strong LD in HS ( r2 = 0 . 85 ) and moderate LD in AA subjects ( r2 = 0 . 60 ) ( Table 2 ) . In a subset of 58 unrelated AA subjects who were genotyped at both rs6677604 and rs16840639 , we found that CFHR3-1Δ was in stronger LD with rs16840639 ( r2 = 0 . 70 ) than with rs6677604 ( r2 = 0 . 60 ) . These results indicated that the association of the CFHR3-1Δ deletion with risk of SLE was tagged by the minor allele of rs6677604 in EA and rs16840639 in AA , respectively , suggesting that CFHR3-1Δ might be a risk variant for SLE . We showed that rs6677604 and CFHR3-1Δ were in the same block in AS ( Figure 1B ) , and the minor allele of rs6677604 could perfectly tag the CFHR3-1Δ deletion ( r2 = 1 . 00 ) . Thus , the lack of significant association of rs6677604 with SLE in our previous AS dataset might be due to insufficient statistical power . To increase power , we further genotyped 787 Chinese SLE cases and 1065 Chinese controls and then assessed the association of rs6677604 with SLE in an enlarged AS dataset ( 2052 cases vs . 2325 controls ) . In the enlarged AS dataset , we detected the significant association of rs6677604 with SLE ( MAF: 7 . 1% vs . 6 . 1% , P = 0 . 0485 , OR[95%CI] = 1 . 19[1 . 00–1 . 40] ) , supporting the hypothesis that CFHR3-1Δ might also be a risk variant for SLE in the AS population . To test whether homozygous deletion of CFHR3-1Δ might confer a higher risk of SLE than heterozygous deletion , we compared the genotypic frequency of homozygous and heterozygous deletion to that of no deletion , respectively . In EA , using rs6677604 as a tag SNP , we found that the homozygous deletion of CFHR3-1Δ conferred a significantly increased risk of SLE ( P = 7 . 5×10−4 , OR[95%CI] = 1 . 47[1 . 17–1 . 84] ) compared to no deletion , which was stronger than that of the heterozygous deletion ( P = 0 . 0018 , OR[95%CI] = 1 . 17[1 . 06–1 . 29] ) ( Table 3 ) , suggesting a dosage dependent risk effect of the CFHR3-1Δ deletion . To confirm , we compared genotypic associations of CFHR3-1Δ in AS and AA using rs6677604 and rs16840639 as tag SNPs , respectively . In these two ethnic groups , we found that only homozygous deletion of CFHR3-1Δ conferred a significantly increased risk of SLE compared to no deletion ( AS: P = 0 . 0021 , OR[95%CI] = 3 . 30[1 . 47–7 . 41]; AA: P = 0 . 0011 , OR[95%CI] = 1 . 40[1 . 14–1 . 71] ) ( Table 3 ) , supporting the hypothesis that homozygous deletion of CFHR3-1Δ conferred a higher risk of SLE than heterozygous deletion . In a meta-analysis combining ORs of EA , AA and AS , we confirmed that the homozygous deletion of CFHR3-1Δ ( Pmeta = 3 . 2×10−7 , OR[95%CI] = 1 . 47[1 . 27–1 . 71] ) had a stronger association with risk of SLE than the heterozygous deletion ( Pmeta = 3 . 5×10−4 , OR[95%CI] = 1 . 14[1 . 06–1 . 23] ) . SLE is a complex disease with heterogeneous sub-phenotypes . To determine whether CFHR3-1Δ had a stronger association with specific clinical manifestations of SLE , we compared its frequency in SLE cases stratified by the presence or absence of each of the eleven ACR classification criteria ( malar rash , discoid rash , photosensitivity , oral ulcers , arthritis , serositis , renal disorder , neurologic disorder , hematologic disorder , immunologic disorder and antinuclear antibody ) and five autoantibodies ( anti-dsDNA , anti-Sm , anti-RNP , anti-SSA/Ro and anti-SSB/La ) . In EA , we found that tag SNP rs6677604 of CFHR3-1Δ was associated with the absence of neurologic disorder ( Table S2 ) . However , in AA , we found that the corresponding tag SNP rs16840639 was associated with the absence of anti-dsDNA and the presence of serositis ( Table S2 ) , the latter of which was found not to be significant after Bonferroni correction for multiple comparisons . Insufficient clinical information for the majority of AS SLE patients precluded us from conducting these analyses . Taken together , these data did not provide evidence for a stronger association of CFHR3-1Δ with specific clinical manifestations of SLE .
In this study , we identified SLE-associated SNPs in the CFH-CFHRs region in three ethnic groups consisting of EA , AA and AS . In addition , we showed that the underlying causal variant was captured by rs6677104 and rs16840639 and could be localized to a ∼146 kb block extending from intron 9 of CFH to the 5′ region upstream of CFHR4 . We demonstrated that the CFHR3-1Δ deletion , which has been associated with AMD and aHUS , could be tagged by the minor risk alleles of rs6677604 ( r2 = 1 . 00 in EA and AS ) and rs16840639 ( r2 = 0 . 70 in AA ) and showed dosage-dependent association with risk of SLE . These data strongly suggested that CFHR3-1Δ , which leads to reduced levels of CFHR3 and CFHR1 proteins , was the causal variant for increased risk of SLE within the SLE-associated block . Multiple CFH exonic SNPs have been associated with various human diseases , but none of them were associated with SLE in this study . Y402H ( rs1061170 ) is the most studied non-synonymous SNP of CFH . Y402H is located in SCR7 and affects the binding of CFH with glycosaminoglycans and CRP [22] , [23] , [24] . Y402H has been strongly associated with risk of AMD and MGPN2 but not associated with aHUS [16] . In this study , we genotyped a tag SNP of Y402H ( rs7529589 , r2 = 0 . 75 with Y402H according to HapMap CEU data ) and detected no statistically significant association with SLE ( Table S1 ) . In a previous study , we had genotyped Y402H directly in 2033 EA cases and 2824 EA controls , and observed a similar result ( 37 . 4% vs . 37 . 7% , P = 0 . 81 , OR = 0 . 99 ) . I62V ( rs800292 ) located in the N-terminal SCR2 is another well-studied non-synonymous SNP of CFH . Although I62V may result in increased binding of CFH with C3b and enhanced CFH co-factor activity and has been associated with decreased risk of AMD , MPGN II and aHUS [16] , [25] , it was not associated with SLE in this study ( Table S1 ) . D936E ( rs1065489 in SCR16 ) was associated with lower host susceptibility to meningococcal disease in a recent GWAS [18] . We genotyped a perfect tag SNP ( rs10489456 ) of D936E and failed to detect an association with SLE ( Table S1 ) . A synonymous SNP A474A ( rs2274700 in SCR8 ) and its tag SNP rs1410996 were strongly associated with risk of AMD independent of Y402H [26] , [27] , but we detected only a marginal association between rs1410996 and risk of SLE in EA ( Table 1 ) , which was eliminated after conditioning on rs6677604 or rs16840639 . In addition , two synonymous SNPs A307A ( rs1061147 in SCR5 ) and Q672Q ( rs3753396 in SCR13 ) that are in strong LD with Y402H and D936E , respectively , were not associated with SLE in our study . These data suggest that the previously described disease-associated CFH exonic SNPs do not contribute to the development of SLE . Compared with SNP genotyping assays , genotyping assays for copy number variation are more labor-intensive and costly . Consequently , CFHR3-1Δ was not specifically genotyped in this study to assess its association with SLE . Instead , we evaluated the effect of the CFHR3-1Δ deletion on SLE development indirectly using tag SNPs that were in strong LD with it . We first confirmed that CFHR3-1Δ was in strong LD with rs6677604 in EA , similar to previous studies of AMD [20] , [28] . Furthermore , we showed that CFHR3-1Δ was also in strong LD with rs6677604 in AS and HS . In addition , we found that CFHR3-1Δ was in stronger LD with rs16840639 than with rs6677604 in AA . Of note , in AA , the most significant association with SLE was detected at rs16840639 rather than rs6677604 , and the risk haplotype H1 in AA was perfectly tagged by the minor allele of rs16840639 rather than rs6677604 ( Figure 2 ) , suggesting that rs16840639 captured the underlying causal variant CFHR3-1Δ in AA . Using these tag SNPs , we deduced that homozygous CFHR3-1Δ deletion conferred higher risk of SLE than heterozygous deletion , which suggested a change in gene dosage of the encoded proteins CFHR3 and CFHR1 might account for the increased SLE risk . The CFHR3-1Δ deletion was associated with the general phenotype of SLE but did not consistently exhibit stronger signals to a specific clinical manifestation in EA and AA , and was not specifically associated with the presence of renal disorder . This is in contrast to the effect of CFH deficiency , which results in the development of glomerulonephritis in CFH knockout mice due to uncontrolled C3 activation [11] , [29] . In addition , the absence of CFH in plasma causes human MPGN II [30] , but an association of the CFHR3-1Δ deletion with MPGN II has not been reported . The absence of an association of CFHR3-1Δ with renal disorder in lupus suggests that CFHR3 and CFHR1 play a different role from CFH in the pathogenesis of lupus , although further studies are required to validate the lack of association between the CFHR3-1Δ deletion and renal disorder in SLE . The CFHR3-1Δ deletion has opposite effects in different diseases [9] , and the underlying mechanism is poorly understood . Activated complement pathways converge to generate C5 convertase , which cleaves C5 into C5a and C5b . C5a is a potent chemoattractant . C5b initiates the formation of the terminal MAC . CFHR1 acts as a complement regulator to inhibit C5 convertase activity and terminal MAC formation [10] , and CFHR3 displays anti-inflammatory effects by blocking C5a generation and C5a-mediated chemoattraction of neutrophils [31] . Increased neutrophils lead to inflammatory injuries in many non-infectious human diseases [32] . It has been shown that immune complex-induced inflammatory injuries are largely mediated by C5a receptor and blocking C5a receptor reduces manifestation of lupus nephritis in mice [33] , [34] . In addition , increased apoptotic neutrophils contribute to autoantigen excess and have been associated with increased disease activity in SLE [35] . The CFHR3-1Δ deletion results in decreased CFHR3 and CFHR1 levels and may therefore lead to uncontrolled production of chemoattractant C5a predisposing to SLE . Of interest , the CFHR3-1Δ deletion also has a risk effect in aHUS and the CFHR3 and CFHR1 deficiency in plasma has been associated with the presence of anti-CFH autoantibodies , which bind to the C-terminus of CFH and block CFH binding to cell surfaces [36] , [37] . It is also possible that CFHR3-1Δ is also associated with the presence of anti-CFH autoantibodies in SLE and thus leads to impaired CFH cell surface regulation . Both CFHR3 and CFHR1 , lacking the CFH N-terminus regulatory activity , were reported to compete with CFH for binding to C3b , and thus CFHR3 and CFHR1 deficiency may lead to enhanced CFH regulation [31] , which may explain the protective effect of the CFHR3-1Δ deletion in AMD . Of interest , as mentioned before , the non-synonymous SNP I62V in the CFH regulatory domain may also increase CFH regulation . I62V confers a protective effect in AMD , aHUS and MPGN II [16] , but it was not associated with SLE in this study . Statistical under-powering might account for the failure to detect a significant association in HS dataset . First , rs6677604 and CFHR3-1Δ were in strong LD and could be defined into a block in HS , which excluded the possibility that the CFHR3-1Δ deletion was not tagged in the HS dataset . In addition , there was no genetic heterogeneity of rs6677604 in the four ethnic groups ( P = 0 . 76 ) , in which the risk minor allele showed consistently higher frequency in cases than in controls . Finally , based on rs6677604 , post hoc analysis indicated a much lower power of 51% in HS to detect association with SLE ( P<0 . 05 ) than the power of 98% in EA and 92% in AA . Thus , the association of CFHR3-1Δ with SLE in HS needs to be further evaluated in a larger dataset . One limitation of this study is that we have not addressed whether rare variants in the CFH-CFHRs region may contribute to the development of SLE . Pathogenic rare variants clustering in CFH C-terminus affect CFH cell surface binding , but they were only found in aHUS patients , not in AMD , MPGN II patients and healthy controls [16] . Deep sequencing of exons in CFH C-terminus in patients with SLE may elucidate whether these rare variants are associated with SLE . To our knowledge , this study is the first to show that genetic variants in the CFH-CFHRs region are associated with SLE susceptibility . Our consistent observations of dose-dependent association between CFHR3-1Δ and SLE across three distinct ancestral populations and no association in CFH exonic SNPs suggest a novel role for CFHR3 and CFHR1 in the pathogenesis of SLE . Further functional studies are required to elucidate the underlying mechanism of CFHR3-1Δ .
The study was approved by the Human Subject Institutional Review Boards or the ethnic committees of each institution . All subjects were enrolled after informed consent had been obtained . To test the association of CFH and CFHRs with SLE , we used a large collection of samples from case-control subjects from multiple ethnic groups . These samples were from the collaborative Large Lupus Association Study 2 ( LLAS2 ) and were contributed by participating institutions in the United States , Asia and Europe . According to genetic ancestry , subjects were grouped into four ethnic groups including European American ( 3 , 936 cases vs . 3 , 491 controls ) , African American ( 1 , 679 cases vs . 1 , 934 controls ) , Asian ( 1 , 265 cases vs . 1 , 260 controls ) and Hispanic enriched for the Amerindian-European admixture ( 1 , 492 cases vs . 807 controls ) . Asians were comprised of Koreans ( 884 cases vs . 994 controls ) , Chinese ( 200 cases vs . 205 controls ) and subjects from other East Asian countries such as Japan and Singapore ( 181 cases vs . 61 controls ) . African Americans included 275 Gullahs ( 152 cases vs . 123 controls ) , who are subjects with African Ancestry . To test LD between CFHR3-1Δ and SLE-associated SNPs , we used 275 unrelated European Americans ( 187 cases vs . 88 controls ) , 106 African Americans ( 88 unrelated subjects [58 cases vs . 30 controls] and 18 subjects from 6 SLE trios families ) , 282 unrelated Chinese ( 218 cases vs . 64 controls ) and 196 Hispanics ( 157 unrelated subjects [91 cases vs . 66 controls] and 39 subjects from 13 SLE trios families ) . All of these subjects were enrolled from UCLA . To enlarge the sample size of Asians for association test , we used 1 , 852 Chinese case-control subjects ( 787 vs . 1065 ) recruited from Shanghai Renji Hospital , Shanghai Jiao Tong University School of Medicine . All SLE patients met the American College of Rheumatology ( ACR ) criteria for the classification of SLE [38] . LLAS2 samples were processed at the Lupus Genetics Studies Unit of the Oklahoma Medical Research Foundation ( OMRF ) . SNP genotyping was carried out on the Illumina iSelect platform . Subjects with individual genotyping call rate <0 . 90 were removed because of low data quality . Subjects that were duplicated or first degree related were also removed . Both principal component analysis and global ancestry estimation based on 347 ancestry informative markers were used to detect population stratification and admixture , as described in another LLAS2 report [39] . After removing genetic outliers , a final dataset of 15 , 864 unrelated subjects ( 8 , 372 cases vs . 7 , 492 controls ) was obtained . Taqman SNP genotyping assay ( Applied Biosystems , California , USA ) was used to genotype rs6677604 for subjects who were not recruited into LLAS2 . MLPA kit “SALSA MLPA KIT P236-A1 ARMD mix-1” was used to genotype the CFH-CFHRs region according to the manufacture's instruction ( MRC-Holland , Amsterdam , The Netherlands ) . ABI 3730 Genetic Analyzer ( Applied Biosystems ) was used to run gel electrophoresis . Software Peak Scanner v1 . 0 ( Applied Biosystems ) was used to extract peaks generated in electrophoresis . Coffalyser v9 . 4 ( MRC-Holland ) was used to readout copy number of target region . The HWE test threshold was set at P>0 . 01 for controls and P>0 . 0001 for cases . SNPs failing the HWE test were excluded from association test . SNPs showing genotyping missing rate>5% or showing significantly different genotyping missing rate between cases and controls ( missing rate>2% and Pmissing<0 . 05 ) were also excluded from association test . In allelic association test ( Pearson's χ2–test ) , the significance level was set at P<0 . 05 . Haploview 4 . 2 was used to estimate pairwise LD values between SNPs , define haplotypes blocks and calculate haplotypic association with SLE . Haplotype-based conditional association analysis was carried out by Plink v1 . 07 . Mantel-Haenszel analysis was performed to generate the meta-analysis P value . CaTS was used to calculate statistical power .
|
Systemic lupus erythematosus ( SLE ) is a complex autoimmune disease , associated with increased complement activation . Previous studies have provided evidence for the presence of SLE susceptibility gene ( s ) in the chromosome 1q31-32 locus . Within 1q32 , genes encoding complement regulator factor H ( CFH ) and five CFH-related proteins ( CFHR1-CFHR5 ) may contribute to the development of SLE , because genetic variants of these genes impair complement regulation and predispose to various human diseases . In this study , we tested association of genetic variants in the region containing CFH and CFHRs with SLE . We identified genetic variants predisposing to SLE in European American , African American , and Asian populations , which might be attributed to the deletion of CFHR3 and CFHR1 genes but not previously identified disease-associated exonic variants of CFH . This study provides the first evidence for consistent association between CFH/CFHRs and SLE across multi-ancestral SLE datasets , providing new insights into the role of complement regulators in the pathogenesis of SLE .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"complement",
"system",
"clinical",
"immunology",
"genetic",
"association",
"studies",
"lupus",
"erythematosus",
"genetics",
"autoimmune",
"diseases",
"immunology",
"biology",
"human",
"genetics",
"genetics",
"and",
"genomics",
"immune",
"system"
] |
2011
|
Association of Genetic Variants in Complement Factor H and Factor H-Related Genes with Systemic Lupus Erythematosus Susceptibility
|
Leprosy , caused by Mycobacterium leprae , can lead to scarring and deformities . Human immunodeficiency virus ( HIV ) , a lymphotropic virus with high rates of replication , leads to cell death in various stages of infection . These diseases have major social and quality of life costs , and although the relevance of their comorbidity is recognized , several aspects are still not fully understood . Two cohorts of patients with leprosy in an endemic region of the Amazon were observed . We compared 40 patients with leprosy and HIV ( Group 1 ) and 107 leprosy patients with no comorbidity ( Group 2 ) for a minimum of 2 years . Group 1 predominantly experienced the paucibacillary classification , accounting for 70% of cases , whereas Group 2 primarily experienced the multibacillary classification ( 80 . 4% of cases ) . There was no significant difference in the prevalence of leprosy reactions among the two groups ( 37 . 5% for Group 1 vs . 56 . 1% for Group 2 ) , and the most frequent reaction was Type 1 . The appearance of Group 1 patients’ reversal reaction skin lesions was consistent with each clinical form: typically erythematous and infiltrated , with similar progression as those patients without HIV , which responded to prednisone . Patients in both groups primarily experienced a single episode ( 73 . 3% in Group 1 and 75% in Group 2 ) , and Group 1 had shorter reaction periods ( ≤3 months; 93 . 3% ) , moderate severity ( 80% ) , with 93 . 3% of the patients in the state of acquired immune deficiency syndrome , and 46 . 7% presenting the reaction at the time of the immune reconstitution inflammatory syndrome . This study used a large sample and makes a significant contribution to the clinical outcomes of patients in the reactive state with comorbid HIV and leprosy . The data indicate that these diseases , although concurrent , have independent courses .
Leprosy , a chronic infectious disease caused by Mycobacterium leprae , can cause scars and deformities , especially if not treated quickly [1] . Brazil is currently responsible for approximately 92% of leprosy cases in the Americas , and is ranked second , behind India , in the number of global cases [2] . Despite the number of detected leprosy cases in the country remaining stable , the North , Midwest , and Northeast regions are the most heavily affected , in proportion to the population [3] . Human immunodeficiency virus ( HIV ) is a lymphotropic virus belonging to the Retroviridae family , which maintains high rates of viral replication , causing cell death in all infection stages [4] . Early diagnosis and clinical management of HIV and its complications are often complex . With the advent of antiretroviral therapy , there has been great improvement in the prognosis and quality of life of people living with HIV [5] . However , due to the increased number of people living with this virus , HIV prevalence continues to increase even in leprosy-endemic countries , which increases the risk of comorbidity [6] . Since the first report of a comorbid infection in a patient with HIV and M . leprae , several questions have been raised regarding the consequences of their interaction , especially considering the direct involvement of T-helper CD4+ lymphocytes in the pathogenesis of both diseases . Early records of this co-infection theory reported that patients developed serious forms of infection due to their immune suppression caused by HIV; however , many studies have shown no or limited alterations in the course of patients with a leprosy and HIV comorbidity [7] . Regarding the interaction conditions of the two infections , a decrease in frequency and intensity was expected , since these are both immune-mediated phenomena . However , research and reports on the subject have shown the continued occurrence of leprosy , including recent data showing that co-infected patients had stronger reactions to the diagnosis ( 31 . 5% vs . 18 . 8% ) compared with the group without HIV [8] . However , during the vigilance period of reaction rates in groups , both were similar ( 59 . 3% vs . 53 . 1% ) . Neural damage was also expected since HIV patients are also at risk of developing lesions in their generalized peripheral nerves , including mono-neuropathy and peripheral neuritis multiplex through both HIV infection and the treatment itself [9] . The introduction of antiretroviral therapy has created , in itself , a new clinical syndrome , which is called reconstitution inflammatory syndrome or immune reconstitution inflammatory syndrome . This syndrome affects HIV-positive patients who are in an advanced stage of the disease ( CD4 <200/ml ) . In these patients , the clinical signs of inflammation associated with opportunistic infections are usually an immune response during the transition process , in which the viral load decreases and T-helper CD4+ cell count increases by more than 20% [7 , 10–14] . Several authors describe the leprosy reaction at the start of the clinical manifestation of leprosy , as part of a demonstration of immune reconstitution inflammatory syndrome [7 , 15–19] . From 2002 to 2009 , 21 cases of reversal reactions occurred as a manifestation of immune reconstitution inflammatory syndrome; of these 21 cases , 13 were diagnosed in Brazil [6] . Altogether , these cases have been reported primarily in areas where antiretroviral therapy is no longer available: 70% in South America ( with 58% being from Brazil ) and 20% in India . Lockwood and Lambert proposed a definition of a leprosy immune reconstitution inflammatory syndrome event in 2010 [9] to facilitate its correct identification . This can even be recognized by the following characteristics: ( 1 ) clinical symptoms of leprosy and/or leprosy reaction starting within 6 months of antiretroviral therapy; ( 2 ) advanced HIV infection; ( 3 ) CD4+ cell counts <200 cells/mm3 before initiating highly active antiretroviral therapy; and ( 4 ) increased CD4+ cells in the peripheral blood after highly active antiretroviral therapy [6] . By observing the magnitude of the two diseases in northern Brazil , particularly in the state of Pará , research on leprosy and HIV comorbidity allows the monitoring of the clinical outcomes of patients , including observations of reaction aspects of the phenomena .
This research was approved by the Ethics Committee on Human Research at the Núcleo de Medicina Tropical da Universidade Federal do Pará under the protocol 001/2011 . Adult subjects had provided written informed consent in agreeing to participate in the study . For those participants who were under 18 years old , a written the informed consent was provided by their parents or guardians .
Of the 40 patients included in Group 1 and 107 patients in Group 2 , 67 . 5% and 67 . 3% , respectively , were male; the predominant age group was 31–59 years for both groups , with the average age being 37 years ( Table 1 ) . The paucibacillary form accounted for 70% of cases found in Group 1 , while the multibacillary form accounted for 80 . 4% of cases in Group 2 ( chi-square , p < 0 . 0001 ) . Patients without HIV infection were more likely to evolve into the multibacillary form of leprosy compared to co-infected patients ( RR = 3 . 0 ) ( Table 1 ) . In Group 1 , the predominant clinical presentation was borderline tuberculoid in 45% of cases , while in Group 2 , the clinical type borderline was primarily expressed in 40 . 2% of cases ( G test , p < 0 . 0001 ) ( Table 1 ) . Only 37 . 5% of patients in Group 1 had a leprosy episode , while 56 . 1% of patients in Group 2 did ( chi-square , p = 0 . 0026 ) . Comorbid patients were less likely to experience leprosy reactions ( RR = 0 . 47 ) ( Table 2 ) . In both groups , the most frequent response was the Type 1 or reversal reaction , accounting for 86 . 7% and 56 . 6% of cases in Groups 1 and 2 , respectively ( G test , p = 0 . 0750 ) . Acute neuritis was observed in 17 . 5% and 25 . 2% of patients in Groups 1 and 2 , respectively , with no statistical difference between the two groups ( Table 2 ) . The treatment of the reaction manifestation was the same in both groups , with prednisone being the drug of choice for the reversal reaction , at a dose adjusted to 1 mg/kg/day; 14 and 36 patients from Groups 1 and 2 , respectively , used this dosage of the drug . In the patients with a Type 2 response , thalidomide was the drug of choice ( Table 2 ) . Regarding the clinical form of the leprosy reactions , it was observed that , in comorbid patients , borderline tuberculoid was the predominant clinical presentation among patients experiencing the Type 1 reaction , at 61 . 6% of patients . Only 2 patients were observed to develop the Type 2 response; both belonged to the group presenting the borderline lepromatous manifestation . Patients without comorbid infections had both Type 1 and 2 reactions; the predominant clinical form in this cohort was the borderline borderline , with 57 . 9% of Type 1 patients and 38 . 5% of Type 2 patients displaying this form ( Table 3 ) . The most prevalent type of reaction in both groups was the reversal reaction . The group of comorbid patients experienced skin lesions consistent with the expectations for each clinical presentation; the lesions were erythematous and infiltrated , with a similar progress and outcome as those found in patients without HIV , and responded appropriately to the use of prednisone . Generally , after 30 days of prednisone , dose-adjusted for weight , the patients' lesions had no infiltration and were in regression . Three of the 13 co-infected patients with reaction showed ulcerated lesions , but also had a good response to prednisone in the expected period of time . The same kind of ulcerated lesion in the Type 1 reaction was observed in 5 patients with leprosy alone . The two coinfected patients experiencing Type 2 reactions showed the classic clinical manifestations , with widespread painful erythematous nodules on the body , as well as fever and arthralgia . Three of the patients without co-infection had ulcerated Type 2 reactions . In those patients with HIV and leprosy experiencing a leprosy reaction , 14 ( 93 . 3% ) were in the acquired immune deficiency syndrome ( AIDS ) state ( chi-square , p = 0 . 0239 ) , all of them on highly active antiretroviral therapy ( G test , p = 0 . 0439 ) . Another 7 patients ( 46 . 7% ) had leprosy with a reversal reaction episode while experiencing immune reconstitution syndrome ( G test , p = 0 . 0855 ) ( Table 4 ) . In these patients , we were able to quantify and observe a significant increase of serum T-helper CD4+ cells at the time of HIV diagnosis , prior to initiation of highly active antiretroviral therapy ( average 141 . 8 ) , compared to the time of diagnosis of the leprosy reactional state ( average 367 . 7 ) ( t-test , p = 0 . 0088 ) . In the first 6 months of observation , there were more leprosy reactions in both groups . At the end of these sixth months , 67 . 5% of Group 1 patients did not have any kind of leprosy reaction , neither did 74 . 77% of Group 2 patients . After the 24 months of observation , both groups behaved similarly and remained stable; 65% and 63 . 55% of patients in Groups 1 and 2 , respectively , had had no reaction during this time , with no new patients experiencing reactions as of 18 months since the beginning of multidrug therapy . Most patients ( 73 . 3% in Group 1 and 75% in Group 2 ) experienced only one cycle of leprosy reaction . However , no patient in Group 1 had more than three episodes , while 10 ( 16 . 7% ) patients in Group 2 did ( G test , p = 0 . 0371 ) ( Table 5 ) . Further , 93 . 3% of patients in Group 1 had relatively short leprosy reaction cycles , of ≤3 months , while 63 . 3% of patients in Group 2 experienced longer cycles of over 3 months' duration ( p >0 . 0001 ) . Patients without co-infection were more likely to have reactional states of over 3 months compared to the comorbid patients ( RR = 7 . 5 ) ( Table 5 ) . As for reaction severity , most of the patients in both groups showed episodes of moderate severity: 80% in Group 1 and 54 . 2% in Group 2 ( G test , p = 0 . 1577 ) ( Table 5 ) . Most of the patients in the two groups also had recurrent episodes , representing 80% and 58 . 3% of patients in Groups 1 and 2 , respectively ( G test , p = 0 . 6540 ) ( Table 5 ) .
The follow-up study of two clinical cohorts of leprosy patients , one experiencing comorbidity with HIV/AIDS and the other not , found that the observed dermatological lesions had a usual aspect with no significant difference between groups , and good clinical progress with the administration of prednisone , the preferred therapeutic for leprosy reaction . There was no significant difference in the prevalence of leprosy reactions between the two groups , and both groups predominantly experienced a Type 1 response , with only one reaction event . Co-infected patients exhibited moderate reaction severity , with predominantly shorter cycles . Although many questions remain in the study of leprosy and HIV comorbidity , particularly regarding leprosy reactions , this work provides information able to confirm assertions that such diseases , when concurrent , are independent in their progression . Future studies may wish to further examine the relationships between the two diseases to corroborate these conclusions .
|
Leprosy and HIV infections , separately , are serious modern public health problems . Many studies have been conducted on these diseases , but knowledge gaps remain . This article provides the first account of important clinical information on a significant sample of patients with leprosy , as well as patients with both leprosy and HIV , who were followed over a period of 24 months . We compared the clinical outcome of both groups , observed the occurrence of reactional episodes , and examined the characteristics of these episodes . The sample consisted of 40 co-infected patients ( Group 1 ) and 107 patients with leprosy only ( Group 2 ) . Group 1 was characterized by high levels of paucibacillary leprosy cases ( 70% ) and the borderline tuberculoid clinical form ( 45% ) , while Group 2 predominantly exhibited multibacillary leprosy ( 86% ) and the borderline clinical form ( 40 . 2% ) . The Type I reaction was present in 13 and 34 patients of Groups 1 and 2 , respectively . The Amazon region , where the study was conducted , is an endemic region for both diseases , which can be useful for conducting studies such as these owing to the generalizability of the results . This study seeks to contribute to the knowledge of the natural history of HIV and leprosy comorbidity .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[] |
2015
|
Leprosy Reactions in Patients Coinfected with HIV: Clinical Aspects and Outcomes in Two Comparative Cohorts in the Amazon Region, Brazil
|
Chagas disease is one of the most important endemic parasitic diseases in Latin America . In its chronic phase , progression to cardiomyopathy has high morbidity and mortality . The persistence of a normal electrocardiogram ( ECG ) provides a similar prognosis to that of a non-diseased population . Benznidazole ( BNZ ) is the only drug with trypanocidal action available in Brazil . A group of 310 patients with chronic Chagas disease who had normal ECGs at the first medical visit performed before 2002 were included . There were 263 patients treated with BNZ and 47 untreated . The follow-up period was 19 . 59 years . Univariate analyses showed that those treated were younger and predominantly male . As many as 79 . 08% of those treated and 46 . 81% of those untreated continued with normal electrocardiograms ( p <0 . 0001 ) . The occurrence of electrocardiographic abnormalities and relevant clinical events ( heart failure , stroke , total mortality , and cardiovascular death ) was less prevalent in treated patients ( p <0 . 001 , p: 0 . 022 , p: 0 . 047 respectively ) . In multivariate analyses , the parasiticide treatment was an independent variable for persistence of a normal ECG pattern , which was an independent variable in the prevention of significant clinical events . The immunofluorescence titers decreased with the parasitological treatment . However , the small number of tests in untreated patients did not allow the correlation of the decrease of these titers with electrocardiographic alterations . These data suggest that treatment with benznidazole prevents the occurrence of electrocardiographic alterations . On the other hand , patients who develop ECG abnormalities present with more significant clinical events .
Chagas’ disease ( CD ) , described by Carlos Chagas in 1909[1] , and caused by a parasite–Trypanosoma cruzi , is one of the most important endemic diseases in Latin America , where there are 10 million people infected ( about two million in Brazil ) . The vectorial transmission has historically been the most important . The disease may also be conveyed by blood transfusion , be congenital , or be transmitted orally ( this is the most prevalent today in Brazil ) , among other types of transmission[2][3] . With globalization , many Latin Americans migrated to other continents , carrying this disease and transmitting it through blood transfusion to the inhabitants of non-endemic countries . Therefore , CD is now present in North America , Europe , Asia , and Oceania , and is becoming a worldwide public health problem[4] . After contamination , the acute phase occurs , characterized by severe inflammation and intense parasitemia , although with limited clinical impact and low mortality . This phase lasts for approximately 8 to 10 weeks , followed by the chronic phase with a decrease of parasitemia and inflammation , but not to extinction . Sixty to 70% of patients remain in the indeterminate form ( positive serum reaction , no clinical signs , normal electrocardiogram , normal Chest X-ray , normal esophagogram , and normal barium enema ) . A total of 40 or 30% evolve to clinical forms , with isolated or concomitant heart , esophagus , and colon involvement[2] . The electrocardiogram ( ECG ) is a very important tool in monitoring patients with CD . Maguire et al[5] , following a population of CD patients for seven years , showed that those who maintained a normal ECG , evolved in a similar way to individuals without the disease . This simple test has important prognostic value , and usually is sufficient for clinical follow-up[6][7] . The parasite's role in the chronic phase remains unclear , even one hundred years after the description by Carlos Chagas[8][9][10] Parasiticide treatment is controversial as to its indication in the chronic phase and as to its real benefits . The criteria for assessing the possible medication benefits and certainty of a cure are not unanimous among authors . Benznidazole is the only drug in Brazil with proven parasiticide action . It is available in 100 mg tablets and the dose recommended for acute patients or children , is 10 mg/kg/day for 60 days of treatment , and in the chronic phase , 5 mg/kg/day , also for 60 days . Major side effects are dermatitis , which occurs in 30% of cases , and polyneuropathy , which is less prevalent . Patients usually tolerate well the side effects described . Significant leukopenia and liver damage are rarely observed , and the occurrence of agranulocytosis is exceptional[11][12][13][14] . The BENEFIT study that randomly evaluated the treatment with BNZ in 2854 ( 1431 BNZ and 1423 placebo ) patients with chronic Chagas cardiomyopathy , NYHA class I , II , III , ( 97% class I and II ) , followed for the short period of 5 . 4 years , showed a significant decrease in parasitemia ( PCR test ) in the BNZ group . However , no difference in the occurrence of events during this period ( death , resuscitated cardiac arrest , insertion of a pacemaker , or an implantable cardioverter–defibrillator ( ICD ) , sustained ventricular tachycardia , cardiac transplantation , new heart failure , stroke or transient ischemic attack , or a systemic or pulmonary thromboembolic event ) [15] . Immunofluorescence titers are stable in untreated patients . However , treated patients showed a decrease of the titers . Negativity of the immunofluorescence titers is infrequent , but may occur persistently after more than a decade of treatment[16][17] . This is a retrospective study that analyzes the electrocardiographic , clinical , and serological evolution of patients with chronic Chagas’ disease , with or without treatment with BNZ , and who had a previous normal ECG .
This study is registered at the Dante Pazzanese Institutional Ethics Committee . All adult patients gave written informed consent to participate in the study .
From a database with approximately 1500 patients with CD , 527 had a normal ECG at their first medical visit . Of these , 379 met the inclusion criteria . Three hundred and ten patients were found ( 81 . 80% ) , and 69 ( 18 . 20% ) could not be reached and were therefore excluded . Table 1 shows baseline characteristics of the group of patients found and the group not found . The presence of DLP ( described in the medical records ) was more prevalent in the group of found patients than in the group not found , with 27 . 40% and 2 . 90% , respectively ( p <0 . 001 ) . It was not possible to detect a statistically significant difference between groups for other variables . We followed the 310 patients included in the study for a period of 10 to 46 years ( 19 . 59 ± 6 . 46 ) , with a median of 18 years; 50% of these patients were followed for 15 to 23 years . Age at the last visit varied from 30 to 84 years ( 57 . 80 ± 10 . 07 ) . Only 107 ( 34 . 52% ) were male , and 231 ( 74 . 52% ) were white . Two hundred and sixty-three patients ( 84 . 84% ) received BNZ and 47 ( 15 . 16% ) did not . The characteristics of the two groups are shown on Table 2 . Treated patients were younger ( 56 . 07 years x 68 . 89 years , p <0 . 0001 ) , predominantly male ( 36 . 90% vs . 21 . 30% , p: 0 . 045 ) , had left the endemic area more recently ( 16 . 77 years vs 19 . 65 years , p: 0 . 012 ) , and 208 ( 79 . 08% ) maintained normal ECGs , compared to 22 ( 46 . 81% ) of the non-treated individuals ( p <0 . 0001 ) . Among the treated patients , 55 ( 20 . 92% ) had ECG changes , as follows: Right Bundle Branch Block in 21 ( 38% ) , nonspecific changes in ventricular repolarization in 20 ( 37% ) , and Blockage of the anterior superior division of the left branch in 11 ( 20% ) . Among untreated patients , 25 ( 53 . 19% ) had worsening of the ECG: Right Bundle Branch Block in two ( 8% ) , nonspecific changes in ventricular repolarization in four ( 16% ) , and Blockage of the anterior superior division of the left branch in six ( 24% ) . Other changes detected had low prevalence . There were no statistically significant differences between groups in the other variables . The side effects observed in treated patients were dermatitis in 92 patients ( 34 . 98% ) , polyneuropathy in 12 ( 4 . 56% ) , and others ( dyspepsia , insomnia , leukopenia less than 4000/mm3 ) in eight ( 3 . 04% ) . Twenty-six ( 9 . 89% ) patients abandoned treatment due to side effects . The analysis of relevant events ( heart failure , stroke , and cardiac death or due to any cause ) described in the medical records were HF in eight cases ( 2 . 58% ) , stroke in four ( 1 . 29% ) , and 12 deaths ( 3 . 87% ) , in which six were men and six were women . The date of occurrence of these events was not available; only if they had occurred or not . Therefore , a longitudinal analysis was not possible , so only the logistic regression was done . In six of them it was possible to assume that the cause was due to CD ( 1 . 93% of all patients studied ) . In only one case was it not possible to determine the cause of death . Among the 80 patients who had worsening of the ECGs , eight ( 10% ) died and among the 230 who maintained normal ECGs , four ( 1 . 7% ) died ( p: 0 . 002 ) . The cause of death related to CD occurred in five ( 6 . 25% ) patients with ECG alterations and in only one ( 0 . 43% ) with a normal ECG ( p: 0 . 001 ) . The eight cases of HF occurred in patients with ECG alterations . Among the four cases with stroke , two ( 2 . 5% of 80 ) had ECG alterations , and two ( 0 . 9% of 230 ) did not ( p = 0 . 274 ) . Combined outcomes ( HF , stroke , and death ) occurred in 24 cases ( 7 . 74% ) , 16 of them ( 20% ) with ECG alterations and eight ( 3 . 48% ) with normal ECGs ( p <0 . 0001 ) . Table 3 shows the occurrence of events in patients untreated and treated with BNZ . It shows that patients treated with BNZ had fewer cardiac deaths and fewer total deaths . There were two or more results of the immunofluorescence test in 171 patients ( Table 4 ) , 11 ( 6 . 43% ) untreated and 160 ( 93 . 57% ) treated . These results remained stable in untreated patients ( 232 . 72 ± 104 . 02 and 254 . 54 ± 93 . 41 ) , whereas in the treated individuals , the titers decreased ( 144 . 90 ± 109 . 80 and 70 . 25 ± 74 . 70: p < 0 . 0001 ) . In the 112 patients who remained with normal ECGs and without any relevant clinical outcomes , the titers of the first and last reactions were 127 . 50 ( ± 104 . 60 ) and 63 . 21 ( ± 65 . 95 ) , respectively ( p <0 . 001 ) . Titers decreased 39 . 93% in patients who had ECG alterations and 50 . 43% in those with normal ECGs ( p: 0 . 863 ) . The difference in years from first to last serology in those who had ECG alterations was 14 . 18 ( ± 4 . 09 ) years , and in those who maintained a normal ECG it was 14 . 04 ( ± 5 . 01 ) years ( p: 0 . 15 ) . The negativity of the immunofluorescence titer ( <1/40 ) occurred in 60 patients treated with BNZ ( 37 . 50% ) , with an average of 14 years follow-up , and in none of the untreated individuals . In the multivariate analysis ( Table 5 ) with dependent variables , the occurrence of combined events ( heart failure , stroke , and total mortality ) and independent variables , treatment with BNZ , follow-up time , males , white ethnicity , and age , it was observed that with the withdrawal of the ECG from this model , the parasiticide treatment was the only protection against events . Table 6 assesses another logistic regression model , analyzing the dependent variable , persistence of a normal ECG , with the independent variables , treatment with BNZ , follow-up time , male , white ethnicity , and age . In this model treatment with BNZ and white ethnicity favored the persistence of a normal ECG , while the evolution of time ( less than average ) favored the appearance of ECG alterations .
The adequate and appropriate assessment of the scientific hypothesis in question should be through a randomized controlled clinical trial , which did not happen in this study . However , the data obtained provides useful information because of the large number of patients evaluated during a two-decade follow-up . From the data obtained , it could be suggested that treatment with BNZ prevents the appearance of ECG abnormalities , and patients with normal ECGs have fewer combined events . Treatment with BNZ decreases immunofluorescence titers . Therefore , only 37 . 5% of the treated patients showed negativity of the immunofluorescence titers .
|
Twenty years of follow-up of patients with Chagas disease treated with benznidazole is presented in this paper . The persistence of a normal electrocardiogram ( ECG ) provides a similar prognosis to that of a non-diseased population . Benznidazole ( BNZ ) is the only drug with trypanocidal action available in Brazil . A group of 310 patients with chronic Chagas disease who had normal ECGs at the first medical visit performed before 2002 were included . There were 263 patients treated with BNZ and 47 untreated . The occurrence of electrocardiographic abnormalities and relevant clinical events ( heart failure , stroke , total mortality , and cardiovascular death ) was less prevalent in treated patients . In multivariate analyses , the parasiticide treatment was an independent variable for persistence of a normal ECG pattern , which was an independent variable in the prevention of significant clinical events . The immunofluorescence titers decreased with the parasitological treatment . However , the small number of tests in untreated patients did not allow the correlation of the decrease of these titers with electrocardiographic alterations . These data suggest that treatment with benznidazole prevents the occurrence of electrocardiographic alterations . On the other hand , patients who develop ECG abnormalities present with more significant clinical events .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"coronary",
"artery",
"disease",
"medicine",
"and",
"health",
"sciences",
"parasite",
"evolution",
"tropical",
"diseases",
"parasitic",
"diseases",
"parasitology",
"electrocardiography",
"bioassays",
"and",
"physiological",
"analysis",
"neglected",
"tropical",
"diseases",
"immunologic",
"techniques",
"cardiology",
"research",
"and",
"analysis",
"methods",
"epidemiology",
"immunoassays",
"protozoan",
"infections",
"electrophysiological",
"techniques",
"heart",
"failure",
"immunofluorescence",
"chagas",
"disease",
"cardiac",
"electrophysiology",
"biology",
"and",
"life",
"sciences",
"vascular",
"medicine",
"ethnic",
"epidemiology"
] |
2016
|
Evaluation of Parasiticide Treatment with Benznidazol in the Electrocardiographic, Clinical, and Serological Evolution of Chagas Disease
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.